重新思考 AI 时代的 SEO | Eli Schwartz(SEO 顾问、作者)
Rethinking SEO in the age of AI | Eli Schwartz (SEO advisor, author)
Meet the Guest
Lenny Rachitsky: You’ve noticed a significant shift in how SEO works with the rise of AI answers being integrated into search results.
Eli Schwartz: Transparently, I thought this was going to be an apocalypse. Up until AI Overviews, whoever won on that long form piece of content would get that first click. But now that doesn’t exist anymore.
How Search Is Changing
Lenny Rachitsky: What should people do to be successful in this new paradigm?
Eli Schwartz: Think of SEO as a product. The product managers are the people that should be thinking about this SEO question because it’s a product question. Product people need to think about how do we position this to the user that is not going to find out about this from a social channel, that’s not going to be attracted by an ad? This is a user that’s doing their own self-discovery journey. If you can’t answer the question about what is it that someone’s going to do a search on, then don’t do SEO.
Impact on SEO
Lenny Rachitsky: To a lot of people, SEO is kind of this dark art.
The Evolving Search Journey
Eli Schwartz: It is not a dark art. It is simple. I think step one is the step that almost everyone misses on SEO, which is.
Real-World SEO Impacts
Lenny Rachitsky: Today my guest is Eli Schwartz. Eli is a growth advisor specializing in SEO and has helped companies like Quora, Coinbase, Tinder, LinkedIn, WordPress, and Zapier develop and execute their SEO strategies. He’s also the author of Product-Led SEO and has a very refreshing take on how to think about SEO and win at SEO. Recently, he’s been spending a lot of his time analyzing how SEO changes with the rise of LLM chatbots, Google giving you the answers straight in the search results, and also how to utilize AI in your SEO strategy.
In this episode, we dive deep into everything that you need to know to be successful in this new AI paradigm. As Eli shares in the conversation, Google is just now rolling out changes to how search works and is greatly increasing how many searches include an AI-generated answer at the top of the search results. So things are going to start shifting under our feed pretty quickly. If you’re at all thinking about SEO, working on SEO, or are just curious about how search is evolving, this episode is for you.
If you enjoy this podcast, don’t forget to subscribe and follow it in your favorite podcasting app or YouTube. It’s the best way to avoid missing future episodes and helps the podcast tremendously. With that, I bring you Eli Schwartz. Eli, thank you so much for being here. Welcome to the podcast.
Impact by Website Type
Eli Schwartz: It’s a real honor to be here. You’ve had some amazing guests, and I’m honored to be counted as one of them.
Lenny Rachitsky: It’s completely my honor. You’ve been working on SEO for a long time. You’ve been helping companies figure out how to win at SEO for a long time, over a decade. And we were chatting about what’s happening in SEO, and you told me that you’ve noticed a significant shift in how SEO works with the rise of LLMs, with the rise of AI answers being integrated into search results.
And so I thought it’d be awesome just to spend an entire episode talking about what people need to know about what’s changing in SEO and how to be successful in this new paradigm of SEO with LLMs and AI being prevalent. How does that sound to you?
Structured vs. Unstructured Data
Eli Schwartz: That’s an awesome idea. I really like what’s happening with AI in general for SEO because it’s causing everyone that cares about SEO traffic, whether that’s a PM or that’s a CMO or that’s a CEO, to really be forced into pivoting their thinking about what SEO traffic means because the tactics around SEO haven’t really changed. It’s always been the exact same thing.
My last full-time job was out at SurveyMonkey, and I was moonlighting on the side. And I was introduced to a CEO of a big company and they were asking me about my approach to SEO. And I wanted to close it. I wanted to get this consulting engagement. And the CEO says to me, “So essentially what you’re telling me is I need to find my keywords, put that into content, and then build some links. Is there anything else you’re going to do for me? Why should I pay you?”
And it stunned me into silence because essentially that is and was SEO, and then that forced me to really pivot my thinking around what SEO might be. And I pivoted my thinking and I’ve worked with many companies around how they should think about SEO and what SEO traffic should mean, but others have not because those tactics did work. And LLMs and AI in general is forcing people to think again, how should SEO work? How should I be driving business from the search channel?
Top vs. Mid-Funnel Differences
Lenny Rachitsky:
First, Pendo is built around product analytics, seeing what your users are actually doing in your apps so that you can optimize their experience. Next, Pendo lets you deploy in-app guides that lead users through the actions that matter most. Then, Pendo integrates user feedback so that you can capture and analyze what people actually want. And the new thing in Pendo, session replays, a very cool way to visualize user sessions.
I’m not surprised at all that over 10,000 companies use it today. Visit pendo.io/lenny to create your free Pendo account today and start building better experiences across every corner of your product. P.S. You want to take your product-led know-how a step further? Check out Pendo’s lineup of free certification courses led by top product experts and designed to help you grow and advance in your career. Learn more and experience the power of the Pendo platform today at pendo.io/lenny.
It’s also a powerful traditional search engine with real innovations versus big tech options. It fights bias and SEO spam. It brings a cleaner results page with fewer ads, Reddit threads in the search engine results page, powerful local results, and even community- driven ranking options. Tired of big tech’s same old list of links? It’s time to try Brave Search. Visit brave.com/lenny to get started. That’s brave.com/lenny. So just to set a little foundation, talk about just what is it that’s changing in search and in SEO with the rise of AI answers and LLMs.
SEO as Product, Not Marketing
Eli Schwartz: So essentially Google and other tech companies had their hand forced by OpenAI and ChatGPT. Google claims to have invented the concept of LLMs and they may or may not have, and some of the early OpenAI employees were Google employees. But ChatGPT came on the scene at the end of 2022 with this ability to ask any question and then get a written out answer. And suddenly people are like, well, I don’t need a Google and click all these results.
So there started this conversation of you don’t need Google anymore. Google is ending. Even more than that, you don’t need SEO. You don’t need to optimize anything because the entire world will just be given to you. We’ll dig into that. I don’t think that’s at all correct, and I don’t think anyone, whether you’re doing SEO as your full-time job or whether you’re receiving SEO traffic as a part of one of your primary marketing channels, I don’t think anybody has to worry about that.
However, Google was worried about it, and I think one of Google’s primary stakeholders is really Wall Street. So if Wall Street suddenly thinks that Google is a has-been company and they’re not interesting anymore and they don’t want to maintain investments in them, their stock price goes down. And that hurts Google’s ability to recruit employees. It hurts Google’s ability to raise money and invest in all the interesting stuff they do.
So Google has to satisfy the curiosity and the interests of the general world and general investor by saying, “Oh, that OpenAI thing, we can do it too. That’s not that big of a deal.” So then Google responded, badly, of course, first, by launching what was then Bard. So they said, “ChatGPT, look, we’ve got our own version,” and they did a public demo. It didn’t work out well at all and their stock price actually went down.
And then they fixed it and then their stock price went up. But they also had to have an answer to this concept of, is search dying? Does anyone need to search anymore when the entire world can just be given to you? So they launched what at the time they called SGE, Search Generative Experience, which is essentially ChatGPT in a search result. So they launched that, but they launched it as a beta.
And there were huge issues with what they were doing there because there’s monetization issues. Everything they do comes from ads. The majority of their revenue comes from ads. So if you’re going to show this AI answer on a search result, then you also can’t have ads. You need to go all in on one of them. That was one issue which they actually have not solved yet. Another issue they had was liability.
So if they have a generative response that tells you to do something awful, like I think there was one where it may or may not have been fake, a lot of people made some fake ones, that told you to jump off the Golden Gate Bridge. Are there liability concerns? Because Google now is the publisher. They’re not a search engine that told you how to find out the answer to that.
They told you to do it themselves. And the third issue there, which is another liability concern, is it plagiarism? So those are the things that they worried about. So they took about a year to test this thing, which was fascinating because we’re tech people and most of the listeners are tech people. Google is very much a launch fast kind of company. They reveal something and then it sort of rolls out very, very quickly.
They’re not the kind of company that says we’re going to launch something and then they take a year to do it. So AI Overviews, what was then called Search Generative Experience, launched at Google I/O this past year, in May of this past year, and they renamed it to AI Overviews. And it’s that. It’s essentially ChatGPT in a search assault. They launched it to great fanfare.
Of course, it was live right away. They said it was only going to be logged in users only in the US. Obviously they always have concerns about launching things in Europe, because Europe’s a little bit more litigious than the US. And once they launched it, suddenly they started getting these screenshots of Google told me to jump off the Golden Gate Bridge, Google told me to put glue on my pizza, which, of course, Google didn’t do that.
They just crawled things from around the web. I’m surprised that Google didn’t predict that because the exact same thing happened with ChatGPT and Google is much bigger and a much more interesting target for people that want to share those things in social media. So Google launches it, and that was a little bit embarrassing, so they rolled it back somewhat.
But what’s even more interesting is they’ve given up on this rollback. Or not given up, they’ve gone back into it and they’ve relaunched it, and now it’s on so many more search results. I’m seeing it on most of the search results that I see. And on top of that, non-logged in users are seeing it, so in incognito users or users without any history with Google are seeing it, and they’ve just launched it in the UK. So this thing is coming and it’s really going to be affecting search results.
Lenny Rachitsky: To maybe help people understand why this may impact SEO, if it’s not obvious, is it pushes results down. People get the answer right there. They don’t have to click your links. And then there’s also the sponsored links at the top of the page that already are pushing you down. So basically your stuff is harder and harder to find, right?
Zapier SEO Case Study
Eli Schwartz: Actually, it’s not that it’s harder to find, your stuff becomes less relevant. And that’s the part about SEO that I’m excited about. So SEO was exactly like that CEO said to me years ago, it’s just about creating some content. And then it’s just this race to the top of getting your ranking results on that top keyword. So years ago, I worked at a startup where we were in the automotive space and we wrote content about cars.
And a word we liked to rank on was cars. The homepage was ranked for the word cars. We bought tons of links, and we were just a cars website. Obviously that’s not something anybody could expect to do now because you don’t search like that. You don’t search, “Oh, I need to buy a vehicle. I’m going to use the word cars and see what comes up,” because cars means a lot of things. Cars is a movie. Cars is a kind of car you drive. Cars could be a go-kart. It could be a lot of things.
So no one’s going to think about those search results. However, SEO is still geared towards those top of funnel, those big keywords that people cared the most about. So now a lot of those keywords are going to be moving into these AI Overviews. I don’t want to just focus on Google, although I think Google’s going to own the entire search page forever, at least for a very long time.
But other engines, whether it’s Perplexity or whether it’s Claude or whether it’s Meta, they all have this opportunity to give AI Overview type responses. That I think, those top of funnel kind of queries, are a great fit for what’s going to come out as an answer. So if you’re looking to go on vacation and you want a beach vacation, you can ask a very explicit question about give me a beach vacation that is not in America, but is a two-hour flight from X airport.
That’s the kind of thing that you could do in a Google search, but would take you a very long time to do. And those are great answers from just getting a paragraph. And then from there, you move into the mid-funnel. So now you do this query and Google suggests to you that you should go to Cancun. Or sorry, not Google, but whatever that answer is suggests to you that you should go to Cancun. And now you’re in the mid-funnel and that’s where SEO begins to matter.
So the reason why I think this disruption is so big is because the journey changes, the discovery changes. So whereas before if you were a travel site and you were able to rank on best beach vacation within two hours of the United States, that’s your ranking result. And then you have your long piece of content and you have your ads and you can monetize that.
That all changes when you can still rank number one on that, but you’re all the way at the bottom of the page. The AI answer, whoever that’s from, whether it’s again from ChatGPT or Google, tells you where to now start doing your deeper search.
Tinder SEO Case Study
Lenny Rachitsky: So just to maybe mirror back what you’re saying is the discovery step of search is going to be swallowed up by LLMs that give you direction. And then once you have a sense of what you want, then you go back to Google and that’s where potentially the opportunity continues to remain.
Eli Schwartz: Absolutely, and I think that’s where things are good. That’s where I’m excited by the user experience for search, because I don’t think the user was best served by, I don’t know, U.S. News or Forbes writing out where the best beach vacations are within two hours of the United States, with that piece of content that was written by a freelancer who’s not a travel expert.
So now you’re going to get that information also from not a travel expert. You’re going to get an AI summarized answer, and that will give you more clues and more ideas to do a deeper research search. I think that’s where the user’s best served.
User-Need-Driven SEO
Lenny Rachitsky: Okay, this is fascinating. Let’s definitely spend more time there so people understand exactly what that means. Before we get into that and what people should do to win here, what impact have you seen on SEO and search results and the space of SEO as these things have rolled out? Have you seen numbers of like, this is declining, this is growing?
Eli Schwartz: Transparently, I thought this was going to be an apocalypse. So I shared about a year ago that it’s going to be an apocalypse. We have not seen it yet, and a lot of people are declaring victory that there’s no apocalypse because we have not seen it yet. However, this thing just launched in a way that I think it will start impacting traffic.
So prior to the last couple weeks, it was not available on logged out search. It was only on logged in search. And even more than that, on that logged in search, Google had rolled it back significantly because of those embarrassing things that happened in May and June when they first launched it. They’re only just now launching it broader. And I do think we’re going to see impacts.
Of course, another challenge with pinpointing and impacts is they’re rolling algorithm updates that are happening, which we’ll also talk about, and that will mask what’s happening. Because if you’ve been hit by an algo update and suddenly you recover from the algo update, you’ll see more traffic, but you might have seen more traffic but on a lower base because these AI Overviews are changing things.
So I think if you take all those steps back and you look at this from a journey perspective, there’s no way that it won’t be impacting search. It’s just going to be hard to find. And there’ll be certain examples where it is extremely prevalent, and there’ll be others where it won’t impact things at all. And I think the riding line is the journey. We’re in the funnel that user is.
So if you are WebMD and you’re writing content about the human body, just generic content that has existed since medical journals were created, AI Overviews are going to be fantastic at giving you that answer. You have a headache and you’re just like, “Well, this headache’s not going away. Should I take more Advil or should I take a nap?”
And then you find that WebMD article and you get to page six and it tells you that 0.2% of people that have a headache actually have a brain tumor, and you don’t need that anymore because Google can tell you you’re good, you should take a nap and drink some more water.
WebMD was competing with Healthline and other, Cleveland Clinic and Dallas Clinic and all these other hospitals, and it’s totally unnecessary to go and look at all these results and go through six pages of information about headaches. So a site like that will be impacted by AI Overviews. An e-commerce site, maybe not. It depends where in the funnel that user is.
A Three-Step SEO Framework
Lenny Rachitsky: As you’re talking, I was reminded of something that I forgot about with my own experience. I used to have a website called whenishanukkahthisyear.com, because it changes every year and it’s just the date. That was the whole website. It gives you the date of Hanukkah this year. And I put ads on there and made 10 bucks a year, and then Google came in and just gave you freaking answers right inside the search result. So I’ve experienced this.
Eli Schwartz: So it’s even more than that. So what you are referencing is structured data. So that’s a very easy thing for Google to tell you when is Hanukkah in 150 years from now? It’s in a data set. What’s happening now is Google’s taking this unstructured data from content and building it into structured data. So you could ask a question of what is the likelihood of a baby needing to go to the hospital because they’re showing this sort of symptom.
And again, instead of reading all that content and making a decision, Google could take all that unstructured data, not just Google, again, ChatGPT or Claude can take all that unstructured data, and give you a statistic based on everything they’ve read. And that’s very helpful to users. And I think, again, users benefit.
And then a user might find out there’s another piece of information where I’d actually like to read a medical paper, or that now I’d like to Google and find the closest doctor to me who has certain hours. That’s a Google search. That’s not an LLM AI search.
When SEO Doesn’t Work
Lenny Rachitsky: And just to clarify, I didn’t intend to say that that was AI oriented. That happened a decade ago. So yeah, I totally, totally agree.
SEO Is Not Free
Eli Schwartz: Well, I just thought that was interesting to really drill into the difference between structured and unstructured, because unstructured is actually where Google is disrupting everything. So this entire idea of SEO, again, up until 2022 was monetize unstructured data. Whoever wrote the longest piece of content on best beach hotels in Miami and then built the best links to it, they would win.
They win whether it’s ads or win whether it’s hotel bookings. Now, Google can start you off at the top of that and say, “These are the best beach hotels based on all the people that have written content,” or actually Google’s own unstructured data from the reviews and give you that. And then you can say, “Okay, I’d like to go to this hotel, or I’d like to stay in this city.”
The Four Growth Engines
Lenny Rachitsky: So help us understand even deeper this distinction between top of funnel and mid-funnel. So when people are maybe winning at one or the other, what does that look like? What are some examples of like, here’s a top of funnel type of search that Google’s going to eat and here’s a mid-funnel experience that you can win that?
Companies That Don’t Need SEO
Eli Schwartz: So in general, SEO has always been more at the top of the funnel generally, because you’re curious about something. So let’s say you’re looking for a new software. You’re looking for new podcasting software. So you search for top podcast tools, and you get back a list on let’s say G2. G2 and all the sites like G2, like the Gartner sites like Capterra, all going to be massively disrupted. So you would get back a list from G2, which would give you out all the software.
And you would look at some and it would say, “This one’s geared towards enterprise. This one’s geared towards small podcasters. This one’s free,” and then now you’ve narrowed down your list to these three tools. And that’s when you start doing those searches. Now you’re mid-funnel. Now let’s say you’ve chosen Riverside and you’ve gotten enough information from those other searches. Now you start searching Riverside price, Riverside capacity, Riverside bandwidth.
That’s bottom of funnel, and that’s where you’ll now go buy Riverside. However, at the top, you were doing top podcast tools. So again, up until AI Overviews, up until this entire concept of LLM, whoever won on that long form piece of content would get that first click. But now that doesn’t exist anymore. You just go to Google and Google tells you, “These are the tools. Oh, you’re looking for a new CRM? This is what a CRM is. You want to know, do you need a CRM or do you just need a calendar?”
Look at that and Google will just tell you in a paragraph, and now you’ve redirected your search somewhere else into middle of the funnel. So I think that’s where SEO always should have been because that’s where conversions could potentially happen. However, SEO never was there before because the way most SEO measures its performance and its success is rankings. So they would say, “Well, it doesn’t really matter if we convert on the word CRM. But look, we’re number one, so we’re winning.”
Lenny Rachitsky: First of all, as a user, this sounds great. I’m so tired of just all search results just being a bunch of SEO pages, just with a bunch of BS answers. And so I really prefer, Google, just tell me. Just tell me what I need to know. The other piece is if you really think about what Google has been trying to do, they’ve been trying to do this is like, here’s our best shot at giving you an answer to this question, and here’s links that’ll point you to an answer.
And this is just a better version of it where it takes all the actual information and just gives you the end result. So it makes a ton of sense that they’re doing this and the technology has finally allowed them to do this. Okay, so let’s get to the, I don’t know, $64,000 question, million dollar question, what should people do? What should people do to be successful in this new paradigm that as you’re describing is in motion?
It’s just starting to now happen. And it may be catching people off guard because they’ve thought it’s already been out and things are okay, and you’re saying it’s actually starting to actually move quicker.
When SEO Shows Results
Eli Schwartz: I think it’s fascinating, I’m honored to be one of the few marketers on your product hard test, but I think of SEO as a product. And I think the product managers are the people that should be thinking about this SEO question because it’s a product question. There are users that are coming in from the search channel. What is the product that I need to create for them?
What is the experience that I need to create for them? So typically, it was thought of as a marketing challenge of the product people have created this thing for me, and now I expect the marketers to go and do the SEO thing for it. But there’s a mismatch. So for example, I find very often when I talk to SaaS companies, I don’t think SaaS in general should do SEO.
But very often when I talk to SaaS companies, they have created a product for whatever user, and then the marketers are expected to make that product fit into the thing that the search results are around. And it doesn’t work because that’s not what they’re looking for. No one asked the question of what it is that they’re looking for. So now that SEO is changing and you really need to think about this mid-funnel and you need to think about a user experience and a buyer experience when they’re doing the search, I think it all comes together.
And this is where the product people now need to think about how do we position this to the user that is not going to find out about this from a social channel, the user that’s not going to be attracted by an ad, the user that’s not going to discover this tool from a trade show. This is a user that’s doing their own self-discovery journey, and this is what is the thing they’re looking for and how do we position this product in a way that they’re going to find it.
So that’s what everyone should be doing, product really collaborating with marketing and discovering what is it that the user wants and showcasing that. I was recently talking to a company in the health space. They have an app. It’s a health app. And we talked about their marketing. Their SEO in general was they write thousands and thousands of blog posts. AI has allowed them to do things they should never have done.
They wrote thousands of blog posts about health in general. I told them they shouldn’t do that in general, but they wanted me to experience their product. So they gave me a code to download their app, and their app is awful. It’s a bad product. So they’re trying to do marketing that doesn’t fit for a product that doesn’t do the thing they say it’s going to do. If they fix the product and understand the user, now it becomes, well, if I’m a user looking for this product, what are they looking for? How do we showcase that, again, in the mid-funnel?
Lenny Rachitsky: Can you help us make this even more real? Maybe go through an example of a product you worked on or one that’s doing this really well in terms of going from, hey, they want our product. What’s the journey look like? What’s a good example of that?
Conversion Metric Expectations
Eli Schwartz: So the earliest example of where I discovered on this process was about 10 years ago, I met the CEO of Zapier, Wade Foster, and he was a co-worker of mine. I had invested in this company, and he asked me to meet with him to just discuss SEO. They were doing something that no one needed. They had this product which zapped things together, but no one needed it because they didn’t know it existed. But people knew that they needed things to work together.
They just weren’t looking for Zapier. So in my discussions with them, in our SEO experiments that we came up with, we said, well, people are looking for let’s say Gmail and they’re looking for Salesforce. They know that Gmail doesn’t connect to Salesforce, they know Salesforce doesn’t connect to Gmail, but they’re looking for ways to pair it together.
So what if we created as a product a marketing way to showcase that this product of zapping Gmail to Salesforce and Salesforce to Gmail exists? And that’s what we did. So we built that at scale for everything that could work, and that was the experiment that I did with them, where everything would be available and it creates this flywheel of, wow, if Gmail can work with Salesforce, what else does Gmail work with?
Can it work with this other tool I have? And that’s a, again, early example where I stumbled upon this idea of people are looking for that other thing that you could do, showcase the fact that your product could do this thing.
Metrics to Measure SEO Success
Lenny Rachitsky: I had no idea that you were involved in Zapier’s SEO work. That’s one of the most legendary successes of SEO. Very cool.
Eli Schwartz: Total accidental. Whenever people ask me about programmatic SEO, and I know we have to dig into that, what programmatic SEO is, they’re like, I want to imitate Zapier. It’s always great to mention that I helped through that.
Time to See SEO Results
Lenny Rachitsky: That’s amazing. Okay, so the lesson there is you realized that the key, the opportunity is to teach people what they could accomplish with this product. There’s this awesome product doing amazing things. They have a need they’re searching for, say it’s like Gmail and Salesforce, and how do we help them see there’s something really interesting here.
Embedding SEO in Buyer Journey
Eli Schwartz: Yeah. I actually have a better one for you.
AI in SEO Content Creation
Lenny Rachitsky: Great. Let’s do it.
When to Use AI Content
Eli Schwartz: So I worked with Tinder for a couple years. One of the biggest challenges I ever had with SEO, especially as a consultant, is getting things done. And I worked with amazing companies and there’s great people, but then they run into this wall. In my book, I talk about one of the earlier stories I had of this where I worked at a company where I was hired by the CMO and we built out a great plan.
And then we go to the CEO who brings in the CTO to discuss our plan, and the CTO says, “I don’t have engineers for this. So you decide whether you want to work on the product or you want to work on this marketing thing you’ve done.” And then we didn’t do anything because they never resourced it. But I worked with this really great product person at Tinder, Udi Milo. He was the chief of growth.
At Tinder, Udi Milo, he was the chief of growth, and he really saw a problem where Tinder had never done SEO. All of Tinder’s inbound came from the word Tinder, and he is like, “There has to be upside for SEO if they’ve never done it before.” So that… he was convinced of that. He was willing to drive forward on this idea. So when we started working together, the first thing we came up with is what is someone going to be searching when they look for Tinder? And it’s not online dating. That’s a single word.
We’re not going to write out long-form content around everything related to dating because that’s not the Tinder product. You’re not reading a piece of content about how to fall in love and then somehow converting into Tinder. But in our user research and in the discussions we had on what the investment we can make into SEO was we discovered that Tinder is a loneliness-solving problem, loneliness-solving solution. So you’re lonely, you’ve gone to a new city, you don’t know anybody you’d like to solve your loneliness problem. So because you’re in a new city, it occurred to us that this is a local thing. So we’re going to look for anything related to local. So what we built out is if you look for online dating in many cities around the world… This is beta. It remained in its beta state, but if you look for online dating in many cities around the world, you’re going to find a Tinder page, which gives some examples of places you can go on a date. And more than that, it gives you Tinder as a solution to solve the loneliness problem you have.
The AI Content Debate
Lenny Rachitsky: So what changes here with the rise of AiO reviews and things like that? Is it this is where… this is how future of SEO looks versus just keywords and endless block posts.
Impact of AI Content on Google
Eli Schwartz: Actually nothing changes here with AiO reviews because if you’re looking for, you’ve gone to Dubai. You’ve just… You’ve gone to Dubai. It’s a brand new country, you’ve never been there before, and you’re lonely. So you look for online dating in Dubai, and you’re going to get… Again, AiO reviews might tell you what the dating scene is like. It might tell you where to go on great dates.
However, it doesn’t allow you to solve the loneliness problem you have. The loneliness problem you have is a mid or bottom-of-funnel problem. You’re still going to click on Tinder’s result, no matter what the AiO review does. So that’s where I think SEO should be is you need to be in the buyer journey with the SEO you’re creating. And it doesn’t matter what AiO reviews does or doesn’t do because you’re still solving the mid-funnel problem with your SEO solution.
Lenny Rachitsky: Let’s go in a direction I was going to say for later, but it might be useful now, which is, say, somebody’s just sitting there at their desk thinking, “Hey, I want to be… I want to start doing some SEO. I want to be successful at SEO. Haven’t really done a ton here.” With this new world of AI and LMs, what would be step one, step two, step three to move down a direction of starting to poke their toe in the water of SEO?
AI Overviews and Brand Building
Eli Schwartz: I think step one is the step that almost everyone misses on SEO, which is be the user. Try to understand who your user is. And there’ve been so many companies, and SurveyMonkey was a great example. One, I spent seven years at SurveyMonkey, and when everyone got onboarded at SurveyMonkey, they gave you a SurveyMonkey account and told you that you should run a survey. I think later, in the later years, they forced everyone in their onboarding time to use a survey… run a survey with people in the company, but before that, no one did. So they had this account, and they never did a survey. So they had zero customer empathy for why people use the tool.
And then, when you think about it from a product and marketing standpoint, you’re thinking about it as a work challenge rather than a customer empathy challenge. So the first thing anybody should really do around anything that they’re trying to do SEO is try to be that customer. So if I am a user of this SaaS, I’m building SaaS, and I want to market this tool, what is a user going to look for? What’s the problem they’re going to look for that would make them want to do a search? And earlier, I referenced that I don’t think most SaaS tools should do SEO. And the reason is because a lot of times, when I talk to SaaS companies about SEO, I ask them this question, and they give me a blank stare.
So if you can’t answer the question about what is it that someone’s going to do a search on, then don’t do SEO because SEO is about appealing to that user. If you can understand what the user should do around looking for whatever product you have, then that’s the first step for SEO. So who is this user? Who am I marketing to? You’re sort of creating this persona in your mind. Step two is think about the asset you’re going to create. So pivoting over to what he talks about Tinder. So we understood that it was a user that was solving a loneliness problem any in the world because that’s what Tinder does. It solves that loneliness problem in any way you want, of course.
And now we had to think about what is it that we’re going to create. We know it should be global, of course. We would know we want it to be programmatic, which is something we should dig into like difference programmatic versus editorial. We know we want to be programmatic because no one wants to in position of writing a page out for every town or every city or every neighborhood in the entire world. So we want it to be programmatic. What things do we need to pull into that? And then the third step is really building, and we’re on a product podcast. Of course, to the product people, you’re building the product for that SEO user. So where do you get those inputs? What does this page need to look like?
So again, the reason I think SEO needs to be on product is because the inputs for SEO aren’t the way many people think of SEO, which is a piece of content optimized for Google based on the keywords I’ve chosen. I think of it as a product, which means you need design resources. You need engineering resources. Of course, you need a product manager to really oversee the building of this. You need user research. So it’s, again, more than just a piece of content. So going through those steps, again, it’s understand who your user is, decide what it is that you need to create for that user. And the third is envision this product.
Lenny Rachitsky: So I think there’s a really powerful point here that might be… people might not get, which is the user needs to be thinking of something they will go to Google to search for to find your product.
So it’s if they’re… if nobody is searching for the thing that you’re building in some way, there’s not going to be an opportunity for you to win an SEO or and benefit from SEO. Are there any examples that come to mind to you of B2B, SaaS companies that just like, “This is not… There’s no SEO opportunity here, just like no one’s searching for this thing.”
Programmatic vs. Editorial SEO
Eli Schwartz: Most of them. [inaudible 00:35:30]. One of my first consulting clients, right, when I left SurveyMonkey was Mixpanel. So, I mean, I just left my job, and I’m going to take any client that I have. I have some awful clients that I took, but Mixpanel was not an awful client. It was just an eye-opening client where we’re trying to do SEO, and I’m doing SEO and exactly the way that CEO had told me SEO should work, and he was going to do it himself. So I told them the keywords. We came up with the content. We built the links. We did all the stuff SEO was supposed to do, and it didn’t work. And we were sitting there, and I asked them to show me the user journey. They have a tool that does it. That’s what Mixpanel does.
So someone clicks on a search result and they land on this piece of content, which we did research and we know what people look for, but why is it that it doesn’t convert? And then I realized that there were other problems to this conversion, which is Mixpanel is a product you integrate in your entire company. You don’t just do a quick Google for it and be like, “Oh, analytics. Yeah. I’m going to tell everyone we got to do it and it’s going to be live tomorrow.” Also, Mixpanel is expensive. So again, it’s the kind of thing that there’s friction of. You don’t just click from a search result and then decide to just purchase it. You can’t even purchase it on your credit card. I don’t believe.
So those are the issues, and that’s why I think SaaS is not the best fit for SEO because if you think about that journey, the problem doesn’t necessarily exist. The SaaS solves the problem, but only once you know that problem exists. And getting people to know that problem exists is typically not an SEO challenge. It’s typically a brand challenge. You make a viral video of you didn’t realize that you could use this tool we’ve created to solve your problem, but no one is necessarily searching for a problem they didn’t know they had or a solution they didn’t know could exist. So that’s where the breakoff is. And then the other issue, of course, is now they know the problem exists, but it’s not an SEO journey.
So I had another eye-opening experience early in COVID when I was… everyone was home, and there wasn’t a lot of things to do. Google reached out to me about a role on their team on Google Cloud for doing SEO. And there wasn’t a chance I was going to take it, and it seemed like an interesting thing to go through a hiring loop and stay at home. So I had a fascinating experience interviewing because every time the… first it was the hiring manager and every time it was someone on that team that started the interview. They said, “Are there any questions that you have for me?” And then I spent the next 45 minutes asking them all the questions like, “Why are you doing SEO? What kind of keywords would you want to do? What does this SEO journey look like?” Because for Google Cloud, they only have two competitors, Amazon and Microsoft, and no one is going to do a search and be like, “Oh, Google’s number one for this term I searched. Let me just go buy it.” There’s a decision-making process. You bring… There’s a committee decision. So SEO sort of made no sense for that even to be included as a part of a marketing channel. And at a grand scale, Google Cloud should not be doing SEO. At a small scale, a lot of SaaS tools shouldn’t be doing SEO because there isn’t necessarily an SEO journey. I get a lot of pushback from companies when I tell them. They’re like, “Look at all the SEO we’ve done.” But that plateaus really quickly.
Lenny Rachitsky: This is really fascinating, and just to maybe reframe what you’re saying because there’s kind of a couple elements of this. It’s not necessarily that people aren’t searching Google for the problem. So I’m thinking of Vanta SOC 2 stuff. It’s not like people aren’t SOC 2 certification. It’s not like they’re searching… they’re not searching.
And what I’m hearing is the bigger issue is they will never buy your product from that experience and that journey. It may educate them a little bit and may teach them, “Oh, Vanta exists,” but they’re never going to become customers. They need to talk to salesperson. They need to involve bunch of stakeholders. So it’s basically a sales motion. It’s not a product-led SEO motion.
Keys to Programmatic SEO Success
Eli Schwartz: Yes.
Antitrust Ruling SEO Insights
Lenny Rachitsky: Is that right?
Eli Schwartz: Yes.
SEO Is Not Dead
Lenny Rachitsky: Cool. Pulling on that thread a little bit further beyond maybe even B2B, just how does one decide if SEO is an opportunity for you, and also just how much should you invest in this opportunity just to explore it?
Eli Schwartz: So we have to really put away the myth that SEO is free because it’s absolutely not free. There’s a cost in time. There’s a cost and resources, and, of course, there’s the direct expense for SEO. So if you’re deciding that you should do SEO, so now we have to go through this evaluation. Again, this is where I think of it as a product where you have all these product ideas and what should you spend time and money on. So SEO is a channel you may or may not want to invest in.
So say there’s a SaaS tool, and they’re convinced there’s an SEO journey, people do search for this solution. When I was at SurveyMonkey, we generated a couple of hundred million dollars a year off of organic traffic because it was a freemium tool where you search for the prodem… the problem, you find the solution, it’s free, you sign up for it. If it works for you, you end up paying, and that’s the revenue we’ve generated from organic. So say there is a journey, so people do search for it, and it makes sense to invest in SEO now is when you’ll decide how much should you invest in it and how you should invest in it.
So typically, [inaudible 00:40:32] again, for SurveyMonkey, it was around creating content. It was around templates. It wasn’t very expensive. But let’s say there’s a company that does not have a content team. They don’t have a product yet that people are going to search for. They don’t have a product manager that’s going to be overseeing this SEO process. So say you need to hire this PM, or you want to hire an agency, a typical SEO agency, I mean, you’re not going to spend 10,000 just because that’s the cost of someone’s time.
So it’s 120,000 a year, gets even more expensive if you have a full-time employee. So that’s one expense you’re going to outline. Then, you’re going to add in all the supporting resources. You need a CMS. That costs money. You need an engineer to support them. That costs money. You need a potentially design. You need content. So it can really add up quickly. And now you look at that investment, and you say, “For this tool, if I invested a million dollars a year in SEO, do I expect to make back a million dollars a year soon, right?
SEO will always make back money if it’s the right fit, but soon. So SaaS tools, especially startups, they need to make that money back soon. Or if they took that exact same million dollars and put it into brand ads, or they put it into influencer campaigns, or they put it into just traditional paid marketing on Meta and Google, would you make that million dollars or $1,000,001 back faster?
And that’s where I think the evaluation should happen instead of this default, “Well, I just got my funding. I need to invest in SEO because it’s free, and everyone does SEO. And look, my competitors do SEO.” It should really be this thoughtful strategic decision-making process of how much will this cost me all in and is this the right use of funds.
Common SEO Myths
Lenny Rachitsky: Amazing. So I think this is really important. It’s not that you won’t benefit from SEO. It’s not that SEO isn’t an opportunity. It’s that you have much bigger opportunities in other areas. Most likely, if you’re a B2B SaaS companies, because the journey isn’t fully online, you’re not going to convert by just reading a bunch of pages.
Getting Started with SEO
Eli Schwartz: Absolutely.
Will TikTok Replace Search?
Lenny Rachitsky: Super fascinating. Yeah.
Eli Schwartz: Yeah. There was once a company I met, there was a SaaS tool, they were in a gardening space. [inaudible 00:42:39] they made a SaaS for gardeners, and they were insistent on doing SEO. And they asked me to look at a proposal they got from an agency for $15,000 a month, and it was all content. And then I asked them how their users found them. How did they get all the customers they had that paid them? How did they find them?
They said they go to these gardening shows around the country, and they have a booth, and each booth costs them 15,000 on SEO, you can get… you can go to all these shows for the exact same budget, and you get users who are interested, they’re in market, they try your tool out at the booth, and then they leave and they’re leads that you could follow up with. Instead of investing in me like, hope it works. It’s sort of free. There are searches for it, but it’s not the right searches.”
Technical SEO Myths
Lenny Rachitsky: The takeaway here is if you’re thinking about whether SEO is worth an investment for you and you see all these other companies doing SEO winning with SEO, think about how much will this actually cost us. And I think the most important takeaway for me here is if you don’t think it’ll convert online if you think sales is the core motion of the process, it’s probably not a good ROI for you.
Page Speed Is Overrated
Eli Schwartz: Absolutely. Yeah. Really think about what are the trade-offs. Again, from a product standpoint, there’s always trade-offs. So what are the trade-offs to investing in this channel versus another channel?
Lenny Rachitsky: And I think what I’ve seen, there’s all the… the way I think about it, there’s four actual growth engines… there are four core growth engines. SEO, paid, virality, and sales. And what I find is, eventually, large companies do them all. Unless you’re a consumer, you don’t do sale. And so it’s not like SEO should you never do. It’s I think the main point here is earlier stage, when you have limited resources, probably not the best use of your time.
The Top-Down SEO Approach
Eli Schwartz: I think there are companies that should probably never do SEO. [inaudible 00:44:24] going back to what I said when Google Cloud, I don’t think that Google could ever say… really pinpoint that there was a customer that they got purely from SEO, and you’re going to do all sorts of weighted attribution. And maybe that person did discover Google Cloud from a piece of content or piece of SEO asset, but they never would’ve converted with all those other things that they’ve touched.
Lenny Rachitsky: And it’s not like you’re saying, “Don’t be on the internet and don’t will make it easy for people to understand who you are, don’t have a great site, don’t have other pages writing about you. It’s just don’t… You don’t need to spend time optimizing the search results for Google Cloud. You’re not going to benefit significantly from that work.
The Value of Keyword Tools
Eli Schwartz: Yes, I mean, again, go really digging into this user journey piece. I don’t think restaurants typically should have a website. Not only should they not do SEO, but I don’t know that they should have a website.
Lenny Rachitsky: Hot take.
External Tools vs. GSC Data
Eli Schwartz: Because if you think about the buyer journey for a person looking for food, they don’t typically go to a website. So if you’re looking for pizza, and again, a lot of pizza shops have websites. But if you’re looking for pizza, you’re going to Google Maps, DoorDash, Uber Eats. You’re not going on Google and saying, “Pizza near me,” and then browsing the websites and just making a decision about where to eat lunch.
Again, if you’re choosing how to cater your kid’s birthday party, that might be a different thing, but you’re not browsing and saying, “These are the places I’m going to go right now.” But more than that, it’s expensive, and there’s no way that that pizza shop knows that the website they have with the soft music playing and their menu that comes up in flash, which is expensive, does a single thing for them. Most of their orders are going to come again through those other platforms.
Lenny Rachitsky: Yeah, I hate restaurant websites. One thing I read once that explains why what restaurant websites are so bad is restaurant owners are big on just like what is the experience when someone enters my restaurant that they go through the vibe, and they do that on the website. They’re like, here’s the music, and here’s the imagery, and here’s the animations.
And nobody wants that in [inaudible 00:46:19] on their website. Just like give me the hours and location and your menu. You mentioned a couple of things that I want to drill into a little bit. So, one is just how long SEO should take for you to see results. Two is just expectations of what does good conversion look like. What does it tells you this is actually us, a journey that could work for us well enough that SEO might work? So maybe those two questions.
The Future of SEO Consulting
Eli Schwartz: So those will be custom for anything. And I hate to use the word, it depends because I think whenever consultants say it depends, they’re just… they’re throwing their hands up and saying, “I don’t really know. I’m not… I don’t have an opinion on this.” And it’s like when you go to a doctor, and the doctor’s like, “Well, it depends.” I mean, you could be dying, or you could just… need to take a nap, right. So there’s a custom answer here, and it comes down to what the company is and what the expectations are. So how long it takes, that depends on what you’re building. So if you’re… if there’s an… Again, with Tinder, it took us a while to actually build anything.
So it took us all this time to build for ideate them, build and then to see results. But once we built, and I love working with big companies that are well-known brands, because what we build, as soon as it’s available on the web, it’s like this turning a huge ship. It starts driving revenue and starts being super effective because it’s there. Something didn’t exist, and now it exists, and it drives revenue. Smaller companies, it may take many months before Google notices. It may take many months before the demand is there. With Zapier, I think it took them a couple of years before they even saw any results because nobody’s looking for it. So it comes down to what it is that you’re building and how quickly users will come and find and need that solution.
Then, as far as what the expectations are on conversions, that also is really depends on what it is that you’re looking for. There’s… Some companies are building media, and they [inaudible 00:48:09]… I mean, again, I think most companies should just not write a lot of content unless they’re a media company. But if you are building media and you’re monetizing that content from a media standpoint, so maybe it leads, maybe clicks off the page into leads, or maybe it’s CPM advertising, so then your conversion is… what you’re trying to do with SEO is get a lot of page views because the more page views you get, the more clicks off that page you get into something else. If you’re a SaaS tool, then your conversion should absolutely be whatever a MQL should be.
So you… And again, I think most people don’t do this correctly with SEO. They used the wrong conversion metric, which is top-of-funnel ranking. “Oh, I’m ranking. My SEO is successful. I’m number one for this.” Instead of, “How does this benefit me?” There was a company I was working with, a SaaS tool. They’re in a two-sided marketplace they worked with… in an HR space, so they only monetized one side of that HR space, but all the traffic was on the other side of the HR space. So when we were talking about their SEO conversion problems, I suggested that they delete all the content that was on the wrong side of that marketplace because they didn’t convert at all.
And there’s no… And this is a common misconception. People will think you get a benefit from traffic. Google sees, “Oh, I get all this traffic from search. They think I’m a very good website. So even though none of this converts for me, I should have a blog because blogs are good.” It doesn’t really work like that. I mean, maybe if you’re this massive website and I don’t know, let’s say, get millions and millions of visits, then you can build some sort of authority and some good experience in Google, and now you can launch something else, and you don’t have to worry about needing to build up that authority. But generally, driving traffic to something that doesn’t convert for you wouldn’t be a good idea.
So really understanding what is it that you’re trying to do with SEO traffic, that’s your conversion metric. So it could be MQLs, could be page use, could be dollar conversions, it could be people picking up the phone or watching the video, but it has to be some sort of conversion. So for a startup, it might mean that they… you’re getting links. I mean, at a minimum, people reading your content deciding to link to you and give you social shares. That might be something you put in a pitch deck. Whatever it is, there has to be a conversion metric that matters for the business.
Quick Lightning Round
Lenny Rachitsky: So just to give someone that’s starting to do this something concrete to look at. When you come to a startup and help them with SEO, what do you usually track as a sign this is working that we should keep investing? What’s… What are the couple metrics that you look at most?
Eli Schwartz: The first is really understanding what is it that they care about. And rankings could be something they care about if someone else cares about it. So, for example, if they’re going to put in their pitch deck, “We’re number one for this tool,” and investors, they don’t know that that doesn’t convert. So that might be something that we should care about. They should be number one for that tool. If it’s MQLs, so they just need people filling out lead forms.
So really understanding who it is that they’re trying to attract with SEO that’s the metric we’re going to use. And again, I don’t know why SEO… this happens with SEO but not any other channel. No one doing paid marketing will ever say, “This is how many times I’m number one on this search, or this is how many clicks I get, or look at how much I spent.” There’s really a, “We spend this, and we’re this efficient. This is what we drive from that.” And SEO really needs that same rigor.
My Favorite Product
Lenny Rachitsky: So what I’m hearing is look at if you’re in B2B leads coming in through SEO, track that, or rank potentially if that’s like a vanity thing, you can show investors and you’re saying traffic alone, because usually in decks I see and investor updates almost always the metric I see is traffic we’re getting through SEO.
Eli Schwartz: Yeah, total waste of time.
Personal Life Motto
Lenny Rachitsky: Amazing.
Proudest SEO Achievement
Eli Schwartz: Yeah, I mean, unless there’s a reason for it. So if you’re a media company and traffic means that that number keeps going up, then yes, it matters. But if you are driving leads and that traffic number goes up, you’re not driving more leads from SEO. You’re getting worthless traffic. It doesn’t help the business.
Final Closing Remarks
Lenny Rachitsky: Awesome. Okay. Back to the question of just how long it takes just to give someone something here that they can tell… that can tell them when it makes sense to keep going or move on. Say you’re just taking a shot at SEO, how long should you give it to make a decision this is working for us or not? Or is it always, “There’s always something here, keep working.” How do you decide keep doing or not?
Eli Schwartz: When you’re building out an SEO effort, and I’m going to keep going back to this, it’s a product. So you’re building out a product and you’re creating milestones. So you say we’re… it’s going to take us this first month to ideate on what it is that we’re building. It’s going to… In the second month, we’re going to build out a PRD for engineers to start working on. In the third month, they’re going to start working, and they’re going to ship this.
Now, if you start missing all those milestones, so I mean, this is a common problem in consulting in general, you miss all those milestones, and then some will say, “What’s… We’ve been working together for six months, and we have nothing to show for it.” And you can point very specifically to all those milestones that were missed. So that’s the result of not shipping or not meeting the milestones that the SEO didn’t work.
So as you’re meeting those milestones, you can say, “Well, we launched, and our expectation was that after the first month of launching, we would have X amount of pages indexed. Did we meet, or did we miss that milestone?” And then you can say, “SEO is working at a small scale.” And I’ve seen in many companies it takes a really long time, but eventually, when you look back, you see this hockey stick, but while you’re in it, you don’t necessarily see that.
Lenny Rachitsky: Awesome. Okay.
Eli Schwartz: Okay. Let me share an example of one that worked really quickly. So this was also an interesting one, and I was at SurveyMonkey, and I was introduced to the CEO of Quora, who was interested in doing more SEO. So Quora had never done SEO or never done SEO effectively from a product standpoint. And there were a couple of things I was able to recommend to them, and within three months they had quadruple traffic. So it really will depend on what is it that’s holding things back. In Quora’s case, they weren’t showing any answers. So they were really pushing to get people to log in before they saw any answers.
So I encourage them to show the answers. Yes, they were going to be giving way answers for free, but they were also going to be giving answers to Google. So that was number one. That they were able to quadruple traffic. And the second thing is they had no way that Google could navigate within the site. So when you came to Quora, and it’s like this again, so they reversed the suggestion I made, but I think it was 12 years ago. So when you came to Quora, and when you come to Quora today, you see related questions. So you see a question and then there’s related questions. And that’s the way a bot or a human will navigate through the site.
If they create a sitemap, they did this in the past, and if Quora’s listening, you should do this again. If you create a categorized sitemap where you can say, “These are all the questions on health and from the sitemap, again, it’s HTML sitemap, not just an XML sitemap and way even a user can navigate through it. This is health, and this is health page one and health page two, and you can navigate through this entire site. Then a search engine can navigate through the entire site, and all of the questions and answers are discoverable.” So when I made that suggestion to them, within a couple of months, they were able to quadruple traffic.
Lenny Rachitsky: There’s so much here, and there’s so many valuable insights. One that I think is really recurring that I think is really important to people is the point that you’re making about your SEO content pages should solve… should be a product, AKA should solve problems for people as they’re trying to understand the space and the potential problem they have.
So correct me if I’m wrong, but kind of the advice you’re sharing is there’s less opportunity in just generating tons of blog posts that just have a bunch of content. And the opportunity is more things like Zapier and Canva with templates and Notion and Quora where it’s just like, “Here’s the answer.” It’s just like actually help them solve the problem. And with that, “Here’s how our product [inaudible 00:56:03] help you solve that problem further.” Is that…
Eli Schwartz: It’s really building a product around what the company is attempting to monetize. So Canva is building templates because then they’re going on to monetize that templates by having other people build off those templates with upgrades and subscriptions. But if a company was… Let’s say there was an enterprise version of Canva, and they didn’t monetize those templates, so putting out a bunch of free content on the internet that just looked good wouldn’t benefit. So it always comes down to what is it. What is the product that you want to use? So going to mention the templates did at SurveyMonkey, so the survey product.
If someone looked for a template of a survey, they wanted to make a survey. So we were not a survey product, and we just monetized off the survey templates that would not have done anything for us. So copying someone else’s version of programmatic doesn’t do anything. And generally, doing programmatic for the sake of programmatic, so just to have content, wouldn’t do anything unless that programmatic fed you into exactly what your product did. So Zapier, they did programmatic. It feeds you into making other zaps between different products. Tinder, we did programmatic.
It fed you into, “Oh, well, you would like to solve your lonely… We showed you that your loneliness problem is solvable in the city you’re in. You have to solve that problem by downloading the Tinder app.” And I’ve never used Tinder, only used it from a marketing standpoint, but [inaudible 00:57:26] Tinder gates you. So if you’d like to get the advanced Tinder experience, you have to pay. And again, they’re… it’s all part of that buyer journey, and anything you’re building SEO for, it has to be a piece of that buyer journey. So programmatic, blog posts, anything you’re doing, if there is no product journey for it, there’s no user journey, it just stops.
Lenny Rachitsky: That’s an amazing clarification. So again, it’s another point you’ve been making of traffic alone is not valuable if it doesn’t convert. And so you can look at a Canva, you can look at Notion at Airtable with all these templates, copy that. But if that isn’t the thing people, will buy from you and monetize it’s not going to be worth doing.
… Will buy from you and monetize. It’s not going to be worth doing.
Eli Schwartz: Yes.
Lenny Rachitsky: Awesome. I’m excited to chat with Christina Gilbert, the founder of OneSchema, one of our longtime podcast sponsors. Hi, Christina.
Christina Gilbert: Yes. Thank you for having me on, Lenny.
Lenny Rachitsky: What is the latest with OneSchema? I know you now work with some of my favorite companies, like Ramp, Vanta Scale, and Watershed. I heard that you just launched a new product to help product teams import CSVs from especially tricky systems, like ERPs.
Christina Gilbert: Yes, so we just launched OneSchema File Feeds, which allows you to build an integration with any system in 15 minutes as long as you can export a CSV to an SFTP folder. We see our customers all the time getting stuck with hacks and workarounds, and the product teams that we work with don’t have to turn down prospects because their systems are too hard to integrate with. We allow our customers to offer thousands of integrations without involving their engineering team at all.
Lenny Rachitsky: I can tell you that if my team had to build integrations like this, how nice would it be to be able to take this off my roadmap, and instead use something like OneSchema, and not just to build it, but also to maintain it forever?
Christina Gilbert: Absolutely, Lenny. We’ve heard so many horror stories of multi-day outages from even just a handful of ad records. We are laser-focused on integration reliability to help teams end all of those distractions that come up with integrations. We have a built-in validation layer that stops any bad data from entering your system, and OneSchema will notify your team immediately of any data that looks incorrect.
Lenny Rachitsky: I know that importing incorrect data can cause all kinds of pain for your customers and quickly lose their trust. Christina, thank you for joining us and if you want to learn more, head on over to OneSchema.co, that’s OneSchema.co. Okay, so coming back to AI and its impact on SEO, can you use AI to help you with this, and use AI to create content for you?
Eli Schwartz: AI is a tool. Everyone says AI is a tool, it’s not a solution. You can use AI to create content if the content you’re creating is a part of that journey. For example, before AI content really came on the scene, because AI’s been around for a while, Jasper’s been around for a while, Writer’s been around for a while before ChatGPT. Before these tools, a lot of ways that companies created content for cheap is they went on Fiverr, and they went on Upwork, and they just created content.
A lot of that content, completely worthless. If you’re just creating content for the sake of content, and you paid someone on Upwork $50 for their content, now you can use an AI tool and just create the content, the same worthless content, for free. AI as a tool is a tool creating something that’s not necessarily useful for the end journey of the company and for the user journey in general. However, if the content you were creating was pretty useful, and now you’re using AI to create really useful content for cheaper and better, of course, you can use it.
An example of a place you can use AI content is if you’re an e-commerce site and you’re selling your own products, of course, you can use AI content to write product descriptions. It’s not a content website. There’s a lot of big companies out there, and if anybody wants to look at some of the e-commerce sites and how they do SEO, a lot of them have large SEO teams. The typical JCPenney, the Nordstroms, the Macy’s, they have a lot of content on their category pages, but if you look at the keywords that drive traffic to those pages, it’s the products on the pages.
If you’re looking for shoes, the fact that a Macys.com page will have a lot of content about what shoes do doesn’t do anything for the user. They’re just looking for shoes, and then there’s shoes on the page. I think Macy’s actually ranks pretty well on those kinds of things. Using AI content to write more fluff content that’s not necessary would just be a waste of time.
Using AI content to feed in a product, and describe what that product is, and maybe some features of that product which help a user, that’s not hurting your SEO, because what you’re trying to optimize for is the product and the product name itself.
Lenny Rachitsky: Got it. Basically, don’t use AI to generate entire blog posts. I know people are doing this all over the place, but absolutely leverage AI to help you add to existing pages, descriptions, titles, things like that, which I can see why Google, Google would’ve no idea that you were helped to write this thing with AI, right, if it’s just a small part of the page versus the entire page?
Eli Schwartz: In their documentation, they say that AI itself is not the problem. It’s the helpfulness, the usefulness of the content that would be a problem.
Lenny Rachitsky: You don’t have to name names or reveal anything, but do you know of a bunch of companies using AI now to generate tons of high successful pages, high converting pages in some way? In this way?
Eli Schwartz: There’s this hatred of AI content by users in general. I think once companies declare that they’re using AI, people get upset about it. It was a Red Ventures company, I think it was CNET that said they were using AI to create content. I think they got in trouble because they said they’re using AI to create content, but if they had not said they’re using AI to create content, I think their entire model would’ve been very successful.
There’s no reason you can’t take in, they take 10 products, and they want to review 10 different products. There’s no reason you can’t tie that all together, have AI write the original piece of content, and then have a human editor just review it. I don’t know, years ago, it’s not AI, but it’s sort of like AI, but there are sports websites that will take in a lot of the things that happen in the sports game to merge it all together into a piece of content.
If you look at any earnings reports on public companies, it’s sort of the same thing. Their company will issue, they’ll file their earnings report, and then now it’s AI. Before AI, it was basically like Mad Libs kind of content. It would extract pieces of the earnings report and then it would write a blog post, which is actually fairly useful. I don’t think there’s an issue with it itself. I don’t think users have an issue with it.
I think if you read that old piece of content, again, on earnings reports, it’s pretty obvious that it wasn’t written by a financial journalist, but it was useful. You got a quick summary, you didn’t have to go into the SEC and read an earnings report.
Lenny Rachitsky: There’s this guy, Noah Smith, he’s on Twitter, he’s got a newsletter called No Opinions. It’s awesome. He had this tweet about how we’re approaching an age of a ton of slop of content, just a lot of really bad content generated by AI for better or worse, this is related to SEO, but it makes me think about just how much bad stuff we’re going to see.
To your point, it doesn’t matter whether it was generated by AI or not, people will gravitate towards and Google will gravitate towards stuff that is useful and good, whether it’s written by people or not.
Eli Schwartz: This is a huge issue for Google, because the amount of content being created is enormous, and it’s growing exponentially. Google’s trying to crawl everything. They now have more content to crawl, which becomes more expensive for Google, and which is why you have this backlash by Google against many websites, because they’re trying to clean the index to save their own costs and to protect the users from having to see this awful content.
Lenny Rachitsky: The crazy thing for LLMs is now to be trained. They’re trained on content on the internet in a big way, and they’re struggling with not training their LLMs on stuff that AI has written because it becomes this, a bad direction that LLMs will go. They’re trained on themselves. Anyway, something I wanted to come back to around AI, which is AI Overviews. I imagine many people are like, “How do I get into that answer? How do I get my product into the answer that Google gives at the top?” Is that something you recommend people try to do? Is that something you can do?
Eli Schwartz: I think of AI or AI Overviews as a branding exercise. Getting into the AI Overview, and Google has links, I mentioned this earlier that Google has links within the content, and I think a lot of what Google does is around potentially liability protection. They’re having links. They’re saying, “Well, we didn’t tell you to jump off the Golden Gate Bridge, this website, which we summarized, told you to jump off the Golden Gate Bridge,” or, “We didn’t plagiarize. We just linked to the piece of content. We may have extracted more from the content than we were supposed to, and the law hasn’t decided that yet, but we linked to it, so it’s fair.”
Right now, Google has these links in the content, and it sort of links off of it. I did a survey on LinkedIn, and I was surprised by the answer. I thought most people did not click on the links. I got a couple hundred responses, and it’s sort of 50/50 and based on my SurveyMonkey experience. It’s not a statistically significant survey in general, but it will never go to 99% and one. That’s fairly indicative that people are clicking the links, which surprises me, because I don’t think that the links are necessarily meant by Google to be useful, which means that AI Overviews is essentially duplicating search results.
You have this AI Overview which summarizes, then you have a link so you can link off of it, and then beneath it, you sort of have the exact same thing. That’s the challenge that Google has to work through. For a company to show up in AI Overview, it’s a brand challenge. If your link shows up there, then you likely were showing up beneath it in the ranking results, so you’ve already done the work you’re supposed to do.
If you’re showing up in the AI Overview mentioned as a brand, like, these are the top CRM tools, you’re showing up as a brand, that means you’ve exercised brand efforts and it’s working out. I don’t know that companies necessarily want their content to show up because they’re giving away their content for free. They want to be showing up in a way that benefits them. If you show up as a brand, you’ve done good branding, but if you show up as a link, it means Google has stolen your content, and people may or may not click off of it.
Lenny Rachitsky: Along this lines of brand and SEO, do you have any… That’s always this question of are we investing in brand? Do we just want people to be aware we exist, or are we trying to actually drive conversion? Any advice on how to think about brand building and SEO, especially with this world of AI?
Eli Schwartz: The biggest myth in SEO is that you could just build links, and link building is the secret source and the secret sauce to growing your SEO footprint. That’s totally wrong, because the way most people build links is they buy these guest posts, or they pay for links on low authority websites that sort of look like they might have authority, and it doesn’t really benefit anyone because no one’s reading these sites.
It’s funny that everyone complained about Google as this really, really smart, all-knowing LLM, and at the same time, they think they’re dumb enough to fall for these guest posts that are on websites that no one really read. The right way to build links is to build a brand. It comes part and parcel of what you’re trying to do. If you’re building links, and going back to all the SaaS examples we talked about, so if you’ve created a bunch of content that is not relevant for the product you’re trying to sell, building links to the content that doesn’t really benefit the product, doesn’t really benefit, the links don’t benefit the product.
If you can build this product that’s awesome and everyone loves and wants to use, and you get not links, but mentions, and links might be mentions now, because the way a link before LLMs and the link 10 years ago was actually an HTML link. Now, Google can read content. They can say, “Well, you’ve been mentioned here,” that’s pretty good. Now, we acknowledge that this might be the brand, and this might be the match for that kind of thing.
That’s helping to build that brand. In general, I think of any SEO effort as promoting a brand, building up a brand, rather than building a product, and then separately building out an SEO effort, which has content that may not be relevant, and links from websites that are certainly not relevant to content that’s not relevant. It all comes together as being one big effort. The same thing you’ll do on PR, you’ll do for SEO.
Lenny Rachitsky: Awesome. Okay. I’m glad we touched that. There’s a few other directions I want to go, so I’m going to bounce around a little bit and cover some of the other stuff I wanted to extract from your deep experience in this space. One is you’ve mentioned this idea of programmatic SEO a couple times versus editorial. Maybe first, just clarify what those two directions are, just for people that don’t exactly know what you mean. Then two, what’s your general advice for which direction to go, programmatic or editorial?
Eli Schwartz: It might be obvious at this point, but I’m going to say it’s dependent on the user, but essentially, what programmatic is is it’s taking a bunch of data sources and building out a page that is a combination of all these data sources. In my book, I talk about two of my favorite programmatic SEO companies. One of them was created by a past, well, it’s not created, the strategy was created by a past guest of yours, Luc Levesque, who introduced us, and that was TripAdvisor.
TripAdvisor took in all these data sources of, “This is the hotel, these are the cities,” and then it allowed UGC to combine into one comprehensive page about each property. TripAdvisor did not write a piece of content on every single hotel. They don’t write a piece of content on, “Here’s this long form blog post as an influencer on the Waldoff Astoria of New York, and here’s a long form blog post about the Marriott Marquis in San Francisco.”
They built it all from a programmatic standpoint. They took all these data sets, they took all the countries in the world, they took all the properties, and they merged into one comprehensive page, which has dominated the top of search results since the beginning of the time. They still own it. That team has cycled through many, many leaders, many different, a lot of different things, and they’re still number ones. It’s the strategy that has allowed them to build that brand and to dominate it.
The second example is Zillow. What Zillow did is they solved an early problem many years ago in the real estate space, which is no one understood what the value of a piece of property was. Zillow took in all of those data sets. They got, some of them are government data sets, some of them are their own comparison data sets based on other home sales. The government can say a house is worth one price, and based on their comparison, Zillow can completely disagree with it.
They’re building in these data sets, and they’re taking photos from realtors, they’re taking neighborhood data, they’re taking school data, and that’s a programmatic SEO page. SEO is the only channel to bring in traffic to each one of these pages. They’re not going to do paid traffic for my house, they’re not going to do paid traffic for your house. The only way someone would find that page on Zillow, either they start at Zillow.com and then just navigate through it, or they Google it. Those are programmatic SEO efforts.
On the flip side is editorial. Editorial would be, in TripAdvisor’s case, writing out that long form piece of content around each hotel, each city, and all the things they’ve done. Zillow would be the same. Zillow would take 500 million properties and write out an editorial piece of content. Neither of those would be a fit for their model, and it would be extremely expensive. Say each piece of content would cost a thousand dollars, that would be cost prohibitive for anyone to do.
Going back to my distinction based on the user, what is it that the user’s looking for? Is the user looking for the value of a home? They don’t need a long form piece of content. They need a piece of content or a page that just says what the value of that home is. TripAdvisor’s case, they’re looking for a single thing, a rating, or maybe they need some of that UGC, but really, they’re looking for that rating. That long form piece of content, there’s a purpose for it, but not on TripAdvisor’s site.
Understanding, what is it that the user needs will help solve, do you approach this from an editorial standpoint, or do you approach this from a programmatic standpoint? Some of the companies I mentioned earlier, like G2 and Capterra, which are completely being disrupted, so they’re being disrupted by editorial, but they’re also being disrupted by Google itself with their AI Overviews.
If you look for the top CRM tools, you have G2 or you have Capterra, but you also have Forbes writing out a long form piece of content on what Salesforce is, what HubSpot is, and that’s probably unnecessary.
Lenny Rachitsky: I think generally, you would, programmatic is the better path and the most common success path is what I’m hearing. What are things you need in place to be successful in programmatic SEO? Like basically, you just need some sort of data source that you own. Is that just… What’s a checklist of like, “Here’s what tells you you have a big opportunity here?”
Eli Schwartz: Programmatic is the right solution when there’s scale and when there’s a user use case. If there is not, you create scale for nothing. An earlier version of programmatic, again, pre a lot of the internet companies now, is building out a piece of content for every single zip code in America. Say you offer some sort of local product. You would build out a page for every single zip code that is, in theory, programmatic, but users don’t look for it, and Google has completely disrupted that with their own local products.
Programmatic can always be something. I see programmatic mistakes all the time, where websites look at, they say, “Look, we’re making reviews of software. There’s 10,000 different softwares in the world. We’re just going to make out a page for each one and we’re going to combine it with, I don’t know, a city.” You can make programmatic pages like that, but there’s no use case. Really knowing what programmatic solution to use comes down to, “Is this something that someone would actually be looking for? Does this page itself provide a solution?” Zapier provided a solution because someone was looking for that combination, but there are many tools like Zapier, which try to be Zapier, but they’re not looking for that solution. Going back to some of the template companies you mentioned, I don’t think a lot of companies that make templates, they’re making programmatic, but there’s not a use case for every single one of those templates. They’ve made programmatic for no [inaudible 01:15:53].
Again, SurveyMonkey was one where we had a fat head of templates we could make, but it wasn’t a long tail. We could make a hundred templates, but we couldn’t make 10,000.
Lenny Rachitsky: I want to do a rapid fire of SEO myths. Before I do that, I know that you spent a lot of time actually reading the recent ruling against Google and their whole deal with Apple, and you were telling me it revealed there were a lot of interesting stuff about their strategy and where search was going. Let’s spend a little time here, just what did you learn doing that exercise?
Eli Schwartz: It’s fascinating. It’s 286 pages. I basically read a book. It was a verdict.
Lenny Rachitsky: Oh, geez.
Eli Schwartz: I would say that the judge or the clerk that wrote this verdict has a better understanding how the SEO works and how digital marketing works than most people I meet in digital marketing.
Lenny Rachitsky: Wow, I did not expect to hear that.
Eli Schwartz: Yeah. This verdict explains how Google decides to rank a page. The verdict explains how the auction model works in ads, in Google Ads. I think that’s fascinating. Talked about the comparison between social media. This court really understood a lot of the questions that were posed to them. For example, Google tried to say that they’re not a monopoly because they’re social media. The court analyzed that question and said, “Social media is not a comparison to Google search.”
Google tried to say a lot of the things they did were non-monopolistic because there’s alternatives like Bing. The court really analyzed a lot of these things. Why I found it fascinating was, again, a lot of the testimony was public, is the summary of how the court put it all together in a verdict. A few of the things I learned is one, the market share, Google’s market share. That’s been a number that no one really liked to talk about.
Google would say, “Oh, we think it’s like 80%,” because they didn’t want to be viewed as a monopoly. Microsoft would say, “Well, maybe we have 20%.” I just found a recent document from Microsoft where they claimed to have a lot of market share in search. In the verdict, it says that Google has 98% of mobile searches. I forget whether it said that the court said it, and it said Google didn’t disagree, or that came from Google’s own documentation, but 98% of mobile searches.
The thing that I thought was most fascinating from the verdict is how much the default partnerships that Google has contributes to Google’s success. The plaintiff in the case was the Department of Justice suing Google for being a monopoly. However, I don’t know what the word is, but one of the complainants in this is this company, Neva, which was a search engine started by a past Google, I think it was like a head of research or something. A senior person at Google created a new search engine.
A lot of the trial was focused on how Neva could not be successful, despite having a better search engine, despite having a better experience for users, because they didn’t have any of these default distribution agreements. That’s the part that I thought was most fascinating. Google has these default agreements with Apple, and they also have Chrome.
They have all these inputs into Google searches, which means that I don’t think there’s a chance, and this is the biggest takeaway, I don’t think there’s a chance that ChatGPT, or Perplexity, or Claude, or any of these other LLM startups, because they’re startups now, even OpenAI is a startup with Microsoft’s investments, still a startup. I don’t think they have a chance at really unseating Google purely from a quality standpoint.
That’s what a lot of the documentation in this verdict is about is how much a force of habit, how much the brand, and how much the distribution agreements Google has, drives searchers to Google. There were a couple of very interesting insights. One is that Bing tried to give itself to Apple for free, and Apple said there was absolutely no price that they would take to use Bing search within Apple.
Another one was Mozilla had a partnership with, I think it was Yahoo, and they did a rev share deal with Yahoo, and everyone switched back to Google. When people say, “Well, SEO is dead, everyone’s going to go to OpenAI,” and you just look at the 25 years of Google’s existence and what they’ve built, I really don’t think so. I really think distribution and the force of brand will keep almost everything Google has today.
Now, why is Google fighting back so hard to build AI Overviews? Well, I think there’s two answers. One is, like I said, Wall Street. They really want Wall Street to think that they care and they’re doing this thing. The other is losing small percentage points to ChatGPT costs Google a lot of money, so it’s worth it for them to invest in it, but I don’t think that they have any fear about not being the most dominant search engine in the world.
Lenny Rachitsky: Well, isn’t the idea with this trial that this may change, and they may lose that default status, whether it happens or not, that’s the whole game here, right? It’s like maybe they can’t be that anymore, and that maybe opens up. I think that’s the opportunity potentially.
Eli Schwartz: I don’t think so. Yeah, I’m no legal expert. I don’t know what the DOJ will do, but what’s interesting, if you saw Google’s press release when the verdict came out, so Google tweeted, “We thank the court for recognizing that we’re the greatest search engine in the world, and more to come.” They’re going to fight back, and it’s almost like a bully being declared the world’s greatest bully. They’re like, “See, told you we’re strong.”
When reading the verdict, it actually said, “There is no competitor. Google is the greatest search engine. They have 25 years of great data. They’ve done a fantastic job.” I don’t see any way that those, if they break the default agreements, Apple actually loses all that money Google’s paying, and then users will still Google. That’s not what will do it. I think the only thing they can do is figure out how to prop up a competitor.
Lenny Rachitsky: Yeah. I think that’s an interesting takeaway. People will go to Google, will switch to Google even if it’s not default. There’s actually a really interesting takeaway from Ben Thompson’s analysis of this stuff, and I will point to the podcast episode. I think that shifted my perspective on this, because when you hear Google is paying Apple to be the default search engine, it sounds nefarious and unfair. The deal is actually revenue share. They give them a percentage of the revenue from the ads, which ends up being $20 billion, something like that, a lot of money for Apple. Anyone else can come to Apple and say, “We will give you a revenue share of our searches also.” It’s just nobody is as big and can pay as much. If you think about it, why can’t Google come to Apple? “Hey, we will send you tons of money by doing the search for you and giving you a large percentage of the ads that we are doing for you.”
It’s weird to not allow that if you really think about it, but it’s also unfair, because no one else will ever be able to compete with that.
Eli Schwartz: Yeah, and that’s what the court said is like Neva didn’t have a chance because of that. One other point is, so I did this survey, this was years ago when I was at SurveyMonkey, I did a search market penetration survey to figure out who used what, like Bing versus Google. On my survey, I was generous. I threw DuckDuckGo on there, and I came back, I had thousands of responses, and came back that I had 1% of people use DuckDuckGo.
I didn’t realize it, that DuckDuckGo didn’t actually know what their market penetration was. When I shared the survey, Gabriel Weinberg, who’s the CEO of DuckDuckGo, hit me on Twitter and he said, “Can you share the data with us? We really would like to see it.” This was, let’s say 10, 12 years ago. In the verdict, it said that DuckDuckGo’s market share is 2%. I came back with 1% 12 years ago. DuckDuckGo, I think they raised a hundred million dollars.
They have all these brand partnerships. They have, at baseball games, they have their logo, they’ve done everything, and all they’ve been able to do is get from 1% to 2%. There’s a lot there, and Google being Google, I don’t see anyone knocking them off.
Lenny Rachitsky: Fascinating. Well, thanks for sharing your takeaway so that we don’t have to read this report. Before I move to rapid fire SEO myths, is there anything else along the lines of AI or SEO that you think is really important to share that you want to leave listeners with?
Eli Schwartz: I think the most important thing to leave people with is SEO is not dying. There will always be a world where users are requesting their own information. One of the reasons I think these home assistants, whether it’s from Google, or Apple, or Amazon, have never taken off, is because you don’t have choice. You talk to your Google, and it gives you only one answer. You talk to Amazon, and it gives you only one answer. Like, “I want to buy toilet paper,” and Amazon just buys you whatever it decides to buy you.
There will always be a world where there needs choice. One other piece from the verdict is Google believes that we’re in the very early days of LLMs, and even with all the machine learning they do and understanding users, you will always need real user data, which Google has on past searches. I think users will always be requesting their own information, and there’s never going to be an AI that understands you so perfectly, it’s going to know that for you personally, you would like to click result number five.
That’s the best fit for what you’re looking for right now, which means that there always needs to be multiple choices to make. Yes, most clicks are probably going through the first one, two, or three results, but there needs to be seven, 10, multiple pages of Google, because some people do go to those other pages. I think that’s the most important thing to underscore is that all of this means that search changes, a lot of top-of-funnel search goes away, but in general, there’s always going to be a world where people are doing the searches.
Then the last piece on that really is when you’re looking to do something with search, you’re looking to take an action. Companies benefit from those actions. Say you’re a hotel, and you want people to sleep in your hotel and pay you to sleep in your hotel. The aggregate number of people needing to take that action and pay you for it does not change even if search volume gets cut. If you’re selling shoes, people still need to buy shoes.
If you’re selling information, though, your media, if you’re WebMD, yes, your revenue declines, because that information that you now aggregated and curated and gave away for free in return for people clicking on ads, Google is now going to give away for free in return for no one clicking on ads.
Lenny Rachitsky: Interesting. I wonder if there’s an investment arbitrage opportunity of predicting which businesses will decline with this and which will thrive. Anyway, I want to talk about SEO myths. I know that you have a few things, and we’ve talked about a few of these, I think, of just things that you know people believe about SEO, but they’re actually wrong. Let me just ask you, what are some myths that people get wrong about SEO?
Eli Schwartz: Probably the biggest one we’ve said over and over is they even need to do SEO. There’s this always default assumption that you’ve raised money, you have a marketing team, you should do SEO. I think that one is really worth putting to bed is think about that user journey. Should you do SEO? It’s a question like, “They ask about other channels, should you do brand…”
It’s a question like they ask about other channels. “Should you do brand marketing? Should you take out an ad in Times Square?” And no one will say, “Well, I just raised a million dollars. I should totally take out an ad in Times Square.” And congratulations on being in Times Square, by the way.
Lenny Rachitsky: Thank you.
Eli Schwartz: No one will just decide, “Oh, I’ve got money to burn, let me just burn it.” So somehow SEO becomes this thing of like, “Oh, I’ve raised money. Now, I need an SEO team.” So that’s I think the first myth that’s really important to just put to bed.
Lenny Rachitsky: Awesome. What else?
Eli Schwartz: Link building. So link building is brand building. So it’s thinking about how you’re going to build a brand and have people mentioned to you. And when you’re doing link building, you’re creating a relationship between the piece of content that has linked to you and the product or whatever you’re trying to monetize. So that should be link-building. Just getting the idea of getting links on… An HTML link will equal SEO success, again, I think that’s completely wrong. And then I mean probably the biggest myth in general is thinking that Google itself is a black box. So I think there’s nuances to how you rank, and no one can really unearth what is it that will you be number one and what it is that will make you be number three. But the basic idea of how you build SEO is very simplistic and Google has a best practices guide on what it is that you should be doing, which is build a website that Google can understand, link to the pages in that, write content that is helpful and that users want to read. And those are the basics. And starting with that is how you’re going to build an SEO strategy that improves upon it. But the idea of if, “I take all three of these steps right now. I’m guaranteed to be number one,” I think is something that is completely incorrect. And there’s an assumption that if you do a lot of things around SEO, you’re going to be successful. And I think that that’s incorrect.
Yes, there are nuances. But for the most part, you get from most… Zero to 10, you can get to step eight by just doing the best practices.
Lenny Rachitsky: Oh, wow. It sounds very empowering. I like the sound of that. Maybe just on that thread briefly. If someone wanted to start down the road of SEO, I know you’re going to be a little bit biased because you help companies with this, but do you recommend bringing someone like you in first or having someone just give it a shot, listening to this podcast, reading some books and blogs, or something else?
Eli Schwartz: Yes, so I am biased. I think they should have the right people give them advice on SEO. So if someone wants to be… If they validated that there is an SEO effort, paying for help on SEO, it goes back to that resource discussion we had before. So if you want to hire a growth advisor, a growth advisor might be very expensive, but you squash a learning curve so you don’t make as many mistakes as fast.
If you hire someone in-house, and a lot of times they see this with roles where I’m helping companies hire someone for SEO, they don’t have a lot of budget, which means they’re going to have someone that doesn’t have a lot of experience. So is that the right decision to make? If you’re hiring an agency, agencies are typically expensive, and agencies a lot of times get paid on deliverables. So are the deliverables they’re doing worth paying for?
An easy deliverable that a lot of agencies produce now is content, but that goes back to our earlier discussion. Do you even need the content? It’s very easy for agencies to sell content as deliverable because it’s something they actually deliver. Delivering strategy the way most of my growth advising works is very difficult because I don’t have a strategy to offer in a proposal because I haven’t developed a strategy, and we don’t actually see the fruits of that until we implement the strategy. But agencies can end-run that by just saying, “Oh, you pay us on a monthly basis and these are the things we’re going to shift to you.”
Lenny Rachitsky: Something that I wanted to touch on, which might be a myth and might not be a myth. But I think you’re going to say it’s a myth based on what you said about Google being so dominant. There’s a lot of talk about TikTok and Instagram replacing Search for people. I actually do search TikTok a lot now for like, “How do I solve this problem? How do I cut a watermelon? How do I, I don’t know, find a cup for my baby?” It’s really good. What’s your take on TikTok and Instagram videos basically replacing a lot of searching for Gen Z, especially, and younger kids?
Eli Schwartz: In the Google Verdict, I think Google’s said this before. 63% of Gen Z uses TikTok to search.
Lenny Rachitsky: Oh, wow.
Eli Schwartz: Yeah. I think that’s just a headline number because it comes down to, you said it over and over in the podcast, user journeys. So there’s going to be things that are more appropriate for TikTok or Instagram, and there are going to be things that are more appropriate for Search. So I think if you’re doing top-of-funnel discovery, you’re going to maybe watch TikTok videos to learn more about the topic. But as you go into the mid-funnel, so say you want to… A popular TikTok search is around travel. So you want to go on a trip. You want to go to Southeast Asia. You don’t know where exactly you want to go. So you might want to watch a bunch of videos and see influencers and see experiences. But now you’re ready to book. And you want to book a hotel. You want to book your flights. All of that is not going to happen on TikTok. All of that’s going to happen on Google Search. So that’s the mid-funnel again.
A lot of search will move different places, but ultimately, Google is still the right place to do those searches, to do those mid-funnel searches. So maybe it was more inappropriate that there ever wasn’t TikTok for you to get those better experiences. For those top-of-funnel things, you had to suffer through reading awful content that was written only for SEO purposes, and now you get rich dynamic content that comes from TikTok.
And one interesting piece on this is if TikTok does get banned in the US, which looks like it’s likely, a company that will pick up all of that users is YouTube. So YouTube has YouTube Shorts. And again, we’re looking at a potential monopoly with Google as Google tries to put those YouTube Shorts to solve the problem that TikTok was now solving. So that will be kind of interesting.
Lenny Rachitsky: Oh, man. Google all the way down.
Eli Schwartz: Yes.
Lenny Rachitsky: A thought I had as you were talking about this idea of mid-funnel versus top funnel is a simple way to think about the mid-funnel is where there’s intent. You actually have intent to now buy a thing, a specific thing. Yeah. Cool.
Eli Schwartz: I mean that’s what it is, but it’s a buyer journey. So at the top of the funnel, you’re curious. You don’t know if you have intent. And in the older way of doing SEO where you only focused on top of the funnel, you only focused on rankings as a KPI, which I think is sort of the incorrect way of doing things. There is no intent. So you showed a KPI that matched with a non-user, which matched with a non-buyer.
As you move down the funnel, there is intent. So there is a user and there’s less traffic. There’s always going to be less traffic at the middle of the funnel. But that doesn’t matter because your KPI is closely related to your business metrics.
Lenny Rachitsky: Let’s do a couple more myths if there’s anything else that we missed, and then we’ll wrap up and get to our very exciting lightning round. Is there anything else that you think is an important myth that people get wrong about SEO?
Eli Schwartz: I’ll touch on one more myth, which is that technical SEO is an easy solve to any SEO problems. So there’s been a lot of Google updates recently. Again, I think… I don’t like to defend Google because in reading the Verdict, I think that Google did some pretty evil things. But a lot of what Google has been doing is pushing back on the things that users hate. There’s a lot of bad SEO content out there. So Google’s trying to get rid of a lot of this bad SEO content in their recent updates.
When sites get hit by these updates, they try to solve the problem that they think happened to them by doing technical SEO. So they’ll reach out to me or to someone like me and say, “Can you do an SEO audit and understand why our traffic has suffered?” The reason their traffic has suffered is because they’ve done things that were not really useful for users and they sort of polluted Google. They need to solve that problem. Technical SEO has a place. So if you have a large website with tens of millions of pages, if you’re Zillow, how you link to each listing, if you’re Airbnb, how you link and allow Google to crawl a site and understand a site, very, very important.
If you have a hundred-page website that sells a SaaS tool, technical SEO is probably going to be less important. So spending money and time on technical SEO would be a waste of time. So I think that’s another big myth is you solve SEO problems with the right SEO solution rather than, “Let’s do an audit or let’s get better links or my page…” Page speed is another one that… Agencies like to sell on… If things related to page speed, Google keeps changing the words of what that is, but it doesn’t matter as much.
Because at the margins, yes, maybe if you’re Kayak and you’re competing, you get Expedia, and Kayak is maybe slightly faster, maybe that matters. But if you’re in a space where you’re competing against lots of slow terrible websites, it doesn’t matter at all. So spending a lot of money on making that fix is not the right thing to do.
Lenny Rachitsky: I like how you just make SEO feel much more approachable and simple and something that you could just do and not have to again figure out the dark arts that…
Eli Schwartz: It is not a dark art. It should be for most things, it is simple. And then there are tons of companies out there get in a lot of trouble, and there’s billions of dollars to be made on doing the right SEO. But a lot can be unlocked by very simplistic SEO. PMs that don’t really have an SEO background can do enough SEO to make a lot of money for the company without needing to be an SEO expert.
Lenny Rachitsky: I love that. That’s actually how it worked at Airbnb. The PMs work in SEO were not like historically SEO legends. They were just PMs figuring out SEO and they had a really big impact. So that really resonates with me. Eli, is there anything else that we haven’t touched on that you think is important to you or something that you think might be useful to people to leave them with before we get to our very exciting lightning round?
Eli Schwartz: Yes. So I think we talked about how to decide whether to resource SEO, but we didn’t talk about how to understand what the expectations are from SEO. So this is something I came up with while I was working at Faire, which is a very interesting company. So it’s F-A-I-R-E. They are a wholesale Shopify. They were launching in a new country, and they wanted to understand what the upside of launching in that country was from an SEO standpoint. So the way most people do SEO forecasting is they do a bottoms-up forecast, which is they look at keywords. So say we want to sell a pair of shoes.
You do a bottoms-up forecast, you go onto a keyword research tool, and they’re all fairly the same. Look at how many people search for shoes every month. And then you try to estimate what your ranking would be on that word shoes, and then you estimate what you click through it would be, and then based on that, that’s your clicks. And then you get a conversion rate, and that’s your bottoms up.
And you make an assumption that the word shoes in that tool didn’t capture all the people that search like white shoes and black shoes and running shoes, so you just gross it up. You just pick a number and you gross it by 10, and that’s your SEO forecast.
Now, the problem with that is that usually it’s too small. So you get this number and now you’re pitching to launch in a new country. I want to launch in Japan and sell shoes. And you say, “Well, I think I can get 500 users because based on the way I got here, this is my click-through rate and this is my rankings and I’m 500 users.” And no one will ever fund that. That’s too small.
Now if you’re launching in a new country, it’s actually fairly easy to figure out what your SEO upside is because you can take the population in the country. So this is not something Pharoah sells, so I don’t have to… So I’m not giving away anything. But say you’re launching shoes. You’re launching a new product website that you’re selling shoes in a country like Japan.
I’m going to guess. I don’t know the actual numbers for Japan. So say Japan has a hundred million people and you’re selling shoes. And you’re not going to sell shoes to a hundred million people. We’re only selling shoes to men. And let’s just divide it in half. 50% of Japan is men, so we got 50 million people that would buy shoes.
But now, we’re also going to say, “Well, there’s older people and younger people and we’re not selling shoes to them.” So now we can take 25 million and cut that in half. And then from there, again, we’re selling shoes on the internet. Not everyone buys shoes on the internet. Some people buy shoes in a store. Some people buy shoes on the internet. So you take some percentage rate of the amount of people that are going to buy shoes on the internet and say, “We’ll take 10% of our 25 billion people in our market. 2.5 million people buy shoes on the internet and we want to do SEO for 2.5 million people. And our expectation is we want to get this number of market penetration, whether it’s 10% or 50% or 100%.” And there’s your number and multiply that number by the amount of shoes they’re going to buy every year and your AOV and there’s your forecast.
And that may be inexact. And you can tweak those numbers up or down. You can say, “Well, my AOV was wrong. My market penetration was wrong. I was wrong on the total population. I didn’t realize that in Japan, no one bought shoes on the internet.” But at any point in time, you can go back and adjust your forecast. Whereas if you’re doing this bottoms-up forecast, it’s actually in many cases wrong to begin with because the keyword research volume is wrong.
I worked with some really fascinating companies where the largest query in their space was the biggest query. So I worked with WordPress. So WordPress, the word WordPress is the biggest query in the web development space itself. There’s no other word that’s as big as WordPress. And the number that every single keyword research tool had was completely wrong of what Google Search Console said for the word WordPress is, “When you’re building these forecasts based on keyword research tools, that first number that you build the entire forecast off of if it’s wrong, your forecast is wrong.” So when you do this top-down, it’s a TAM forecast essentially. When you do a top-down, you’re closer to the truth. Now, you probably aren’t going to get to the truth. I’ve never seen a product plan get to the truth of what it could do, but it will help you make a better decision than if you just guess.
Lenny Rachitsky: Wow, this is a massive insight you’re sharing. So you’re saying the keyword research tools are not actually that accurate in terms of the opportunity there. And why is that? That begs the question. Why are they so off? What are they doing wrong? What are they missing?
Eli Schwartz: They have their own secret sauce for estimations. So even Google, again, for past monopolistic reasons, Google’s not allowed to give the real number that they see on Google Ads for how many people search it. They have to buy it from another data source, and then give that out. I forget why. They’re all basically guessing. So they’re using whatever sort of proprietary algorithms to guess, which is why a lot of them are not aligned because they have their own algorithms. So you’re using, whether it’s Semrush or Ahrefs, one of my favorite tools is Similarweb, which Similarweb has a lot of browser plugins, which snoop on the way people are searching. So whatever it is that they’re… Because Similarweb has browser plugin, but they’re not seeing every single person’s search, so they have to use algorithms to estimate how much the entire world would be.
And again, I don’t know that any of them could get close to truth was when I… I worked with big companies where there were keywords that I could look at Google Search Console, some of them, the tools were overestimating by 10 times. Sometimes they’re underestimating by 10 times. So I’m not saying they didn’t get the exact monthly number wrong. I’m saying they got it wrong by many factors.
Lenny Rachitsky: Wow, that’s crazy. So your advice is… Do you just ignore those numbers? Or is it just, “Check it out but don’t roll it. Don’t use that number”?
Eli Schwartz: They’re indicative. So I would say if you want to know like, “Do most people spell WordPress with a space or WordPress without a space?” It’s pretty indicative that people spell it without a space. But if I were building a forecast and say, “Oh, I absolutely choose this based on this defined number.” I don’t think so. I can use it for normalization. And I use them only for normalization to understand how do people search. And I really like user journeys and they’re helpful for understanding user journeys. But as an exact science, it’s hard to really use that.
Lenny Rachitsky: That’s wild. And so it’s mostly useful for order magnitude and comparing one keyword to another relative in this bit. Okay, interesting.
Eli Schwartz: Yes. I mean I have a great example of how wrong it could be in estimating traffic. So there was a company I was working with early in COVID. The public company, their board member emailed them and said, “You guys are getting crushed by your competitors in COVID because look at your competitors. I’m looking at one of these tools and you’re at the bottom and you’re doing everything wrong.” And the CMO says, “What do I tell this board member?” I said, “The board member is completely wrong. This is the Google Search Console.” And Google Search Console is not perfect, but again, it’s real data and our Google Search Console shows that our traffic has quadrupled in COVID, so it doesn’t really matter what this external tool shows. So they’re helpful. Tools are helpful. But I don’t think they are a source of actual truth.
Lenny Rachitsky: Wow. That’s an awesome point in closing. One thing I had noted here that I want to make sure you have a chance to talk about is just you’re really passionate about helping people get into SEO and also just becoming advisors the way you are. Share what you think might be useful to people along these lines.
Eli Schwartz: As my own forecast and prediction, I think the need for SEO expertise is going to explode because a lot of what is happening in the search layouts is going to mean that companies have to pivot their approaches. So again, companies focus on rankings, they focus on traffic, and a lot of that goes away. Suddenly the layout changes, the traffic changes. It’s not necessarily going to impact their bottom lines if their SEO wasn’t a right fit. So this is going to create a lot of interest in SEO help, and I’m seeing this.
Over the last year, my own inbound has really grown because there’s a lot of questions as things shake out. So there will be people that want to go into SEO consulting, and I think there’s going to be a huge need for it in growth advising in general.
What I would say is, and I had amazing mentors along the way, some of your past guests like Casey Winters and Yuriy Timen, Ethan Smith, they’re all guests with great episodes, but they advised me and they shared with me not on how to be a better operator, but on how to do sales better, how to propose better.
And for anyone out there that wants to become a growth advisor, I’d say that that’s the skill you really want to perfect. Communication, sales, proposals, and not really worry about being the best operator. You should be the best operator, but that’s probably the skills you already have. And don’t make an assumption that because you’re a good operator today, you’ll also be a great growth advisor. Build that growth advising muscle by staying on your day job. Don’t quit your day job, and moonlighting and practice selling, closing, working, retaining, and that’s where if you’re successful there, you can be successful on your own.
Lenny Rachitsky: Amazing. And I know you have some posts that get into this stuff that we’ll link to you, right?
Eli Schwartz: Yep.
Lenny Rachitsky: Awesome. It’s kind of think of yourself as a product and the journey of working with you.
Eli Schwartz: Yes. You’re building a brand. You’re not building a consultancy.
Lenny Rachitsky: I love that. Eli, this has been amazing. And with that, we’ve reached our very exciting lightning round. Are you ready?
Eli Schwartz: Absolutely.
Lenny Rachitsky: Here we go. First question, what are two or three books that you’ve recommended most to other people?
Eli Schwartz: As has been apparent, I really like user journeys and understanding people. So there’s a book called Small Data by I think Martin Lindstrom. I don’t know if I’m pronouncing his name right, where he talks about understanding people and understanding how people buy. And he digs into that entire process. He actually goes and lives in people’s houses and watch how they use different tools and toys. I always find that book to be fascinating, and recommend to people who want to understand users.
Simon Sinek’s Start With Why. Again, the same idea. Really understand what a product a business is supposed to be doing to understand users. This is specifically for growth advisors and not necessarily for PMs, but Million Dollar Consulting. It’s a book I read, recommended actually by Ethan Smith many years ago. A fascinating book on how to build a brand and become a growth advisor. I ended up working with the author as a coach for almost a year. Amazing book.
And then the last book that’d be remise in not recommending is my wife’s book, which is How To Stop Caring What Others Think: For Real. So it is a book for precisely that. Understanding your own successes, not worry about what other people think about you.
Lenny Rachitsky: Beautiful. Is that in the background, by the way? And if not, you got to put it in your background, your wife’s book.
Eli Schwartz: My wife made the background. So yes, it is in the background.
Lenny Rachitsky: Which one is it? Which color am I… Just so people recognize it. Oh, the big one right there. I was wondering why that one was a little taller than the rest.
Eli Schwartz: Yes, this is the book.
Lenny Rachitsky: There it is. Wait, move it up a little bit so we can see the full cover. Oh, beautiful. Stop Caring What Others Think. Beautiful. I need that. For Real, I like the ends, but for real.
Eli Schwartz: For Real.
Lenny Rachitsky: I love it. Okay.
Eli Schwartz: It may or may not be a Google Search query.
Lenny Rachitsky: Oh, man. I see. I see what happened there. Genius. Okay, next question. Do you have a favorite recent movie or TV show you’ve really enjoyed?
Eli Schwartz: I was on a plane and I saw this movie Blackberry and it looked like a documentary, so I don’t know if you’ve seen it, but it was fascinating. Because it’s one of those movies, it has a cliffhanger, and you know exactly how it ends. And it is such a fascinating movie. I didn’t know all that history, but they owned the entire smartphone industry, and they went to zero. And it really charted that journey and it makes you think like, “You can never really rest on your laurels. You have to create a product people want and understand your users and keep selling that product.” And not be like, “Well, we’re number one, this will never go away.” And great movie.
Lenny Rachitsky: Funny, it was just recommended by another guest very recently, so there’s trending up. It’s been around for maybe a year at this point. I watched it.
Eli Schwartz: Yeah, I never would’ve watched it if it wasn’t on a plane. It was totally… It seemed like a plane movie, but it was perfect.
Lenny Rachitsky: I watched that at home. And I love just the technical… Their ability to find clever ways to use the cell networks, that felt impossible, and that’s what allowed them to all these. I like the messaging and things like that. That was really fascinating because I didn’t realize they basically reverse-engineered the way the cell networks worked to allow what they allowed. Yeah, awesome movie, and crazy story.
Next question, do you have a favorite recent product that you’ve recently discovered that you really love? It could be an app. It could be some physical.
Eli Schwartz: It’s not a recent product and it’s the kind of thing I fall in love with over and over again. And it might be cliche, but it’s my phone. So I recently traveled in Southeast Asia. I lived there for a little bit. So I haven’t been back in eight years. And the things I was able to do with my phone like traveling, Google Maps, Waze, and ordering food. I went from multiple countries and I was able to use the same app to book rides and make payments. It was so useful.
Eight years ago, and I was there and I had to buy a SIM card and my phone didn’t really work and I couldn’t make payments. It changed my entire experience. I almost didn’t need a computer. So falling in love with my phone again.
And then another one that… It’s not necessarily recent, but I absolutely love, which is Grammarly. So I like writing. I write all my own stuff. I don’t write with AI. Grammarly helps me to be a better writer.
Lenny Rachitsky: You know what I’ve realized about Grammarly recently? I just upgraded to their pro plan. They’re like the best product at upselling you on their paid plan because they’re just right there in your face all the time. “Hey, we have stuff. We have so many tips for you. We have so much we can improve. Just pay a hundred whatever bucks and we have so much advice to make all your writing so much better.” It’s right there in your face all the time. They’re so good at it. And they got me. They got me. And I’m happy. It’s not like that much money in the scheme of things if you’re doing this full-time.
Eli Schwartz: I’m so embarrassed when I use Grammarly on my book and it just shows up all blue and green. I’m like, “Oh, man, if only I had known.”
Lenny Rachitsky: I actually have a copywriter on my newsletter who’s incredible. She finds a hundred things every time to improve that even when I think it’s perfect. And interestingly, she doesn’t do what Grammarly suggests in more cases than I would expect. So that’s kind of interesting I’m finding. But anyway, yeah, Grammarly is great. I use it all the time.
Two more questions. Do you have a favorite life motto that you often use yourself, share with friends or family, find useful in work or in life?
Eli Schwartz: I don’t know that it fits into a sentence, but it’s something I always encourage people, which is to really think big and think long. So I just started working with LinkedIn as a consulting client, and I’ve talked to LinkedIn about working with them for six years. So, never give up. You meet someone and propose something and suggest an idea you never know where it’ll go. I had that all through my career with living in different places, really not think about the moment of where things will go, but just it’s a relationship you’re building. You don’t ever know where anything will happen. So look at the big picture.
Lenny Rachitsky: Love that. And kind of following along those lines, last question. Curious if something comes up here. What SEO win are you most proud of?
Eli Schwartz: I don’t know if I could say most proud of. But I really like what I do with Tinder because it brought the entire journey together. There’s users out there that didn’t know that Tinder would solve a very specific problem, and it’s just there to solve that problem. And it’s not the way Tinder thought of themselves. They thought of themselves as a dating app, but it’s a loneliness solution problem that answers a Google Search.
Lenny Rachitsky: This has been awesome. I think this is going to help a lot of people think through SEO and especially as they realize how things are changing, they’re going to have this resource now to be like, “I see. This is what I should change. This is what I should be doing.” And I just love, especially the pattern and the thread of, “It’s not actually that complicated. You can do it even if you’ve never really done it before.” So I really appreciate you being here and sharing all this wisdom with us. Two final questions. Where can folks find you if they want to reach out to work with you? Also, check out your book. And finally, how can listeners be useful to you?
Eli Schwartz: You can find me on LinkedIn. So search Eli Schwartz, and you should definitely look for my book. So the book is Product-Led SEO. Actually, one piece on just your own personal brand and your own personal rankings. It doesn’t matter where you rank if you search your own name. So if LinkedIn shows up first for your name, that’s great because they’re finding you. And a lot of times, brands and people will be very focused on where they’re positioned, but it’s about the journey. If they find you, they find you. It doesn’t matter if it’s number one.
And that also underscores that it’s not all about links. I believe for my own name and probably for you too, I outranked LinkedIn. And my book, so if you search Product-Led SEO, my own personal website, which does not have the best domain authority, it outranks Amazon. So that should right away disprove that it’s all about links. That’s all about SEO metrics. It’s really all about the right fit. So be the right fit, and you’ll show up where you’re supposed to.
And then users can be most helpful for me by subscribing to my newsletter and giving me ideas and feedback on what to write. I really enjoy writing, and I really enjoy hearing from people. So my newsletter is Productledseo.com.
Lenny Rachitsky: It’s so consistent across all your things. I love it. Productledseo. Eli, thank you so much for being here.
Eli Schwartz: Thanks for having me.
Lenny Rachitsky: Bye, everyone.
Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com. See you in the next episode.
Glossary
| English | 中文 |
|---|---|
| Ahrefs | Ahrefs(SEO 分析工具) |
| AI Overviews | AI 概览(Google 的 AI 搜索摘要功能) |
| Airbnb | Airbnb(短租住宿平台) |
| algo update | 算法更新 |
| AOV | AOV(客单价,Average Order Value) |
| Ben Thompson | Ben Thompson(科技行业分析师,Stratechery 作者) |
| beta | beta(测试版) |
| Bing | Bing(微软搜索引擎) |
| bottoms-up forecast | 自下而上预测 |
| buyer journey | 购买旅程 |
| Capterra | Capterra(软件评测平台) |
| Casey Winters | Casey Winters(增长与产品顾问) |
| Cleveland Clinic | Cleveland Clinic(克利夫兰医学中心) |
| click-through rate | 点击率 |
| CMS | CMS(内容管理系统) |
| CNET | CNET(科技媒体网站) |
| CPM | CPM(千次展示费用,Cost Per Mille) |
| CRM | CRM(客户关系管理) |
| default agreements | 默认协议(搜索引擎与设备厂商/浏览器的预装合作协议) |
| DOJ | DOJ(美国司法部,Department of Justice) |
| DuckDuckGo | DuckDuckGo(注重隐私的搜索引擎) |
| editorial | 编辑型(内容) |
| Ethan Smith | Ethan Smith(SEO 顾问,Graphmate 创始人) |
| Expedia | Expedia(在线旅行预订平台) |
| Faire | Faire(批发电商平台) |
| fat head | 肥头(与”长尾”相对,指搜索量集中在少数热门关键词的分布形态) |
| Fiverr | Fiverr(自由职业者服务平台) |
| flywheel | 飞轮 |
| freemium | 免费增值(免费使用+付费升级的商业模式) |
| funnel | 漏斗(营销漏斗,指用户从发现到转化的旅程阶段) |
| G2 | G2(软件评测平台) |
| Gabriel Weinberg | Gabriel Weinberg(DuckDuckGo CEO) |
| Google Search Console | Google Search Console(Google 搜索管理工具) |
| Healthline | Healthline(医疗健康信息网站) |
| hockey stick | 曲棍球棍式增长(指先平缓后急剧上升的增长曲线) |
| inbound | inbound(入站流量/渠道) |
| index | 索引库(搜索引擎的网页数据库) |
| Jasper | Jasper(AI 写作工具) |
| JCPenney | JCPenney(美国百货零售商) |
| Kayak | Kayak(旅行搜索引擎) |
| keyword research tool | 关键词研究工具 |
| KPI | KPI(关键绩效指标,Key Performance Indicator) |
| lightning round | 闪电问答 |
| LLM | LLM(大语言模型,Large Language Model) |
| Luc Levesque | Luc Levesque(SEO 从业者,TripAdvisor 早期增长策略负责人) |
| Marriott Marquis | 万豪侯爵酒店 |
| Martin Lindstrom | Martin Lindstrom(品牌顾问与作家) |
| mid-funnel | 漏斗中部 |
| Mixpanel | Mixpanel(产品数据分析工具) |
| MQL | MQL(营销合格线索,Marketing Qualified Lead) |
| Neva | Neva(前 Google 高管创立的搜索引擎公司) |
| No Opinions | No Opinions(Noah Smith 的 newsletter 名称) |
| Noah Smith | Noah Smith(经济学家、时事评论作者) |
| Nordstrom | Nordstrom(美国高端百货零售商) |
| OpenAI | OpenAI(AI 研究公司) |
| organic traffic | 自然流量(非付费搜索带来的访问量) |
| page speed | 页面速度 |
| Perplexity | Perplexity(AI 搜索引擎) |
| persona | 用户画像 |
| PM | PM(产品经理,Product Manager) |
| PRD | PRD(产品需求文档,Product Requirements Document) |
| product-led SEO | 产品驱动型 SEO |
| programmatic SEO | 程序化 SEO |
| Quora | Quora(问答社区平台) |
| Red Ventures | Red Ventures(数字营销与媒体公司) |
| rev share | 收入分成 |
| SaaS | SaaS(软件即服务) |
| SEC | SEC(美国证券交易委员会,Securities and Exchange Commission) |
| Semrush | Semrush(SEO 与数字营销分析工具) |
| SEO | SEO(搜索引擎优化) |
| SEO audit | SEO 审计 |
| Similarweb | Similarweb(网站流量分析工具) |
| Simon Sinek | Simon Sinek(领导力与激励领域作家) |
| SOC 2 | SOC 2(服务组织控制报告,一种安全合规认证) |
| structured data | 结构化数据 |
| SurveyMonkey | SurveyMonkey(在线调查平台) |
| TAM | TAM(总可达市场规模,Total Addressable Market) |
| top of funnel | 漏斗顶部 |
| top-down | 自上而下 |
| TripAdvisor | TripAdvisor(旅行点评与预订平台) |
| Udi Milo | Udi Milo(Tinder 增长负责人) |
| UGC | UGC(用户生成内容,User-Generated Content) |
| unstructured data | 非结构化数据 |
| Upwork | Upwork(自由职业者服务平台) |
| Vanta | Vanta(自动化合规平台) |
| Wade Foster | Wade Foster(Zapier 联合创始人) |
| Waldorf Astoria | 华尔道夫酒店 |
| WebMD | WebMD(医疗健康信息网站) |
| Writer | Writer(AI 写作平台) |
| Yuriy Timen | Yuriy Timen(增长顾问) |
| Zillow | Zillow(房地产信息与估值平台) |
Reformatted by reformat_english.py
重新思考 AI 时代的 SEO | Eli Schwartz(SEO 顾问、作者)
文字记录
Lenny Rachitsky: 你注意到随着 AI 回答被整合进搜索结果,SEO 的运作方式发生了重大变化。
Eli Schwartz: 坦白说,我曾以为这将是一场末日。在 AI Overview 出现之前,谁在那篇长篇内容中胜出,谁就能获得第一次点击。但现在这一切不存在了。
Lenny Rachitsky: 人们在这个新范式下应该怎么做才能成功?
Eli Schwartz: 把 SEO 当作一个产品来看待。产品经理才应该是思考这个 SEO 问题的人,因为这是一个产品问题。产品人员需要思考的是:对于那些不会通过社交渠道发现我们、不会被广告吸引的用户,我们该如何向他们定位?这是一个正在进行自主探索之旅的用户。如果你无法回答”用户会在搜索引擎上搜索什么”这个问题,那就不要做 SEO。
Lenny Rachitsky: 对很多人来说,SEO 有点像一种黑魔法。
Eli Schwartz: 它不是黑魔法。它很简单。我认为第一步是几乎所有人都忽略的那个 SEO 环节,那就是……
嘉宾介绍
Lenny Rachitsky: 今天的嘉宾是 Eli Schwartz。Eli 是一位专注于 SEO 的增长顾问,曾帮助 Quora、Coinbase、Tinder、LinkedIn、WordPress 和 Zapier 等公司制定和执行其 SEO 战略。他也是《Product-Led SEO》一书的作者,对如何思考 SEO 并在 SEO 中取胜有着非常新颖的见解。最近,他花了大量时间分析 SEO 随着 LLM 聊天机器人的崛起会发生怎样的变化,分析 Google 如何直接在搜索结果中给出答案,以及如何在 SEO 战略中利用 AI。
在本期节目中,我们深入探讨了在这个新的 AI 范式下取得成功所需了解的一切。正如 Eli 在对话中分享的那样,Google 目前正在推出搜索机制的重大调整,大幅提升了在搜索结果顶部包含 AI 生成回答的搜索比例。因此,变化很快就会在我们脚下发生。如果你正在考虑 SEO、正在做 SEO,或者只是对搜索的演变方向感到好奇,这期节目就是为你准备的。
如果你喜欢这档播客,别忘了在你喜欢的播客应用或 YouTube 上订阅和关注。这是避免错过未来节目的最佳方式,也对播客帮助极大。话不多说,我请到了 Eli Schwartz。Eli,非常感谢你的到来,欢迎来到播客。
Eli Schwartz: 能来到这里真的很荣幸。你邀请过一些非常棒的嘉宾,我很荣幸能被算作其中之一。
Lenny Rachitsky: 完全是我的荣幸。你在 SEO 领域工作了很长时间,帮助公司弄清楚如何在 SEO 中取胜已经超过十年了。我们之前聊过 SEO 领域正在发生什么,你告诉我你注意到随着 LLM 的崛起、随着 AI 回答被整合进搜索结果,SEO 的运作方式发生了重大变化。所以我想到用完整的一期节目来谈谈人们需要了解 SEO 正在发生什么变化,以及在 LLM 和 AI 盛行的新范式下如何取得成功。你觉得怎么样?
Eli Schwartz: 这个主意太棒了。我真的很喜欢 AI 对 SEO 带来的整体影响,因为它迫使所有关心 SEO 流量的人——不管是 PM、CMO 还是 CEO——真正开始重新思考 SEO 流量意味着什么,因为围绕 SEO 的策略其实并没有真正改变。它一直都是同样的事情。
我上一份全职工作是在 SurveyMonkey,当时我还在业余时间做副业。有人把我介绍给一家大公司的 CEO,他问我对 SEO 的方法论。我想拿下这个项目,想促成这次咨询合作。那位 CEO 对我说:“所以你 basically 告诉我的是,我需要找到我的关键词,把它们放进内容里,然后建一些链接。除此之外你还要为我做什么?我凭什么付你钱?”
这让我愣住了,因为这确实就是 SEO,过去也是 SEO。这件事迫使我真正重新思考 SEO 可能是什么。我转变了思路,并和许多公司合作,帮他们重新思考 SEO 以及 SEO 流量应该意味着什么。但其他人并没有这样做,因为那些老策略确实有效。而现在,LLM 和 AI 整体上正在迫使人们再次思考:SEO 应该如何运作?我应该如何从搜索渠道驱动业务?
搜索正在发生什么变化
Lenny Rachitsky: 先打一点基础,聊聊随着 AI 回答和 LLM 的崛起,搜索和 SEO 究竟发生了什么变化。
Eli Schwartz: 本质上,Google 和其他科技公司是被 OpenAI 和 ChatGPT 逼着行动的。Google 声称发明了 LLM 的概念,他们也许确实发明了,也许没有,而且 OpenAI 的一些早期员工就是从 Google 出来的。但 ChatGPT 在 2022 年底横空出世,具备了提问任何问题并获得一个完整书面回答的能力。突然之间人们觉得,我不再需要 Google 了,不需要去点击那些搜索结果了。
于是,“不再需要 Google”的讨论开始兴起。Google 要完蛋了。不仅如此,你也不再需要 SEO 了。你不需要去优化任何东西,因为全世界的信息都会直接呈现在你面前。这一点我们后面再深入讨论。我认为这完全是不对的,而且我认为无论你是全职做 SEO 的人,还是作为主要营销渠道之一在接收 SEO 流量的人,都不需要为此担心。
Google 的焦虑与应对
然而,Google 确实紧张了。而我认为 Google 最重要的利益相关者其实是华尔街。如果华尔街突然认为 Google 是一家过气的公司,觉得它不再有趣,不愿意继续持有对它的投资,那它的股价就会下跌。这会影响 Google 招募人才的能力,也会影响 Google 筹集资金、投资各种有趣项目的能力。
所以 Google 必须向整个市场和投资者证明,“那个 OpenAI 做的东西,我们也能做。没什么大不了的。“于是 Google 做出了回应——当然,一开始搞砸了——推出了当时叫 Bard 的产品。他们说,“ChatGPT,看,我们也有自己的版本,“然后做了一次公开演示。效果非常糟糕,股价反而跌了。
后来他们修好了,股价又涨回去了。但他们还必须回应另一个质疑:搜索要死了吗?当全世界的信息都可以直接呈现在你面前时,还有谁需要搜索?于是他们推出了当时叫做 SGE(Search Generative Experience)的产品,本质上就是在搜索结果里嵌入 ChatGPT。他们推出了这个功能,但只是以 Beta 版的形式。
AI Overviews 的推出与风波
他们在这件事上面临着巨大的挑战,首先是商业化问题。他们所做的一切都依赖广告,大部分收入都来自广告。所以如果你要在搜索结果里展示 AI 回答,那就不能同时放广告,你必须二选一。这是他们至今仍未能解决的问题。第二个问题是法律责任。
如果他们的生成式回答让你去做一些可怕的事情——比如有一条,可能是真的也可能是伪造的,很多人伪造了不少——告诉你去从金门大桥上跳下去。这会不会有法律责任?因为此时 Google 不再是搜索引擎,告诉你去哪里找答案,它自己成了发布者,是它亲口告诉你去做那件事的。
第三个问题,同样涉及法律责任,就是抄袭。这些都是他们担忧的问题。所以他们花了大约一年时间来测试这个东西,这其实很有意思,因为我们是做技术的,大多数听众也是做技术的。Google 通常是一家快速上线的公司,他们发布一个东西,然后很快就会大规模铺开。
他们不是那种说要发布一个东西然后花一年时间才上线的那种公司。AI Overviews,也就是当时叫 Search Generative Experience 的产品,是在去年五月的 Google I/O 上正式发布的,并更名为 AI Overviews。它就是这个东西,本质上就是搜索结果里的 ChatGPT。发布时声势浩大。
当然,它立刻就上线了。他们说只面向美国地区的已登录用户。很显然,他们对在欧洲发布产品一直很谨慎,因为欧洲比美国的诉讼风险更高。上线之后,突然就开始出现各种截图——Google 让我去跳金门大桥,Google 让我在披萨上涂胶水。当然,Google 并没有真的这样做。
它只是从网上抓取了各种内容。我很惊讶 Google 没有预料到这种情况,因为 ChatGPT 上发生了完全一样的事情,而且 Google 是一个更大、更有话题性的靶子,人们更愿意在社交媒体上分享这些东西。Google 上线之后有点尴尬,于是做了一定程度的回撤。
但更有意思的是,他们放弃了这次回撤。或者说不是放弃,而是重新推进,再次上线了,现在它出现在了更多的搜索结果中。我在大多数搜索结果中都能看到它。而且,未登录的用户也能看到了,比如隐身模式的用户或者与 Google 没有任何历史记录的用户都能看到。他们还在英国上线了。所以这个东西正在到来,而且真的会影响搜索结果。
对 SEO 的影响
Lenny Rachitsky: 为了帮助大家理解为什么这会影响 SEO,如果这不是那么显而易见的话——就是它把搜索结果往下推了。用户直接在那里就得到了答案,不需要点击你的链接了。而且页面顶部还有赞助链接,已经在把你往下挤了。所以基本上,你的内容越来越难被看到了,对吧?
Eli Schwartz: 其实不是更难被看到,而是你的内容变得不那么相关了。而这正是我对 SEO 感到兴奋的地方。SEO 过去就像那位 CEO 多年前对我说的那样,就是创造一些内容,然后拼命争夺那个热门关键词的排名。多年前,我在一家汽车领域的创业公司工作,我们写关于汽车的内容。
我们想要排名的一个词就是”cars”。主页就是为”cars”这个词排名的。我们买了大量链接,我们就是一个关于 cars 的网站。显然现在没人能指望这样做了,因为人们不再那样搜索了。你不会想,“哦,我要买一辆车,我去搜’cars’看看会出现什么。“因为 cars 可以指很多东西。Cars 是一部电影。Cars 是一种你开的车。Cars 可以是卡丁车,可以有很多意思。
所以没人会去看那些搜索结果。然而,SEO 仍然在追逐那些漏斗顶端的、人们最关心的大关键词。而现在这些关键词中的很大一部分将会转移到 AI Overviews 里。我不想只谈 Google,不过我认为 Google 将永远——至少在很长很长一段时间内——拥有整个搜索页面。
但其他搜索引擎,不管是 Perplexity 还是 Claude 还是 Meta,都有机会提供类似 AI Overviews 的回答。我认为那些漏斗顶端类型的查询,非常适合由一段话来给出答案。比如你想去度假,想要一个海滩度假,你可以非常明确地问:给我推荐一个不在美国、但距离 X 机场两小时飞行距离的海滩度假方案。
这种事情你在 Google 搜索里也能做,但会花很长时间。而用一段话直接得到答案就非常方便。然后你进入漏斗中段。你做了这个查询,不管是 Google 还是其他什么回答,建议你去坎昆。现在你进入了漏斗中段,而这正是 SEO 开始发挥作用的地方。
搜索旅程的变化
所以我认为这次颠覆之所以如此重大,是因为用户旅程变了,发现的方式变了。以前,如果你是一个旅游网站,你能在”距离美国两小时航程的最佳海滩度假”上获得排名,那就是你的排名结果。然后你有你的长篇内容,有你的广告,你可以借此变现。
这种情况下,即便你仍然排名第一,但你的位置已经跌到了页面最底部。AI 的回答——不管来自谁,不管是 ChatGPT 还是 Google——会告诉你接下来该从哪里开始深入搜索。
Lenny Rachitsky: 所以我想复述一下你的意思:搜索中的发现环节将被 LLM 吞噬,由它们为你指明方向。而一旦你明确了自己想要什么,你再回到 Google,这时候机会才可能继续存在。
Eli Schwartz: 完全正确,而且我认为这正是好的地方。这正是我对搜索用户体验感到兴奋的原因,因为我觉得用户并没有被最佳地服务——比如 U.S. News 或 Forbes 雇一个并非旅游专家的自由撰稿人,写一篇”距离美国两小时航程的最佳海滩度假”之类的内容。
现在你依然会从非旅游专家那里获取信息——你会得到一个 AI 摘要式的回答,但它会给你更多线索和灵感去做更深入的研究。我认为这才是对用户最好的服务方式。
SEO 的实际冲击
Lenny Rachitsky: 这太有意思了。我们一定要在这方面多花些时间,让大家准确理解这意味着什么。在深入讨论以及人们应该如何应对之前,你观察到 SEO、搜索结果以及整个 SEO 领域随着这些东西的推出受到了什么影响?你有没有看到具体的数据——比如什么在下降,什么在增长?
Eli Schwartz: 坦白说,我原本以为这将是一场末日。大约一年前我就说过这会是一场末日。但目前我们还没有看到,很多人因此宣布胜利,说末日没有来。然而,这个东西才刚刚以一种我认为即将开始影响流量的方式推出。
就在几周之前,它还不对未登录用户开放,只有登录用户才能使用。而且不仅如此,即使在登录状态下,Google 也因为五六月份刚推出时出现的那些令人尴尬的事件而大幅回撤了。他们现在才开始更广泛地推出。我确实认为我们将会看到影响。
当然,准确定位影响的另一个挑战是,他们还在持续推出算法更新——这个我们后面也会谈到——这会掩盖正在发生的事情。因为如果你之前被算法更新打击了,突然又从算法更新中恢复了,你会看到流量增长,但你可能是在一个更低的基数上增长的,因为 AI Overviews 正在改变局面。
所以我觉得,如果退后一步,从用户旅程的角度来看,它不可能不影响搜索。只是这种影响很难被察觉。在某些场景下它会非常明显,而在另一些场景下则完全不会产生影响。而分界线就是用户旅程——就是用户处于漏斗的哪个阶段。
不同类型网站受到的影响
比如,如果你是 WebMD,你在写关于人体的内容——就是那种自从医学期刊诞生以来就存在的通用内容——AI Overviews 会非常擅长直接给你答案。你头疼,心想:“这头疼老不好,我是该多吃点 Advil 还是该睡一觉?”
然后你找到那篇 WebMD 的文章,翻到第六页,它告诉你 0.2% 的头疼患者实际上患有脑肿瘤。你不再需要这个了,因为 Google 可以直接告诉你没事,去睡一觉、多喝点水就好了。
WebMD 在和 Healthline 以及其他——Cleveland Clinic、Dallas Clinic 等各种医院竞争,而完全没有必要去看所有这些搜索结果、翻阅六页关于头疼的信息。所以这类网站会受到 AI Overviews 的冲击。而电商网站则未必,取决于用户处于漏斗的哪个阶段。
Lenny Rachitsky: 你这么一说,我想起了自己的亲身经历,之前我都忘了。我曾经有一个网站叫 whenishanukkah.com,因为光明节每年日期不同,而这个网站就只告诉你日期。整个网站就这些——告诉你今年光明节是哪天。我在上面放了广告,每年赚十块钱。后来 Google 直接在搜索结果里给你答案了。所以我亲身经历过这种事。
结构化数据与非结构化数据
Eli Schwartz: 甚至还不止于此。你提到的这个是结构化数据。Google 很容易就能告诉你 150 年后光明节是哪天——它在一个数据集里。而现在发生的事情是,Google 正在把内容中的非结构化数据转化为结构化数据。比如你可以问:婴儿出现某类症状需要去医院的概率是多少。
同样地,与其阅读所有那些内容然后自己做判断,Google——不只是 Google,ChatGPT 或 Claude 也可以——可以提取所有那些非结构化数据,根据它所阅读的一切给你一个统计数据。这对用户非常有帮助。我认为,用户是受益的。
然后用户可能会发现还有另一条信息——比如我其实想读一篇医学论文,或者现在我想 Google 一下找离我最近的、营业时间符合要求的医生。这就是 Google 搜索,不是 LLM AI 搜索。
Lenny Rachitsky: 顺便说明一下,我不是在说那个例子跟 AI 有关。那是十年前的事了。所以我完全同意。
Eli Schwartz: 我只是觉得那个例子很有意思,值得深入讲讲结构化和非结构化之间的区别,因为非结构化才是 Google 颠覆一切的地方。SEO 这整套理念,直到 2022 年之前,本质上是将非结构化数据变现。谁写了最长的”迈阿密最佳海滩酒店”内容,然后给它建了最好的外链,谁就赢了。
无论通过广告还是通过酒店预订,他们都能赢。而现在,Google 可以直接把你带到那个环节的起点,说”这些是根据所有写过内容的人评出的最佳海滩酒店”,或者实际上利用评论中 Google 自己拥有的非结构化数据来给你答案。然后你可以说,“好,我想去这家酒店”或者”我想住在这个城市”。
漏斗顶端与漏斗中段的区别
Lenny Rachitsky: 那请再帮我们更深入地理解漏斗顶端和漏斗中段这个区分。当人们可能在其中一端获胜时,具体是什么样的?能不能举些例子——比如这是一个漏斗顶端类型的搜索,Google 会把它吃掉;而这是一个漏斗中段的体验,你可以赢?
Eli Schwartz: 总体来说,SEO 一直以来更多集中在漏斗顶端,因为你对某个事物感到好奇。比如你在找新软件,你在找新的播客软件。你搜索”top podcast tools”,然后从 G2 得到一个列表。G2 以及所有类似 G2 的网站——比如 Gartner 旗下的 Capterra——都会被大规模颠覆。你会从 G2 得到一个列表,里面列出所有软件。
你会逐一查看,上面写着”这款面向企业,那款面向小型播客创作者,还有一款免费”,然后你把选择范围缩小到三个工具。这时你才开始做那些具体搜索——现在你进入了漏斗中段。假设你选定了 Riverside,并且已经从前面的搜索中获取了足够信息,现在你开始搜索 Riverside 价格、Riverside 容量、Riverside 带宽。
这就是漏斗底端,也是你最终去购买 Riverside 的地方。而在顶端,你搜索的是”top podcast tools”。所以说,在 AI 概览出现之前,在 LLM 这一整套概念出现之前,谁在那篇长篇内容上赢了,谁就能获得第一个点击。但现在这一切不存在了。你直接打开 Google,Google 会告诉你:“这些就是工具。哦,你在找新 CRM?CRM 是这样的。你想知道你是需要一个 CRM,还是只需要一个日历?”
看一眼,Google 直接用一段话告诉你答案,然后你的搜索就被引导到漏斗中段的其他地方了。所以我认为 SEO 本来就应该聚焦在那里,因为那才是潜在转化可能发生的地方。然而,SEO 之前从未真正涉足那个位置,因为大多数 SEO 衡量绩效和成功的方式是排名。他们会说,“好吧,我们在 CRM 这个词上能不能转化并不重要。但你看,我们排第一,所以我们赢了。”
Lenny Rachitsky: 首先,作为一个用户,这听起来太棒了。我真的厌倦了所有搜索结果都只是一堆 SEO 页面,里面全是些废话回答。所以我真的很希望 Google 直接告诉我,直接告诉我我需要知道什么。另一方面,如果你认真想想 Google 一直在努力做什么,他们一直在做的事情就是——这是我们对你这个问题的最佳回答,这里是一些指向答案的链接。
而这只是它的一个更好版本,把所有实际信息汇总起来,直接给你最终结果。所以他们做这件事完全说得通,而且技术终于让他们有能力做到了。好,那我们聊聊那个——不知道该说六万四千美元的问题还是百万美元的问题——人们应该怎么做?在这个正如你所描述的、正在发生变化的新范式中,人们应该怎么做才能成功?
这一切才刚刚开始发生。而且可能让很多人措手不及,因为他们以为 AI 概览已经推出了,情况还行,而你说的其实是它现在才开始真正加速。
SEO 是产品问题,不是营销问题
Eli Schwartz: 我觉得这很迷人,也很荣幸能成为你播客上为数不多的营销人之一,但我是把 SEO 当作一个产品来看的。我认为产品经理才是应该思考 SEO 问题的人,因为这是一个产品问题。有用户从搜索渠道进来,我需要为他们打造什么样的产品?
我需要为他们创造什么样的体验?过去通常的做法是,把它看作一个营销挑战——产品团队已经做出了这个东西,现在期望营销人员去为它做 SEO。但这里存在一个错位。比如,我经常和 SaaS 公司交流时发现,我认为 SaaS 整体上不应该做 SEO。
但很多时候,当我跟 SaaS 公司交谈时,他们已经为某个用户群打造了产品,然后期望营销人员把这个产品硬塞进搜索结果所围绕的需求里。这行不通,因为那不是用户在找的东西。根本没人问过用户到底在找什么。现在 SEO 正在发生变化,你确实需要思考漏斗中段,需要思考用户在搜索时的体验和购买体验,我觉得这一切开始整合到一起了。
这就是产品人员现在需要思考的问题——如何把这个产品呈现给那些不会通过社交渠道发现它、不会被广告吸引、不会在展会上发现的用户。这是一个在进行自主探索之旅的用户,他们在找什么,我们如何以他们能找到的方式来呈现这个产品。
所以这就是每个人都应该做的——产品真正与营销协作,发现用户想要什么,然后把它展示出来。我最近跟一家健康领域的公司聊过。他们有一个健康 App。我们谈了他们的营销。他们的 SEO 策略基本上就是写成千上万篇博客文章。AI 让他们做了他们根本不该做的事情。
他们写了成千上万篇关于健康的通用博客文章。我当时告诉他们不应该这样做,但他们还是让我体验了他们的产品,给了我一个下载 App 的兑换码,结果这个 App 体验很差,是个不好的产品。他们试图用不适合的营销去包装一个做不到它所宣称功能的产品。如果他们先把产品做好,理解用户需求,那问题就变成了——如果我是寻找这类产品的用户,他们在找什么?我们如何在漏斗中段把这个展示出来?
Lenny Rachitsky: 能不能帮我们再具体化一点?能不能走一遍你做过的产品案例,或者一个在这方面做得特别好的案例——从”嘿,他们想要我们的产品”到整个用户旅程,好的案例是什么样的?
Zapier 的 SEO 实战案例
Eli Schwartz: 我最早发现这个过程的例子大约是十年前,我认识了 Zapier 的 CEO Wade Foster,他是我的前同事。我投资了这家公司,他约我见面讨论 SEO。他们做的东西当时没人需要。他们有一个把各种工具”连接”起来的产品,但没人需要它,因为没人知道它存在。但人们知道自己需要各种工具之间能互相协作。
他们只是在找 Zapier 时找不到。所以在和他们的讨论中,在我们设计的 SEO 实验中,我们想到——人们在找 Gmail,也在找 Salesforce。他们知道 Gmail 不能连接 Salesforce,知道 Salesforce 不能连接 Gmail,但他们在找把两者对接起来的方法。
那如果我们作为一种产品,用营销的方式展示”把 Gmail 连接到 Salesforce、把 Salesforce 连接到 Gmail”这个功能确实存在呢?我们就是这么做的。我们把这个模式大规模复制到所有能连接的组合上,这就是我跟他们做的实验——所有可能的连接都展示出来,由此创造了一个飞轮效应——“哇,Gmail 能跟 Salesforce 搭配,那 Gmail 还能跟什么搭配?”
能跟我拥有的其他工具配合吗?这又是一个早期案例,让我偶然发现了这个理念——人们在找的是那个”其他东西”,而你能做的事情就是展示你的产品能够实现这件事。
Lenny Rachitsky: 我完全不知道你参与过 Zapier 的 SEO 工作。那可是 SEO 领域最传奇的成功案例之一。太酷了。
Eli Schwartz: 完全是意外。每当有人问我关于 programmatic SEO 的事——我知道我们后面得深入聊聊 programmatic SEO 是什么——他们说”我想模仿 Zapier”,这时候提到我曾经帮过忙就特别好。
Tinder 的 SEO 案例
Lenny Rachitsky: 太棒了。那这个案例的启示是——你发现关键的机会在于教会人们用这款产品能实现什么。产品本身很厉害,能满足真正的需求,用户会去搜索——比如”Gmail 和 Salesforce”——而你们要做的,就是帮他们看到这里有一个很有意思的解决方案。
Eli Schwartz: 对。我其实还有一个更好的例子。
Lenny Rachitsky: 好,说来听听。
Eli Schwartz: 我曾和 Tinder 合作过几年。在 SEO 领域,尤其是在做顾问的时候,我遇到的最大挑战之一就是让事情真正落地。我合作过的公司都很优秀,团队里也有很棒的人,但他们总会撞上一堵墙。在我的书里,我讲过一个早期的故事——我在一家公司工作,是被 CMO 聘请的,我们制定了一份很好的方案。然后我们去找 CEO,CEO 又叫来了 CTO 一起讨论我们的方案。CTO 说:“我没有工程师可以做这个。所以你自己决定,是要做产品,还是做你搞的这个营销的东西。“于是我们什么也没做,因为他们从来没有给过资源。
但我后来在 Tinder 遇到了一位非常出色的产品负责人,Udi Milo,他是增长负责人。Tinder 以前从未做过 SEO。Tinder 所有的 inbound 流量都来自”Tinder”这个词本身。他觉得,既然从来没做过 SEO,那一定有上升空间。他对此深信不疑,也愿意推动这件事向前走。所以当我们开始合作时,首先要解决的问题是:用户搜索什么的时候会找到 Tinder?答案不是”线上约会”——那只是一个单独的词。我们不可能围绕约会相关的所有话题写长篇内容,因为那不是 Tinder 的产品。用户不会先读一篇关于如何坠入爱河的文章,然后就莫名其妙地转化成 Tinder 用户。
从用户需求出发的 SEO 思路
但在用户调研和讨论中,在探讨我们可以在 SEO 上做什么投资的时候,我们发现 Tinder 解决的是一个孤独感的问题——Tinder 是一个解决孤独感的方案。你感到孤独,你搬到了一座新城市,谁都不认识,你想解决孤独的问题。既然你身处一座新城市,我们意识到这是一个本地化的需求。所以我们要找的就是所有与”本地”相关的搜索。于是我们做了这样的页面:如果你在世界各地很多城市搜索”线上约会”——这个功能是 beta 版,一直保持在 beta 状态——你会找到一个 Tinder 的页面,上面列举了一些适合约会的地方。不仅如此,它还把 Tinder 作为解决你孤独问题的方案呈现给你。
Lenny Rachitsky: 那 AI 概览的出现对这些有什么影响吗?这是不是就是——这就是 SEO 未来要走的路,而不是单纯堆关键词和发无尽的博客文章?
Eli Schwartz: 实际上 AI 概览对这些没有任何影响。因为假设你去了迪拜,你刚到迪拜,一个全新的国家,你从没去过,你感到孤独。于是你搜索”迪拜线上约会”。AI 概览可能会告诉你那里的约会场景如何,可能会推荐一些好的约会地点。但它并不能帮你解决孤独的问题。你的孤独感是一个漏斗中部或底部的问题。无论 AI 概览做了什么,你还是会点击 Tinder 的结果。所以我认为 SEO 应该做的,就是让你通过 SEO 进入用户的购买旅程。AI 概览做或不做什么都不重要,因为你用 SEO 方案解决的始终是漏斗中部的问题。
SEO 实操三步法
Lenny Rachitsky: 我们换个方向聊聊吧——我本来打算稍后再谈的,但现在可能正好合适。假设有个人坐在办公桌前想:“我想开始做 SEO,想在 SEO 上取得成功,但之前没怎么做过。“在 AI 和大语言模型的这个新世界里,想要开始涉足 SEO,第一步、第二步、第三步分别应该做什么?
Eli Schwartz: 我认为第一步是几乎所有人在做 SEO 时都会忽略的一步——成为你的用户,尝试理解你的用户是谁。这样的公司太多了,SurveyMonkey 就是一个很好的例子。我在 SurveyMonkey 待了七年,每当有人入职,公司会给你一个 SurveyMonkey 账号,告诉你说你应该去发一份问卷。后来的几年里,他们要求每个人在入职培训期间用问卷在公司内部做一次调研,但在那之前,没人这样做过。所以大家有了账号,却从不发问卷,对人们为什么使用这个工具毫无共情。
然后,当你从产品和营销的角度去思考时,你把它当作一个工作任务来对待,而不是一个需要用户共情的挑战。所以无论你想对什么做 SEO,任何人真正应该做的第一件事就是尝试成为那个客户。如果我是某个 SaaS 的用户——我在做一个 SaaS 产品,想要推广这个工具——那用户会搜索什么?他们会带着什么问题去搜索?我之前提到过,我认为大多数 SaaS 工具不应该做 SEO。原因就是,很多时候我和 SaaS 公司谈 SEO 时,我问他们这个问题,他们一脸茫然。
如果你回答不了”用户会搜索什么”这个问题,那就不要做 SEO,因为 SEO 的核心就是触达那个用户。如果你能理解用户在寻找你产品时会做什么,那就是 SEO 的第一步。这个用户是谁?我在向谁营销?你在脑海中构建出这个用户画像。第二步是思考你要创建什么资产。回到刚才讲的 Tinder 的例子——我们理解了用户是在世界任何地方解决孤独感的人,因为 Tinder 做的就是这件事,它以你想要的任何方式解决孤独感的问题。
然后我们需要思考的是要创建什么。我们知道它应该是全球化的,当然。我们希望它是程序化的——这涉及到程序化和编辑型内容之间的区别,值得深入聊聊。我们确定要程序化,因为没有人愿意手写世界上每个城镇、每个城市、每个街区的页面。我们希望它是程序化的。那我们需要往里面放哪些东西?第三步就是真正去构建。我们在一档产品播客上,对产品人来说,你就是在为那个 SEO 用户构建产品。你的输入从哪里来?这个页面需要长什么样?
所以,我认为 SEO 应该归属于产品范畴,原因就在于——SEO 所需的投入,并不是很多人以为的那种”基于选定的关键词、针对 Google 优化的一篇内容”。我把它看作一个产品,这意味着你需要设计资源,需要工程资源,当然还需要产品经理来真正把控构建过程。你需要做用户调研。所以它远不止是一篇内容。总结一下这几个步骤:第一步,理解你的用户是谁;第二步,决定你需要为这个用户创建什么;第三步,把这个东西当作产品来规划和构建。
SEO 不适用的场景
Lenny Rachitsky: 我认为这里有一个很有力量的观点,可能……大家未必能意识到,那就是——用户必须在想某个问题时,会去 Google 搜索,然后才能找到你的产品。所以,如果没有任何人以某种方式搜索你正在构建的东西,那你就没有机会赢得 SEO,也不可能从 SEO 中获益。你脑海中有没有什么 B2B SaaS 公司的例子,就是那种”这里没有 SEO 机会,根本没人搜索这个东西”的情况?
Eli Schwartz: 大多数都是。我离开 SurveyMonkey 后最早接的咨询客户之一就是 Mixpanel。我刚开始做独立咨询,什么客户都接。确实接了一些糟糕的客户,但 Mixpanel 不算糟糕——它只是让我大开眼界。我们在尝试做 SEO,我完全按照 CEO 告诉我的方式去做,他原本打算自己来做。所以我给他们关键词,我们一起策划内容,我们建设外链,我们做了 SEO 应该做的所有事情,但没有效果。我们就坐在那里,我让他们给我展示用户旅程。他们有工具可以做这件事——这正是 Mixpanel 的产品。
有人点击了搜索结果,落在了我们创建的内容页面上——我们做了调研,知道用户在找什么——但为什么不转化呢?然后我意识到转化方面还有其他问题:Mixpanel 是一个需要整合到整个公司的产品,你不是随便 Google 一下就觉得”哦,分析工具,行,我告诉大家咱们上吧,明天就能上线”。而且 Mixpanel 价格不低,所以它天然存在摩擦。你不会从搜索结果点进去就直接决定购买。你甚至没法用自己的信用卡直接购买,我记得。
这些就是问题所在,也是为什么我认为 SaaS 不是 SEO 的最佳适用场景——因为想想那个旅程,问题本身不一定已经被意识到。SaaS 确实解决了问题,但前提是你知道那个问题存在。而让人们意识到问题存在,通常不是 SEO 能解决的,它通常是品牌层面的挑战。你做一个病毒式传播的视频,告诉人们”你可能没意识到我们创造的这个工具能解决你的问题”——但没有人会去搜索一个他们不知道自己有的问题,也不会搜索一个他们不知道可能存在的解决方案。这就是断裂的地方。另一个问题当然就是,即便他们知道了问题存在,这也不是一个 SEO 驱动的旅程。
我在 COVID 初期还有一个开眼界的经历。当时大家都待在家里,没什么事可做。Google 找我,说 Google Cloud 团队有一个 SEO 职位。我根本不可能去,但经历一下面试流程、待在家里也挺好。所以整个面试过程非常有趣——每次面试开始,不管是招聘经理还是团队成员,他们都会问”你有什么问题想问我吗?“然后接下来的 45 分钟全是我问他们:“你们为什么要做 SEO?你们想做什么类型的关键词?这个 SEO 旅程是什么样的?“因为 Google Cloud 只有 Amazon 和 Microsoft 两个竞争对手,没有人会搜索一个词,看到 Google 排第一,就直接说”好的我买了”。这里有一个决策流程,有委员会决策。所以 SEO 作为营销渠道根本说不通。宏观来看,Google Cloud 不应该做 SEO。微观来看,很多 SaaS 工具也不应该做 SEO,因为根本不存在 SEO 旅程。我跟公司说这些时经常遭到反驳,他们会说”看看我们做的 SEO 成果”,但那些东西很快就到天花板了。
Lenny Rachitsky: 这真的很有意思,我试着把你说的重新梳理一下,因为这里面有几个层面。不一定说人们不会在 Google 上搜索那个问题。我在想 Vanta 做 SOC 2 的例子——不是说人们不会搜索 SOC 2 认证……好吧,他们确实不会搜索。
我听到更大的问题是:用户永远不会在那个搜索旅程中直接购买你的产品。也许这能起到一点教育作用,让他们知道”哦,Vanta 存在”,但他们不会因此成为客户。他们需要找销售人员谈,需要牵涉到一堆利益相关者。所以这本质上是一个销售驱动的动作,不是产品驱动型 SEO 的动作。对吗?
Eli Schwartz: 对。
Lenny Rachitsky: 顺着这个思路再往下拉一点——甚至超出 B2B 的范畴——一个人怎么判断 SEO 到底是不是自己的机会?以及,应该投入多少资源去探索这个机会?
SEO 并不免费
Eli Schwartz: 我们必须彻底抛弃”SEO 是免费的”这个迷思,因为它绝对不免费。时间上有成本,资源上有成本,当然还有 SEO 的直接支出。所以如果你在考虑是否应该做 SEO,就需要进行评估。这也是我把它当作产品来看待的原因——你有很多产品想法,要决定把时间和钱花在哪里。SEO 只是你可能投资、也可能不投资的一个渠道。
比如说有一个 SaaS 工具,他们确信存在 SEO 旅程,人们确实会搜索这个解决方案。我在 SurveyMonkey 的时候,我们每年通过自然流量创造了几亿美元的收入,因为它是一个免费增值模式——你搜索问题,找到解决方案,免费注册,如果适合你,你最终会付费,这就是我们从自然流量中产生的收入。假设确实存在这样的旅程,人们会搜索,投资 SEO 是合理的——接下来就是决定投入多少、以及怎么投入。
拿 SurveyMonkey 来说,主要是围绕创建内容、围绕模板展开的,花费并不高。但假设有一家公司,既没有内容团队,也没有能让人们搜索到的产品,也没有产品经理来监督这个 SEO 流程。那么你需要招聘这个 PM,或者找一家代理商——一家典型的 SEO 代理商,你不可能花 500 美元就能请到,起步至少 10,000 美元一个月,因为那就是人力时间的成本。
一个月 10,000 美元,一年 120,000 美元,如果招全职员工会更贵。这是一笔你要列出来的开支。然后还要加上所有配套资源:你需要 CMS,这要花钱;你需要工程师来支持,这也花钱;你可能还需要设计师,需要内容。这些加起来真的很快。然后你看着这笔投入,问自己:“对于这个工具,如果我每年在 SEO 上投入 100 万美元,我预期能在短期内收回这 100 万美元吗?”
Eli Schwartz: SEO 如果方向对,总会回本,但关键是要”快”。SaaS 工具,尤其是创业公司,需要尽快回本。如果把同样的一百万美元投入到品牌广告、网红营销,或者直接投到 Meta 和 Google 上的传统付费广告,你会不会更快地赚回这一百万,甚至一百万零一美元?
我认为评估应该在这个层面进行,而不是默认地想”我刚拿到融资,需要投资 SEO,因为 SEO 是免费的,而且大家都在做,你看,竞争对手也在做。“它应该是一个深思熟虑的战略决策过程——总共要花多少钱,这是不是对资金的正确使用。
Lenny Rachitsky: 总结得很好。我认为这一点非常重要。不是说 SEO 不会带来收益,不是说 SEO 不是一个机会,而是你在其他领域有更大的机会。大多数情况下,如果你是 B2B SaaS 公司,因为购买旅程并不完全在线上完成,用户不会仅仅因为读了一堆页面就转化。
Eli Schwartz: 完全同意。
Lenny Rachitsky: 太有意思了。
Eli Schwartz: 是的。我曾经遇到过一家公司,一个 SaaS 工具,做的是园艺领域。他们做了一个面向园艺师的 SaaS,并且坚持要做 SEO。他们让我看一份从代理商拿到的方案,每月 15,000 美元,全部是内容。然后我问他们,用户是怎么找到他们的?他们那些付费客户都是怎么来的?
他们说去参加全国各地的园艺展会,设一个展位,每个展位成本 10,000 美元。所以我说:“与其花 15,000 美元做 SEO,你可以用同样的预算去参加所有这些展会,获得有真实兴趣的用户——他们有需求、在现场,在展位上试用你的工具,离开后成为你可以跟进的线索。而不是把钱投给我,抱着’希望有效’的心态。看起来好像是免费的,确实有人搜索,但不是正确的搜索。”
Lenny Rachitsky: 这里的启示是,如果你在考虑 SEO 是否值得投资,看到其他公司都在做 SEO、靠 SEO 取得成功,先想想这实际上要花多少钱。我认为这里最重要的启示是——如果你认为不会在线上转化,如果你认为销售才是整个流程的核心环节,那 SEO 对你来说大概率不是一个好的投资回报。
Eli Schwartz: 完全正确。真的要想清楚权衡是什么。再说一次,从产品的角度来看,总是有权衡的。那么投资这个渠道与投资另一个渠道之间的权衡是什么?
四大增长引擎
Lenny Rachitsky: 根据我的观察……我的理解是,有四个真正的增长引擎——四个核心增长引擎:SEO、付费广告、病毒传播和销售。我发现,最终大公司会把它们全做了。除非你是消费品,不需要做销售。所以不是说永远不要做 SEO,我想这里的核心观点是:在早期阶段,资源有限的时候,SEO 大概率不是你时间最好的用途。
Eli Schwartz: 我认为确实有些公司可能永远不应该做 SEO。回到我之前说 Google Cloud 的例子,我认为 Google 永远无法准确指认某个客户是完全通过 SEO 获得的,即使你做各种加权归因分析。也许那个人确实是通过某篇内容或某个 SEO 资产发现 Google Cloud 的,但如果没有他们接触到的所有其他触点,他们永远不会转化。
Lenny Rachitsky: 而且你并不是在说”不要出现在互联网上,不要让人们轻松了解你是谁,不要做一个好的网站,不要有其他页面写关于你的内容”。只是不需要花时间去为 Google Cloud 优化搜索结果,你不会从这项工作中获得显著收益。
什么样的公司不需要做 SEO
Eli Schwartz: 是的,我的意思是,再深入看一下这个用户旅程。我认为餐厅通常不需要有网站——不仅不应该做 SEO,我甚至觉得它们不需要有网站。
Lenny Rachitsky: 热辣观点。
Eli Schwartz: 因为如果你想想一个人找食物时的购买旅程,他们通常不会去访问网站。比如你想吃披萨——再说一次,很多披萨店确实有网站——但如果你在找披萨,你会去 Google Maps、DoorDash、Uber Eats。你不会在 Google 上搜”我附近的披萨”,然后浏览各个网站再决定去哪里吃午饭。
当然,如果你在考虑怎么给孩子办生日派对的餐饮,那可能不一样,但你不会这样浏览并说”这些是我现在要去的地方”。更重要的是,这很花钱,而那家披萨店根本无从知道他们那个配着轻柔音乐、用 Flash 动画弹出菜单的网站——这本身就很贵——到底有没有任何效果。他们的大部分订单还是会通过那些其他平台来的。
Lenny Rachitsky: 是的,我讨厌餐厅网站。我曾经读到过一个解释,说明为什么餐厅网站都那么糟糕——餐厅老板很在意有人走进他们餐厅时的体验和氛围,然后他们把这种思路用在了网站上。他们会说:这里有音乐,有图像,有动画。但在网站上没人想要这些,给我营业时间、地址和菜单就行了。
你提到了几点我想深入聊聊。第一,SEO 大概需要多久才能看到效果。第二,对于好的转化率应该有什么样的预期,什么信号能告诉你”这确实是一条适合我们的购买旅程,SEO 可能会奏效”?就这两个问题。
SEO 多久能看到效果
Eli Schwartz: 这些答案因情况而异。我很讨厌用”看情况”这个词,因为我觉得每当顾问说”看情况”的时候,他们就是在两手一摊说”我真的不知道,我对此没有看法”。就像你去看医生,医生说”嗯,看情况”——你可能快死了,也可能只是需要睡一觉。所以这里确实需要定制化的答案,取决于这家公司是什么样的、预期是什么。
需要多长时间,取决于你在构建什么。再拿 Tinder 来说,我们花了一段时间才真正开始构建任何东西。从构思、构建到看到结果,花了我们很长时间。但一旦构建完成——我很喜欢和知名品牌的大公司合作,因为我们构建的内容一上线,就像推动一艘巨轮,开始驱动收入,变得非常有效,因为它就在那里。之前不存在的东西现在存在了,就开始带来收入。
小公司的话,可能需要好几个月 Google 才会注意到,可能需要好几个月需求才会出现。拿 Zapier 来说,我觉得他们花了好几年才看到任何效果,因为没有人会搜索它。所以归根结底取决于你在构建什么,以及用户能多快找到并需要那个解决方案。
关于转化指标的期望
Eli Schwartz: 至于对转化的期望是什么,这也确实要看你具体在追求什么。有些公司在做媒体,他们……我的意思是,我还是那句话,我觉得大多数公司根本就不应该写大量内容,除非他们是一家媒体公司。但如果你确实在构建媒体,并且从媒体的角度通过内容变现——也许是获取线索,也许是引导用户点击离开页面转化为线索,又或者是 CPM 广告——那么你的转化就是……你通过 SEO 想做的事情就是获取大量页面浏览量,因为页面浏览量越多,从那个页面点击出去进入其他地方的人就越多。如果你是一个 SaaS 工具,那你的转化指标绝对应该是 MQL(营销合格线索)。
所以……而且我再说一遍,我觉得大多数人在 SEO 上都没有做到这一点。他们用了错误的转化指标,那就是漏斗顶部的排名。“哦,我排名上去了,我的 SEO 很成功,我是这个关键词的第一名。“而不是问:“这对我有什么好处?“我曾经合作过一家公司,一个 SaaS 工具。他们在一个双边市场里运营……在人力资源领域,他们只在 HR 市场的一侧变现,但所有的流量都在 HR 市场的另一侧。所以当我们讨论他们的 SEO 转化问题时,我建议他们删掉市场上错误那一侧的所有内容,因为那些内容根本不会转化。
而且这里有一个常见的误解。人们会认为流量本身能带来好处。Google 看到,“哦,我从搜索获得了这么多流量。它认为我是一个非常好的网站。所以即使这些流量完全没有转化,我也应该有一个博客,因为博客是好的。“事情并不是这样运作的。我的意思是,也许如果你是一个巨大的网站,比如说,获得数以百万计的访问量,那你可以建立某种权威性和在 Google 中的良好体验,然后你可以推出其他东西,而不需要担心建立权威性的问题。但总体来说,把流量引向对你不会转化的东西,不是一个好主意。
所以真正理解你想通过 SEO 流量做什么,那就是你的转化指标。可以是 MQL,可以是页面使用量,可以是金额转化,可以是人们拿起电话或者观看视频,但必须是有某种形式的转化。对于一家创业公司来说,可能意味着……你获得了链接。我的意思是,至少,人们阅读你的内容后决定链接到你、给你社交分享。那可能是你可以放进融资演示文稿里的东西。不管是什么,必须有一个对业务有意义的转化指标。
判断 SEO 是否有效的指标
Lenny Rachitsky: 那么为了给刚开始做这件事的人一些具体可参考的东西——当你来到一家创业公司帮他们做 SEO 时,你通常追踪什么指标来判断这是有效的、我们应该继续投入?你最看重的几个指标是什么?
Eli Schwartz: 首先是要真正理解他们在乎什么。排名也可能是他们在乎的东西,如果其他人因此而在乎的话。比如,如果他们要在融资演示文稿里写”我们是这个工具领域的第一名”,而投资者并不知道那并不会转化,那这也许就是我们应该关心的东西——他们应该在那工具词上排第一。如果是 MQL,那他们就需要人们填写线索表单。
所以真正理解他们想通过 SEO 吸引什么样的人,这就是我们要使用的指标。而且我再说一遍,我不知道为什么 SEO……这种情况发生在 SEO 上但不会发生在其他渠道上。做付费营销的人绝不会说,“我在这个搜索上排第一多少次了”或者”我获得了多少次点击”或者”看我花了多少钱”。他们关注的是”我们花了这些钱,效率是这样的,这是我们从中获得的成果”。SEO 也需要同样的严谨。
Lenny Rachitsky: 所以我听到的是,如果你在 B2B 领域,就追踪通过 SEO 进来的线索;或者如果排名是一个可以展示给投资者的面子指标,也追踪排名。而你是说单独看流量是没有意义的——因为我通常在融资演示文稿和投资者更新中看到的几乎都是 SEO 带来的流量这个指标。
Eli Schwartz: 对,完全是浪费时间。
Lenny Rachitsky: 有意思。
Eli Schwartz: 是的,除非有特殊原因。如果你是一家媒体公司,流量意味着那个数字持续增长,那当然有意义。但如果你是在获取线索,而流量数字在增长,你并没有从 SEO 获得更多线索。你获得的是无价值的流量,对业务没有帮助。
SEO 需要多长时间见效
Lenny Rachitsky: 好的。回到需要多长时间这个问题——也给听众一个判断标准,让他们知道什么时候值得继续、什么时候该转向。假设你在尝试 SEO,你应该给它多长时间来决定它对我们是否有效?还是说”总会有效果的,继续做就好了”?你怎么决定是继续还是放弃?
Eli Schwartz: 当你启动一个 SEO 项目时——我会不断回到这个比喻——它就是一个产品。你在构建一个产品,你在设定里程碑。你说我们第一个月要做的是构思我们要构建什么。第二个月,我们要为工程师写出 PRD(产品需求文档)让他们开始工作。第三个月,他们开始动手,然后上线这个东西。
现在,如果你开始错过所有这些里程碑——这在咨询行业是一个很常见的问题——你错过了所有里程碑,然后有人说,“怎么回事……我们已经合作了六个月,什么成果都没有。“而你可以非常具体地指出所有那些被错过的里程碑。所以 SEO 没有见效,是因为没有上线、没有达成里程碑的结果。
随着你达成这些里程碑,你可以说,“好,我们上线了,我们的预期是上线后的第一个月会有 X 数量的页面被索引。我们达成了还是错过了那个里程碑?“然后你可以说,“SEO 在小规模上是有效的。“我在很多公司见过,它需要很长时间,但最终当你回头看时,你会看到那个曲棍球棍式的增长曲线,但当你身处其中的时候,你不一定能看到。
Lenny Rachitsky: 很好。好的。
Eli Schwartz: 好。让我分享一个见效非常快的例子。这也是一个很有趣的案例,我当时在 SurveyMonkey,有人把我介绍给了 Quora 的 CEO,他对做更多 SEO 很感兴趣。Quora 之前从来没有做过 SEO,或者从产品的角度从来没有有效地做过 SEO。我给他们提了几条建议,三个月内他们的流量翻了四倍。所以这确实取决于是什么在阻碍事情推进。在 Quora 的案例中,他们没有展示任何答案。他们极力要求用户先登录才能看到任何答案。
所以我建议他们展示答案。没错,他们等于是在免费给出答案,但同时他们也在把答案展示给 Google。这是第一点,让他们得以把流量翻了四倍。第二点是,Google 没有任何方式可以在站内导航。所以当你来到 Quora——再说一次,他们后来其实撤销了我的建议,但那是大约十二年前的事了——当你来到 Quora,无论是过去还是现在,你都会看到相关问题。你看到一个问题,下面就有相关问题。这就是一个机器人或人类在站内导航的方式。
如果他们创建一个站点地图——他们以前做过这件事,如果 Quora 的人在听,你们应该再做一次——如果创建一个分类站点地图,你可以说”这些都是关于健康的问题”,而且这个站点地图是 HTML 站点地图,不只是 XML 站点地图,连用户都可以通过它来导航。这是健康类,这是健康第一页、健康第二页,你可以浏览整个网站。那么搜索引擎也能导航整个网站,所有的问题和答案都可以被发现。所以当我给他们提了这个建议之后,几个月内他们的流量就翻了四倍。
Lenny Rachitsky: 这里边信息量太大了,有很多有价值的洞察。有一个反复出现的观点我觉得对大家来说非常重要,就是你一直在强调的——SEO 内容页面应该是一个产品,也就是说,应该帮助用户解决他们在理解领域和潜在问题时的实际需求。
所以如果我理解没错的话,你给出的建议是:单纯大量生成一堆内容的博客文章,机会已经不多了。真正的机会更多在于像 Zapier 和 Canva 的模板、Notion 和 Quora 那样的做法——就是直接给出答案,真正帮用户解决问题,同时顺势展示”我们的产品可以进一步帮你解决这个问题”。是这个意思吗?
SEO 必须嵌入购买旅程
Eli Schwartz: 本质上是要围绕公司想要变现的东西来打造产品。比如 Canva 做模板,是因为他们要让其他人在这些模板的基础上进行升级和订阅来变现。但如果一家公司——假设有一个企业版的 Canva,他们并不从模板本身变现,那么在互联网上放一堆只是好看但免费的内容就没有意义了。所以归根结底要看你的产品是什么,你想让用户使用的是什么。举个例子,我在 SurveyMonkey 也做过模板,就是那个调查问卷产品。
如果有人搜索调查问卷模板,那说明他想做一个调查问卷。如果我们不是一个调查问卷产品,而仅仅从调查模板上变现,那就毫无意义。所以照搬别人的程序化 SEO 方式是没用的。一般来说,为了做程序化而做程序化——只是为了有内容——也不会有什么效果,除非这些程序化内容能把你精确引导到你自己的产品。比如 Zapier,他们做了程序化 SEO,它引导你去做不同产品之间的 zap 连接。Tinder,我们也做了程序化 SEO。
它引导到的是——“哦,你想解决孤独问题,我们展示了在你所在的城市这个问题是可解决的。你要解决这个问题,就下载 Tinder 吧。“我从来没用过 Tinder,只是从营销角度用过——不过 Tinder 会设门槛。如果你想获得高级 Tinder 体验,就得付费。这整个都属于购买旅程的一部分。你做 SEO 所构建的任何东西,都必须是购买旅程的一部分。不管是程序化的、博客文章,还是其他任何形式,如果没有对应的产品旅程,没有用户旅程,它就到此为止了。
Lenny Rachitsky: 这个澄清太精彩了。这又印证了你一直在说的一个观点——如果流量不能转化,单纯追求流量是没价值的。所以你可以看到 Canva、Notion、Airtable 有那么多模板,然后去照搬。但如果那不是人们会向你付费购买并让你变现的东西,那就不值得做。
Eli Schwartz: 对。
(——以下为广告,已跳略——)
AI 与 SEO 内容创作
Lenny Rachitsky: 好,我们回到 AI 对 SEO 的影响这个话题。能不能用 AI 来帮助你做这件事,用 AI 来为你创建内容?
Eli Schwartz: AI 是一个工具。大家都说 AI 是工具,不是解决方案。如果你的内容是那段旅程的一部分,你可以用 AI 来创建内容。比如,在 AI 内容真正兴起之前——AI 其实已经存在一段时间了,Jasper 已经存在一段时间了,Writer 也已经存在一段时间了,都在 ChatGPT 之前。在这些工具出现之前,很多公司低成本创建内容的方式是去 Fiverr 和 Upwork 上找人写内容。
那些内容中很大一部分完全没有任何价值。如果你只是为了有内容而创建内容,之前花钱请 Upwork 上的人花 50 美元写的内容,现在你可以用 AI 工具免费生成同样没有价值的内容。AI 作为工具,它创造的东西对公司最终的业务旅程和用户旅程来说不一定有用。但是,如果你之前创建的内容本身是有价值的,现在你用 AI 来以更低成本、更高质量地创建真正有用的内容,那当然可以用。
AI 内容的合理使用场景
Eli Schwartz: 你可以使用 AI 内容的一个场景是,如果你是一个电商网站,卖自己的产品,当然可以用 AI 来写产品描述。这不是一个内容型网站。很多大公司都有这样的做法——如果有人想看看电商网站怎么做 SEO,很多都有大型 SEO 团队。典型的比如 JCPenney、Nordstrom、Macy’s,它们的分类页上有大量内容,但如果你看那些页面的流量关键词,真正驱动流量的是页面上的产品本身。
如果你在找鞋子,Macy’s 页面上有大量关于鞋子功能的内容,这对用户来说没什么意义。他们就是想找鞋子,然后页面上有鞋子。我觉得 Macy’s 在这类搜索上排名还不错。用 AI 去写更多不必要的填充内容,纯粹是浪费时间。
用 AI 来辅助录入产品信息,描述产品是什么,以及一些帮助用户了解产品的功能特性——这不会损害你的 SEO,因为你要优化的就是产品和产品名称本身。
Lenny Rachitsky: 明白了。基本上就是,不要用 AI 生成整篇博客文章。我知道到处都有人这么做,但一定要善用 AI 来帮助你补充现有页面——描述、标题之类的。我能理解为什么 Google 完全看不出来你用了 AI 辅助,因为它只是页面的一小部分,而不是整个页面。
Eli Schwartz: 他们在官方文档中说,AI 本身不是问题。问题在于内容是否有帮助、是否有用。
AI 内容的争议与公众反应
Lenny Rachitsky: 你不用点名,也不用透露什么,但你知道有没有一些公司现在正在用 AI 大量生成非常成功的页面、高转化页面之类的?
Eli Schwartz: 普通用户普遍对 AI 内容有一种反感。我觉得一旦公司公开宣布自己在使用 AI,人们就会对此不满。之前有一家 Red Ventures 旗下的公司,好像是 CNET,公开说他们在用 AI 创建内容。我觉得他们惹上麻烦就是因为公开说了这件事。如果他们没说,我觉得他们的整个模式会非常成功。
你完全可以这样做——拿 10 个产品,想做 10 个产品的评测,完全可以让 AI 先写一篇初稿,然后由人工编辑来审核。几年前——虽然不完全是 AI,但有点类似——有一些体育网站会把比赛中的各种数据自动整合成一篇内容。
如果你看上市公司的财报报道,其实也是类似的操作。公司发布财报,然后现在是用 AI 来生成报道。在 AI 出现之前,基本就是填空式的模板内容——从财报中提取关键信息,然后写出一篇博客文章,其实还挺有用的。我觉得这本身不是问题,用户也不觉得是问题。
如果你去看那种关于财报的老文章,很明显能看出不是专业财经记者写的,但它是有用的。你能快速得到一份摘要,不用去 SEC 网站上翻阅完整财报。
Lenny Rachitsky: Twitter 上有个叫 Noah Smith 的人,他有一个叫 No Opinions 的 newsletter,非常棒。他发过一条推文,说我们正在进入一个海量内容垃圾的时代——AI 生成的大量劣质内容。不管好坏,这跟 SEO 有关,但这也让我想到我们会看到多少糟糕的东西。
正如你所说,关键不在于是不是 AI 生成的,人们和 Google 都会倾向于那些有用、优质的内容,不管是不是人写的。
AI 内容爆炸对 Google 的影响
Eli Schwartz: 这对 Google 来说是一个巨大的问题,因为现在创建的内容数量是海量的,而且还在指数级增长。Google 要爬取所有内容,现在有了更多内容需要爬取,这对 Google 来说成本更高。这也是为什么 Google 对很多网站进行了打击——他们试图清理索引库来节省自己的成本,同时保护用户免受这些糟糕内容的影响。
Lenny Rachitsky: 更疯狂的是,LLM 现在的训练数据很大程度上来自互联网内容,而它们现在面临一个困境——很难避免用 AI 写过的内容来训练 LLM,因为这会形成一个恶性循环:LLM 在自己生成的输出上训练自己。好了,我想回到 AI 的另一个话题,就是 AI 概览。我想很多人会问:“我怎样才能出现在那个回答里?怎样让我的产品出现在 Google 顶部给出的回答中?“这是你建议人们去做的事情吗?这是可以做到的吗?
AI 概览与品牌建设
Eli Schwartz: 我认为 AI 概览本质上是一个品牌建设的事情。出现在 AI 概览中——Google 在其中放了链接,我之前提到过 Google 在内容里嵌了链接——我觉得 Google 的很多做法都是出于潜在的免责考虑。他们放了链接,等于是说:“我们没让你跳金门大桥,是我们摘要的那个网站让你跳金门大桥的。“或者说:“我们没有剽窃,我们只是链接到了那篇内容。我们可能从中提取了超过应有限度的内容,法律还没有对此做出裁决,但我们链接了,所以是合理的。”
目前 Google 在内容中放了这些链接,可以点击跳转。我在 LinkedIn 上做过一个调查,结果让我意外。我原本以为大多数人都不会点击那些链接。我收到了几百份回复,结果基本是一半一半。根据我做 SurveyMonkey 的经验,虽然这不算统计意义上的严谨调查,但结果不可能是一边倒的九一开。这说明确实有人在点击链接,这让我挺惊讶的,因为我不认为 Google 设计那些链接的初衷是为了实用,这意味着 AI 概览本质上是在复制搜索结果。
你有一个 AI 概览做摘要,然后有链接可以跳转,再往下看,基本是同样的内容。这就是 Google 需要解决的问题。对于一个公司来说,要出现在 AI 概览中,这是一个品牌挑战。如果你的链接出现在那里,你大概率本来就在下面的排名结果中了,所以你已经做了你应该做的工作。
如果你在 AI 概览中作为一个品牌被提及——比如”这些是顶级 CRM 工具”——你作为一个品牌出现,这意味着你的品牌建设投入正在见效。我不确定公司是否一定希望自己的内容出现在那里,因为这等于是免费把内容送出去了。他们希望以对自己有利的方式出现。如果你以品牌形式出现,说明你的品牌建设做得好;但如果你以链接形式出现,意味着 Google 拿走了你的内容,而人们不一定会点击跳转。
Lenny Rachitsky: 沿着品牌和 SEO 这条线,你有没有什么……这个问题一直存在:我们在品牌上投资,到底是只希望人们知道我们的存在,还是想真正推动转化?关于如何思考品牌建设和 SEO 之间的关系,尤其是在现在这个 AI 时代,你有什么建议?
Eli Schwartz: SEO 领域最大的神话是,你只需要建设链接,链接建设是扩大 SEO 版图的秘诀和法宝。这完全是错误的,因为大多数人建设链接的方式是购买客座文章,或者付费在一些低权威度的网站上获取链接——这些网站看起来似乎有点权威性,但实际上对谁都没有好处,因为根本没有人会去读这些网站。
有意思的是,所有人都抱怨 Google 是一个极其聪明、无所不知的 LLM,但同时又认为它蠢到会被那些发布在根本没人读的网站上的客座文章所蒙骗。正确的链接建设方式是打造品牌,这是你整体努力中不可分割的一部分。如果你在做链接建设,回到我们之前提到的那些 SaaS 案例——如果你创建了一堆与你要销售的产品无关的内容,然后为这些内容建设链接,这些链接对产品并没有真正的好处,并不能带来实际价值。
如果你能打造出一款令人惊叹、人人都喜欢并愿意使用的产品,你获得的不是链接,而是提及。而且链接现在可能就等同于提及,因为在 LLM 出现之前,链接就是十多年前意义上的 HTML 链接。现在 Google 能够读懂内容,它可以说”好吧,你在这里被提到了”,这就很不错。我们现在承认这可能是一个品牌,也可能是某类查询的匹配结果。
这有助于打造品牌。总的来说,我认为任何 SEO 工作都应该以推广品牌、建设品牌为目标,而不是先做产品,再单独搞一套 SEO——搞一堆可能不相关的内容,再从不相关的网站获取指向不相关内容的链接。这一切应该是一个统一的整体。你在公关上做的事情,同样适用于 SEO。
Lenny Rachitsky: 太好了。很高兴我们聊到了这个话题。还有几个方向我想深入探讨,所以我会稍微跳跃一下,从你在这个领域的深厚经验中再挖掘一些其他内容。其中一个是你已经几次提到了程序化 SEO 与编辑型内容这两个方向。首先,也许可以先解释一下这两个方向分别是什么,让不太了解的人也能明白。其次,你对于应该走程序化还是编辑型路线有什么总体建议?
Eli Schwartz: 到目前为止大家可能已经很清楚了,但我还是要说这取决于用户。基本上,程序化 SEO 就是将大量数据源组合起来,构建一个整合了所有这些数据源的页面。在我的书中,我谈到了两家我最欣赏的程序化 SEO 公司。其中一家的策略是由你之前的一位嘉宾 Luc Levesque 设计的——正是他介绍我们认识的——那就是 TripAdvisor。
TripAdvisor 整合了所有这些数据源——“这是酒店,这些是城市”——然后让用户生成内容(UGC)融合在一起,形成关于每个物业的综合页面。TripAdvisor 并没有为每家酒店单独撰写内容。他们不会写一篇”某位网红对纽约华尔道夫酒店的长篇评测”,也不会写”某位网红对旧金山万豪侯爵酒店的长篇评测”。
他们完全是基于程序化的方式构建的。他们整合了所有数据集,囊括了世界上所有国家、所有物业,合并成一个综合页面,从互联网之初就一直占据搜索结果榜首。他们至今仍然占据着。那个团队经历了许许多多轮的领导者更迭、各种变动,但他们依然是第一名。正是这套策略帮助他们打造了品牌并持续称霸。
第二个例子是 Zillow。Zillow 解决了多年以前房地产行业的一个早期问题——没有人了解一处房产的真正价值。Zillow 整合了所有这些数据集。其中一些是政府数据集,另一些是他们自己基于其他房屋销售得出的对比数据。政府可能说一套房子值某个价格,但根据 Zillow 自己的对比分析,他们可以完全不同意这个估价。
他们不断整合这些数据集,采集房产经纪人的房源照片、社区数据、学校数据,这就是一个程序化 SEO 页面。SEO 是为这些页面引入流量的唯一渠道。他们不会为我的房子投放付费广告,也不会为你的房子投放付费广告。人们在 Zillow 上找到某个页面的唯一方式,要么是从 Zillow.com 首页开始然后层层导航进去,要么就是通过 Google 搜索。这些就是程序化 SEO 的实践。
程序化与编辑型的对比
另一面是编辑型内容。以 TripAdvisor 为例,编辑型就是为每家酒店、每个城市撰写长篇内容。Zillow 也是同理——他们需要为五亿处房产各写一篇编辑型内容。这两种方式都不适合他们的商业模式,而且成本极其高昂。假设每篇内容花费一千美元,这对任何人来说都是无法承受的。
回到我之前说的”取决于用户”这个判断标准——用户到底在寻找什么?用户是在查询房屋价值吗?他们不需要长篇大论的内容,他们需要一个页面直接告诉他们那栋房子的价值是多少。TripAdvisor 的例子中,用户寻找的是单一信息——一个评分,或者可能需要一些用户评论,但本质上他们要的就是那个评分。长篇内容有其存在的意义,但不应该出现在 TripAdvisor 的网站上。
理解用户真正的需求,就能帮助你判断应该从编辑型角度切入,还是从程序化角度切入。我前面提到的一些公司,比如 G2 和 Capterra,正在被全面颠覆——它们既受到编辑型内容的冲击,也受到 Google 自身 AI 概览的冲击。
如果你搜索顶级 CRM 工具,你会看到 G2 或 Capterra,但同时你也会看到 Forbes 写了一篇关于 Salesforce 是什么、HubSpot 是什么的长篇内容,这其实可能是多余的。
Lenny Rachitsky: 我觉得总的来说,程序化是更好的路径,也是最常走通的成功路径——至少我听下来是这样。要在程序化 SEO 上取得成功,你需要具备哪些条件?基本上你是不是只需要拥有某种数据源就行了?能不能给出一个清单:“以下这些条件告诉你这里有一个巨大的机会”?
程序化 SEO 的成功要素
Eli Schwartz: 当存在规模和用户使用场景时,程序化就是正确的解决方案。如果没有这两点,你创造规模就是白费力气。程序化的一个早期版本——同样是在现在这些互联网公司出现之前——是为美国每一个邮政编码创建一个内容页面。假设你提供某种本地产品,你会为每个邮政编码建一个页面,理论上这是程序化的,但用户并不会这样搜索,而且 Google 已经用自己的本地产品彻底颠覆了这种做法。
Eli Schwartz: 程序化总能做出点什么来。我经常看到程序化的错误做法——网站会看一眼自己的数据,说:“你看,我们做软件评测,世界上有一万种不同的软件,我们就给每种软件单独建一个页面,再组合上一个什么城市。“你确实可以这样生成程序化页面,但根本没有使用场景。真正判断该用哪种程序化方案,归根结底就是两个问题:“有没有人真的在搜这个东西?这个页面本身是否提供了解决方案?“Zapier 提供了解决方案,因为确实有人在搜那种组合。但市面上有很多类似 Zapier 的工具试图模仿 Zapier,却没有人去搜它们对应的解决方案。回到你之前提到的一些模板公司,我认为很多做模板的公司在做程序化,但并非每一个模板都有对应的使用场景。它们做了程序化,却是无的放矢。
话说回来,SurveyMonkey 的情况是,我们有一个可以做模板的”肥头”(fat head),但它不是长尾。我们能做一百个模板,但做不了一万个。
谷歌反垄断判决中的 SEO 洞察
Lenny Rachitsky: 我想做一个 SEO 迷思快问快答环节。不过在此之前,我知道你花了大量时间阅读最近那份针对 Google 的裁决书——关于 Google 和 Apple 之间那整套合作的案子。你之前跟我说,那份文件揭示了很多关于 Google 战略和搜索未来走向的有趣信息。我们稍微展开聊聊,你从那次阅读中收获到了什么?
Eli Schwartz: 非常精彩。286 页,我基本上读了一本书。那是一份判决书。
Lenny Rachitsky: 天哪。
Eli Schwartz: 我想说,写这份判决书的法官或书记员对 SEO 运作机制和数字营销的理解,比我遇到的大多数数字营销从业者都要深。
Lenny Rachitsky: 哇,没想到会听到这个评价。
Eli Schwartz: 是的。这份判决解释了 Google 如何决定页面排名,解释了 Google Ads 的竞价模型如何运作。我觉得非常精彩。判决还讨论了社交媒体的对比问题。法庭对提交给他们的很多问题理解得非常透彻。比如,Google 试图说自己不是垄断者,因为它属于社交媒体领域。法庭分析后得出结论:“社交媒体不能与 Google 搜索相提并论。”
Google 还试图辩称他们的很多做法不具有垄断性,因为存在 Bing 这样的替代选择。法庭对所有这些论点都进行了深入分析。我觉得这份判决之所以引人入胜,还因为很多证词本来就是公开的,而判决展示了法庭如何将这些材料整合在一起。我学到的东西之一是市场份额——Google 的市场份额。这个数字过去没人愿意公开谈论。
Google 会说:“哦,我们觉得大概 80% 吧,“因为他们不想被视为垄断者。Microsoft 则会说:“嗯,我们大概有 20% 吧。“我最近看到 Microsoft 的一份文件,声称自己在搜索领域拥有很大的市场份额。而判决书中写的是,Google 占据了 98% 的移动端搜索量。我忘了是法院说的、Google 没有反驳,还是这来自 Google 自己的文件,但数据就是 98% 的移动端搜索。
判决中我认为最令人震撼的部分,是 Google 的默认合作伙伴关系对其成功的贡献有多大。这起案件的原告是美国司法部(DOJ),起诉 Google 构成垄断。但案件中还有一个投诉方,是一家叫 Neva 的公司——它是一家搜索引擎,创始人是从 Google 出来的,我记得好像是研究部门的负责人之类的,一位 Google 高管创建的新搜索引擎。
庭审很大一部分围绕 Neva 为什么无法成功展开——尽管它有更好的搜索引擎,尽管它有更好的用户体验,但它没有任何默认分发协议。这是我觉得最引人深思的部分。Google 与 Apple 有默认协议,同时还拥有 Chrome 浏览器。
Google 拥有所有这些导向搜索的入口,这意味着——这也是我最大的结论——我认为 ChatGPT、Perplexity、Claude,或者任何其他 LLM 初创公司都没有机会真正撼动 Google,单纯从产品质量角度来看是不可能的。
这些公司现在都还是初创公司,即便是 OpenAI,虽然有 Microsoft 的投资,仍然是一家初创公司。判决书中的大量文件都在说明:使用习惯的力量有多大,品牌的力量有多大,Google 的分发协议有多强大——正是这些把搜索用户源源不断地推向 Google。还有几个非常有意思的细节。一个是,Bing 曾试图免费把自己的搜索提供给 Apple,而 Apple 说无论出什么价格,他们都不会在 Apple 产品中使用 Bing 搜索。
另一个是,Mozilla 曾与 Yahoo 合作,做了收入分成协议,结果用户全部切换回了 Google。所以当人们说”SEO 已死,大家都会转去用 OpenAI”的时候,你只需要看看 Google 25 年来建立的一切,我真的不这么认为。我真的认为分发渠道和品牌惯性将保住 Google 今天拥有的几乎一切。
那么,为什么 Google 如此拼命地反击、打造 AI 概览?我认为有两个答案。第一,正如我所说,是华尔街。他们需要让华尔街相信自己很重视这件事、正在积极布局。第二,流失几个百分点给 ChatGPT 对 Google 来说代价巨大,所以值得投入。但我不认为他们对失去全球最大搜索引擎的地位有任何真正的担忧。
Lenny Rachitsky: 但这场审判的意义不就在于,现状可能会改变,他们可能会失去那个默认地位吗?不管最终会不会发生,这才是整个博弈的核心,对吧?也许他们不能再维持现状了,而这可能打开了机会之门。我觉得这才是潜在的机会所在。
Eli Schwartz: 我不这么认为。是的,我不是法律专家,我不知道 DOJ 会怎么做。但有意思的是,判决出来后 Google 发布的新闻稿——Google 发推说:“我们感谢法庭认定我们是世界上最伟大的搜索引擎,后续还有更多消息。“他们会反击的,这几乎就像一个恶霸被宣布为世界上最厉害的恶霸,然后他说:“看吧,我早就说过我很强。”
读判决书的时候,它确实写着:“不存在竞争者。Google 是最伟大的搜索引擎。他们拥有 25 年的优质数据。他们做得非常出色。“我看不出有什么办法——如果真的打破默认协议,Apple 会损失 Google 支付的那一大笔钱,然后用户还是会去用 Google。光靠这一点不会改变什么。我认为他们唯一能做的就是想办法扶植一个竞争对手。
Lenny Rachitsky: 对,我觉得这是一个很有启发性的结论。即便 Google 不是默认搜索引擎,人们也会主动切换到 Google。Ben Thompson 对此事的分析中其实有一个很有意思的观点,我会把那期播客链接放出来。那个分析确实改变了我的看法,因为当你听到 Google 付钱给 Apple 成为默认搜索引擎,听起来就很阴暗、很不公平。但实际上这笔交易是收入分成。Google 把广告收入的一定比例分给 Apple,最终大概是 200 亿美元,对 Apple 来说是一大笔钱。其他任何人也可以去找 Apple 说,“我们也愿意把搜索的收入分成给你。“只是没有谁有 Google 那么大的体量,付不出那么多钱。你仔细想想,为什么 Google 不能去找 Apple 说,“嘿,我们帮你做搜索,广告收入的大部分分给你,这样你能赚一大笔钱”?如果你真的仔细想想,禁止这种做法其实挺奇怪的。但这同时也不公平,因为没有任何其他竞争者能与之抗衡。
Eli Schwartz: 对,法庭也正是这个意思——正因为如此,Neva 根本没有机会。还有一点,我以前做过一个调查,那是很多年前我在 SurveyMonkey 的时候,我做了一个搜索引擎市场渗透率调查,想了解人们用的是什么,比如 Bing 对 Google。在调查里我很大方地把 DuckDuckGo 也放了进去,回收了数千份问卷,结果显示只有 1% 的人在使用 DuckDuckGo。
我当时不知道的是,DuckDuckGo 自己其实也不知道他们的市场渗透率是多少。当我分享这份调查结果时,DuckDuckGo 的 CEO Gabriel Weinberg 在 Twitter 上找到了我,说:“你能把数据分享给我们吗?我们很想看看。“这大概是十、十二年前的事了。而在这次判决书中写道,DuckDuckGo 的市场份额是 2%。十二年前我测出来是 1%。DuckDuckGo 融了大概一亿美元,做了各种品牌合作,在棒球比赛场上都能看到他们的 logo,该做的都做了,结果只从 1% 涨到了 2%。这里面信息量很大。以 Google 的地位,我不认为有谁能撼动它。
SEO 不会消亡
Lenny Rachitsky: 太有意思了。感谢分享你的解读,这样我们就不用自己去读那份报告了。在进入快速问答 SEO 迷思环节之前,关于 AI 或 SEO,你还有什么觉得特别重要、想留给听众的内容吗?
Eli Schwartz: 我认为最重要的一点是:SEO 不会消亡。用户主动获取信息的需求会永远存在。我认为这些智能音箱——不管是 Google 的、Apple 的还是 Amazon 的——之所以一直没有真正普及,原因之一就是你没有选择权。你对 Google 助手说话,它只给你一个答案。你对 Amazon 的说,它也只给你一个答案。你说”我要买卫生纸”,Amazon 就直接给你买了它决定给你买的那一款。
用户需要选择权的场景永远存在。判决书中还有一点:Google 认为我们目前仍处于 LLM 的极早期阶段,即使他们运用了大量机器学习来理解用户,始终还是需要真实的用户数据——而 Google 拥有过去搜索的海量真实用户数据。我认为用户永远会想自己主动获取信息,不可能出现一个 AI 完美地理解你到了这种程度——它知道就你个人而言,你会想点击第五条搜索结果,而那条结果恰好是你当下最需要的。这意味着搜索结果中永远需要提供多个选项。没错,大多数点击确实集中在前一、二、三条结果上,但必须要有第七条、第十条,必须有多页结果,因为确实有人会翻到后面的页面。我想强调的最重要的一点就是:这一切意味着搜索的形式会改变,大量漏斗顶部的搜索会消失,但总体而言,人们进行搜索的行为会一直存在。
最后再补充一点:当你通过搜索去做某件事时,你是在采取一个行动,而企业正是从这些行动中获益的。假设你经营一家酒店,你希望人们来住你的酒店并为此付费。即使搜索量下降了,需要采取这个行动并向你付费的总人数并不会改变。如果你在卖鞋,人们还是需要买鞋。
但如果你卖的是信息——比如你是媒体,你是 WebMD——那确实,你的收入会下降。因为你过去把信息聚合、整理后免费提供出来,换取用户点击广告;而 Google 现在也把这些信息免费给用户了,但不需要任何人点击广告。
SEO 迷思
Lenny Rachitsky: 有意思。我在想,这是不是存在一个投资套利机会——预测哪些业务会因此衰落、哪些会因此繁荣。不管了,我想聊聊 SEO 迷思。我知道你有一些想法,其中一些我们之前已经聊过了,就是那些人们深信不疑但实际上是错误的 SEO 观念。让我直接问你:人们对 SEO 有哪些常见的误解?
Eli Schwartz: 最大的一个我们其实反复说过了,就是”公司到底需不需要做 SEO”。人们总是默认一个假设:你融了资,有了一个营销团队,就应该做 SEO。我觉得这个观念真的该彻底放下去了——想想用户旅程,你到底该不该做 SEO?这个问题就像人们会问其他渠道一样:“我应该做品牌营销吗?我该在时代广场打广告吗?“没有人会说:“我刚融了一百万美元,我完全应该在时代广场打广告。“顺便说一下,恭喜你在时代广场上亮相了。
Lenny Rachitsky: 谢谢。
Eli Schwartz: 没有人会决定说,“哦,我钱多到烧不完,那就烧吧。“但 SEO 不知怎么就变成了这样一种东西:“哦,我融了资,现在我需要一个 SEO 团队。“所以我认为这是第一个需要彻底打破的迷思。
Lenny Rachitsky: 很好。还有呢?
Eli Schwartz: 外链建设。外链建设其实就是品牌建设——要思考如何打造品牌、让别人提到你。当你做外链建设时,你在建立一种关系:从那个链接到你的内容,到你试图变现的产品或其他东西。这才是外链建设应有的样子。如果只是想着获取链接——觉得一个 HTML 链接就等于 SEO 成功——我认为这完全是错的。
然后,可能最大的迷思总体上是认为 Google 本身是一个黑箱。确实,排名的细节有微妙之处,没有人能真正揭示到底是什么让你排第一、什么让你排第三。但构建 SEO 的基本思路其实非常简单,Google 也有一份最佳实践指南,告诉你应该做什么:建一个 Google 能理解的网站,为页面建立链接,撰写有用的、用户想读的内容。这些就是基础,从这个基础出发去构建你的 SEO 策略并在此基础上持续优化。但那种”我现在把这三步全做了,就一定能排第一”的想法,我认为是完全不正确的。还有一种假设是,只要你在 SEO 上做了很多事,你就一定会成功,我认为这也是不对的。
SEO 入门建议
Eli Schwartz: 是的,确实有微妙之处。但在大多数情况下,你能从零到十的过程中,光靠遵循最佳实践就能走到第八步。
Lenny Rachitsky: 哇,听起来很鼓舞人心,我喜欢这个说法。顺着这个话题简单问一下——如果有人想开始做 SEO,我知道你会有点偏心,因为你就是帮企业做这个的——你是建议先请一位像你这样的专家,还是自己先试试看,听听这期播客、读几本书和博客,还是别的什么?
Eli Schwartz: 是的,我确实有偏心。我认为应该让对的人来给你提供 SEO 建议。如果有人想做……如果他们已经验证了确实需要投入 SEO,花钱请人帮忙做 SEO,这就回到了我们之前讨论的资源问题。如果你想请一位增长顾问,增长顾问可能很贵,但你能压缩学习曲线,避免犯那么多错误。
如果你招一个内部员工,我帮很多公司招 SEO 人员时经常看到这种情况——他们的预算不多,这意味着招到的人经验不会太丰富。那这个决策对吗?如果你请代理公司,代理公司通常很贵,而且很多时候是按交付物收费的。那他们做的这些交付物值得你付这个钱吗?
一个很容易做的交付物——现在很多代理公司都在产出——就是内容。但这又回到了我们之前的讨论:你真的需要这些内容吗?代理公司很容易把内容作为交付物来卖,因为它确实是一个看得见、摸得着的东西。像我的增长顾问工作大部分是在交付策略,这就很难卖,因为我在提案阶段还没有一个策略可以交付——策略还没开发出来,而且我们在执行之前看不到成果。但代理公司可以绕过这个问题,直接说”你按月付钱,我们每月交付这些东西给你。“
TikTok 会取代搜索吗?
Lenny Rachitsky: 有个话题我想聊聊,它可能是一个迷思,也可能不是。但根据你之前说的 Google 主导地位的观点,我觉得你会说这是个迷思。现在很多人在讨论 TikTok 和 Instagram 正在取代搜索。我自己现在确实经常用 TikTok 搜索,比如”怎么解决这个问题?""怎么切西瓜?""不知道,给我宝宝找个什么杯子?“效果真的很好。你怎么看 TikTok 和 Instagram 视频正在替代大量搜索行为这件事,尤其是对 Z 世代和更年轻的孩子来说?
Eli Schwartz: 在 Google 那个反垄断判决中,我想 Google 之前也说过——63% 的 Z 世代使用 TikTok 来搜索。
Lenny Rachitsky: 哇。
Eli Schwartz: 对。但我觉得这只是一个标题数字,因为归根结底,就像你在播客里反复提到的那样——购买旅程。有些东西更适合 TikTok 或 Instagram,有些东西则更适合搜索。如果你在做漏斗顶部的探索,你可能会看 TikTok 视频来了解某个话题。但当你进入漏斗中部,比如——TikTok 上很热门的搜索是旅行类。你想去旅行,想去东南亚,但不确定具体去哪。那你可能会看一堆视频,看网红分享的体验。但现在你准备预订了,你要订酒店、订机票。这些都不会在 TikTok 上完成,全部都会在 Google 搜索上完成。这就是漏斗中部。
大量搜索会流向不同的地方,但最终,Google 仍然是做这些搜索的正确位置——做那些漏斗中部的搜索。也许以前没有 TikTok 的时候才是不正常的,你不得不忍受阅读那些纯粹为 SEO 目的而写的糟糕内容,而现在你可以获得来自 TikTok 的丰富、动态的内容。
还有一个有趣的点是,如果 TikTok 真的在美国被封禁——看起来很可能会发生——那么接手所有这些用户的公司会是 YouTube。YouTube 有 YouTube Shorts。然后我们又可能看到一个潜在的垄断局面,Google 会把这些 YouTube Shorts 用来解决 TikTok 之前在解决的问题。这会很有意思。
Lenny Rachitsky: 天哪,到处都是 Google。
Eli Schwartz: 是的。
Lenny Rachitsky: 听你说漏斗中部和漏斗顶部的区别时,我有一个想法——理解漏斗中部的一个简单方式就是那里有意图。你确实有了购买某个具体东西的意图。嗯,不错。
Eli Schwartz: 意图确实是关键,但这本质上就是购买旅程。在漏斗顶部,你只是好奇,你不知道自己是否有意图。而在过去那种只关注漏斗顶部、只以排名为 KPI 做 SEO 的方式——我认为这是不太正确的做法——那里没有意图。你展示了一个与非用户匹配的 KPI,与非买家匹配的指标。
当你沿漏斗向下移动,意图出现了。那里有真正的用户,但流量更少。漏斗中部的流量总是会更少。但这没关系,因为你的 KPI 与业务指标紧密相关。
技术 SEO 的迷思
Lenny Rachitsky: 如果还有什么我们遗漏的,再来几个迷思,然后我们就收尾,进入非常精彩的闪电问答环节。还有什么你觉得是人们容易误解的关于 SEO 的重要迷思吗?
Eli Schwartz: 我再讲一个迷思——技术 SEO 是解决一切 SEO 问题的灵丹妙药。最近 Google 做了很多算法更新。再说一次,我不太喜欢替 Google 辩护,因为在读了那份判决之后,我觉得 Google 确实做了一些相当不地道的事情。但 Google 最近做的很多事情,其实是在回应用户讨厌的东西。网上有大量糟糕的 SEO 内容,所以 Google 在近期的更新中试图清除这些劣质 SEO 内容。
当网站被这些更新打击后,他们会试图通过做技术 SEO 来解决他们自认为出现的问题。他们会联系我或像我这样的人,说”你能做一个 SEO 审计,看看我们的流量为什么下降了吗?“他们流量下降的原因是他们做了一些对用户并不真正有用的事,污染了 Google。他们需要从根上解决这个问题。技术 SEO 有它的用武之地——如果你有一个数千万页面的大型网站,如果你是 Zillow,你如何链接到每个房源;如果你是 Airbnb,你如何链接并让 Google 抓取和理解你的网站——这非常重要。
但如果你有一个一百页的网站,卖一个 SaaS 工具,技术 SEO 可能就没那么重要了。把钱和时间花在技术 SEO 上就是浪费。所以我认为另一个很大的迷思是——你觉得可以用正确的 SEO 方案来解决 SEO 问题,而不是去想”我们做个审计吧”或者”我们去获取更好的外链”或者”我的页面……”页面速度也是一个——代理公司喜欢拿页面速度相关的指标来卖服务,Google 也一直在改这个指标的名字叫什么,但它其实没那么重要。
页面速度的重要性被高估
Eli Schwartz: 因为在边际上,是的,也许你是 Kayak,你在和 Expedia 竞争,Kayak 可能稍微快一点,也许这有影响。但如果你所在的领域竞争对手都是些又慢又糟糕的网站,那就完全无所谓了。所以花很多钱去修复那个问题不是正确的做法。
Lenny Rachitsky: 我喜欢你刚才把 SEO 说得很平易近人、很简单的样子,让人觉得这是自己就能做的事,而不需要再去钻研什么黑暗魔法之类的……
Eli Schwartz: 这不是什么黑暗魔法。在大多数情况下,它应该是简单的。确实有很多公司在 SEO 上遇到了很多麻烦,做正确的 SEO 可以带来数十亿美元的收入。但很多价值可以通过非常简单的 SEO 来释放。没有 SEO 背景的产品经理也能做出足够的 SEO 成果,为公司赚很多钱,而不需要成为 SEO 专家。
Lenny Rachitsky: 我很喜欢这个说法。其实 Airbnb 就是这么运作的。负责 SEO 的产品经理并不是什么历史上的 SEO 传奇人物。他们就是普通的产品经理,自己摸索 SEO,却产生了非常大的影响。所以这让我很有共鸣。Eli,在我们进入非常精彩的闪电问答之前,还有什么我们没有涉及到的、你觉得很重要的事情,或者你想留给听众的?
SEO 预测:自上而下法
Eli Schwartz: 有的。我们之前谈了如何决定是否要给 SEO 投入资源,但还没有谈如何理解对 SEO 的预期。这是我在 Faire 工作时想出来的一个方法,Faire 是一家非常有意思的公司。它的拼写是 F-A-I-R-E。它本质上是一个批发版的 Shopify。当时他们要进入一个新的国家,想了解从 SEO 的角度来看,在那个国家上线能有多大的增长空间。
大多数人做 SEO 预测用的是自下而上的方法,也就是从关键词出发。假设我们要卖一双鞋。
你做一个自下而上的预测,打开一个关键词研究工具——这些工具大同小异。看看每月有多少人搜索”鞋”。然后你估算自己在这个词上的排名会是多少,再估算点击率会是多少,基于这些算出点击量。然后代入一个转化率,这就是你的自下而上预测。
然后你再做一个假设:工具里”鞋”这个词没有覆盖到所有搜索了”白鞋”、“黑鞋”、“跑鞋”之类的人,所以你按比例放大。随便乘个 10,这就是你的 SEO 预测了。
这个方法的问题在于,得出的数字通常太小了。你拿到了这个数字,然后你去提案说要进入一个新的国家——我想进入日本卖鞋。你说,“嗯,我觉得我能获得 500 个用户,因为根据我的计算方式,这是我的点击率,这是我的排名,所以是 500 个用户。“没有人会为此买单。太少了。
但如果你要进入一个新的国家,其实有一个相当简单的方法来判断你的 SEO 上限——你可以用那个国家的人口。这不是 Faire 卖的那种产品,所以我不算在泄露什么。假设你在卖鞋,你正在上线一个新的产品网站,在日本卖鞋。
我估算一下,我不知道日本的实际数字。假设日本有一亿人口,你要卖鞋。你不可能把鞋卖给一亿人。我们只卖男鞋。对半分,日本 50% 是男性,所以我们有 5000 万潜在买家。
但我们还要说,“有老年人,有年轻人,我们不会把鞋卖给他们。“所以把 2500 万再砍一半。然后,我们是在互联网上卖鞋,但不是所有人都在网上买鞋。有些人在店里买,有些人在网上买。所以你取一个比例,说”我们拿 2500 万市场中的 10%。250 万人在网上买鞋,我们要对这 250 万人做 SEO。我们的预期市场渗透率是多少——不管是 10%、50% 还是 100%。“这就是你的数字,然后乘以他们每年买多少双鞋,再乘以 AOV,这就是你的预测。
这个方法可能不够精确。你可以上下调整那些数字。你可以说,“嗯,我的 AOV 不对。我的市场渗透率估计错了。我对总人口估算有误。我没想到在日本根本没人网购鞋子。“但在任何时点你都可以回去修正你的预测。而如果你做的是自下而上的预测,在很多情况下它从一开始就是错的,因为关键词研究量本身就是错的。
我合作过一些非常有意思的公司,他们领域里最大的查询词就是那个品类本身最大的词。比如我合作过 WordPress。WordPress 这个词是网站建设领域最大的查询词,没有其他任何词比 WordPress 更大。而每一个关键词研究工具上关于 WordPress 这个词的搜索量数字,和 Google Search Console 显示的完全不同。所以”当你基于关键词研究工具做这些预测时,你构建整个预测所依赖的那个初始数字如果错了,你的整个预测就错了。“所以当你用这种自上而下的方法时,本质上就是一个 TAM 预测。自上而下的方法让你更接近真相。当然,你可能到不了真相——我从没见过一个产品计划能准确预测最终结果——但它能帮你做出比拍脑袋更好的决策。
Lenny Rachitsky: 哇,你分享的这个洞察太重要了。你是说关键词研究工具在衡量机会方面其实并没有那么准确。那问题就来了——为什么会这样?为什么它们差这么多?它们做错了什么?遗漏了什么?
Eli Schwartz: 它们各有各的秘密配方来做估算。即便是 Google——再次说明,出于过去垄断地位的原因——Google 也不能把他们在 Google Ads 上看到的真实搜索量数字直接公布出来。他们必须从另一个数据源购买数据,然后再发布出来。具体原因我忘了。它们本质上都是在猜。它们用各自的专有算法来估算,这就是为什么很多工具之间数据不一致——因为各自的算法不同。所以无论你用的是 Semrush 还是 Ahrefs,我最喜欢的工具之一是 Similarweb,Similarweb 有很多浏览器插件,可以窥探人们的搜索方式。但 Similarweb 有浏览器插件,并不意味着它能看到每个人的搜索,所以它还是要用算法来估算全球的总量。
而且,我不知道它们中有没有任何一个能接近真相——我曾经和大型公司合作,有些关键词我可以对照 Google Search Console 来看,有些工具高估了 10 倍,有些低估了 10 倍。我不是说它们把具体的月度数字搞差了一点点,我是说它们差了好几个数量级。
Lenny Rachitsky: 哇,这太疯狂了。所以你的建议是……你是直接忽略那些数字?还是说,“可以看看,但别太当真,别拿那个数字作为依据”?
关键词工具的参考价值
Eli Schwartz: 它们具有参考价值。比如说,你想知道”大多数人拼写 WordPress 是带空格还是不带空格”,工具很清楚地表明人们不带空格。但如果我要做一个预测,说”我绝对要基于这个具体数字来做决策”,我不这么认为。我可以把它们用于归一化。我只用它们做归一化,来理解人们的搜索方式。我非常喜欢用户旅程,这些工具对理解用户旅程很有帮助。但作为一门精确科学,它们很难真正派上用场。
Lenny Rachitsky: 太不可思议了。所以它主要用在数量级判断上,以及在同一个工具内比较不同关键词的相对关系。好的,有意思。
外部工具 vs. Google Search Console 的真实数据
Eli Schwartz: 对。我有一个很好的例子,说明这些工具在估算流量方面可以错得多离谱。COVID 初期我合作过一家公司,是家上市公司。他们的董事会成员给他们发邮件说,“你们在疫情期间被竞争对手打得一败涂地,看看你们的竞争对手,我在用其中一款工具看,你们排在最底下,你们什么都做错了。“CMO 来问我,“我该怎么回复这位董事?“我说,“这位董事完全错了。来看看 Google Search Console。“Google Search Console 虽然也不完美,但再说一次,它是真实数据。我们的 Google Search Console 显示我们的流量在疫情期间翻了四倍,所以外部工具显示什么其实并不重要。所以这些工具是有帮助的。工具是有帮助的。但我不认为它们是事实真相的来源。
Lenny Rachitsky: 哇。这个例子太好了,很适合作为收尾。我这边还记了一件事,想确保你有机会聊聊——你对帮助人们进入 SEO 领域,以及像你一样成为顾问这件事非常热心。分享一下你觉得在这方面对大家可能有用的建议。
SEO 顾问行业的未来与建议
Eli Schwartz: 作为我自己的预测,我认为对 SEO 专业人才的需求将会爆发式增长,因为搜索结果页面中正在发生的很多变化,意味着公司必须调整自己的策略。再说一次,很多公司关注的是排名,关注的是流量,而其中很大一部分将会消失。页面布局突然变了,流量也变了。如果他们的 SEO 本来就不是正确适配的,这未必会影响他们的实际营收。而这将激发出大量对 SEO 帮助的需求,我已经看到了这一点。
过去一年里,我自己的 inbound 大幅增长,因为很多东西在重新洗牌,大家有很多疑问。所以会有人想要进入 SEO 咨询领域,我认为整个增长顾问领域也将有巨大的需求。
我想说的是,我一路上得到了很多优秀的导师帮助,你之前的一些嘉宾比如 Casey Winters、Yuriy Timen、Ethan Smith,他们都有很棒的节目,但他们给我的建议不在于如何成为更好的运营者,而是如何更好地做销售、如何更好地写提案。
对于任何想成为增长顾问的人,我想说这才是你真正需要打磨的技能——沟通、销售、提案,而不是纠结于成为最好的运营者。你当然应该成为最好的运营者,但这可能已经是你具备的技能了。不要想当然地认为,因为你是今天的好运营者,你就会自然而然成为优秀的增长顾问。在你现有的工作中锻炼增长顾问的肌肉。不要辞掉你的本职工作,利用业余时间兼职练习——练习销售、签单、执行、留住客户。如果你在这些方面成功了,你独立发展也能成功。
Lenny Rachitsky: 太棒了。我记得你有一些文章深入讲了这些内容,我们会在节目笔记里附上链接,对吧?
Eli Schwartz: 对。
Lenny Rachitsky: 很好。可以这样理解——把自己当作一个产品,把与你合作的过程当作一次旅程。
Eli Schwartz: 是的。你在建立的是品牌,不是咨询公司。
Lenny Rachitsky: 我很喜欢这个说法。Eli,今天的对话太精彩了。接下来,我们进入了非常令人期待的闪电问答环节。准备好了吗?
Eli Schwartz: 绝对的。
闪电问答
Lenny Rachitsky: 开始吧。第一个问题:你最常向别人推荐的两三本书是什么?
Eli Schwartz: 大家应该也看出来了,我非常喜欢用户旅程和了解用户。有一本书叫《小数据》(Small Data),作者大概是 Martin Lindstrom,不确定发音对不对。他在书中讲的是如何理解人和理解人们如何购买。他深入探讨了这整个过程,实际上他会住进别人家里,观察他们使用各种工具和玩具的方式。我一直觉得这本书非常迷人,会推荐给想了解用户的人。
Simon Sinek 的《从为什么开始》(Start With Why),同样的道理。真正理解一个产品、一个企业应该做什么,才能理解用户。这本特别适合增长顾问,不一定适合 PM——《百万美元咨询》(Million Dollar Consulting)。这本书是很多年前 Ethan Smith 推荐给我读的。关于如何建立品牌和成为增长顾问的一本精彩著作。后来我甚至和作者合作了近一年,请他做我的教练。非常好的书。
最后一本不推荐说不过去的,是我妻子的书——《如何真正做到不在意别人的看法》(How To Stop Caring What Others Think: For Real)。这本书就是讲这件事的——认识自己的成功,不必在意别人怎么看你。
Lenny Rachitsky: 太美了。顺便问一下,那本书是不是在你身后的背景里?如果没有的话,你得把它放到你背景里,你妻子的书。
Eli Schwartz: 我妻子做的这个背景。所以是的,它在背景里。
Lenny Rachitsky: 哪一本?哪个颜色的那个……好让大家认出来。哦,就是那本大的。我还纳闷为什么那本比其他的稍微高一点。
Eli Schwartz: 对,就是这本书。
Lenny Rachitsky: 就是它。稍微往上挪一点让我们看到完整封面。哦,很漂亮。《不在意别人的看法》,很漂亮。我也需要这本。For Real,我喜欢这个副标题,for real。
Eli Schwartz: For Real。
Lenny Rachitsky: 我喜欢。好。
Eli Schwartz: 它也许是一个 Google 搜索词,也许不是。
Lenny Rachitsky: 哈哈,我懂了。我明白你在这儿做了什么。天才。好,下一个问题。你最近有没有特别喜欢的一部电影或电视剧?
Eli Schwartz: 我在飞机上看了一部电影叫《黑莓》(Blackberry),看起来像纪录片,所以我不知道你看过没有,但它非常引人入胜。因为它是那种有悬念的电影,而你完全知道结局。这部电影太精彩了。我之前并不了解全部历史,但他们曾经占据整个智能手机行业,然后归零了。电影完整地展现了这段历程,让你不禁思考——“你永远不能躺在功劳簿上。你必须创造用户想要的产品,理解你的用户,持续销售那个产品。“而不是觉得,“我们是第一名,这永远不会变。“很棒的电影。
Lenny Rachitsky: 有意思,最近刚好有另一位嘉宾也推荐了这部片子,所以它正在上升势头。这部电影大概已经出来一年了。我看过。
Eli Schwartz: 对,如果不是在飞机上,我永远不会看。完全……感觉像是那种适合在飞机上看的片子,但真的很完美。
Lenny Rachitsky: 我是在家看的。我特别喜欢里面那些技术细节——他们能找到巧妙的方式来利用蜂窝网络,当时看来简直不可能,而正是这些让他们实现了所有那些功能。我喜欢那个消息功能之类的东西。真的很吸引人,因为我之前不知道他们基本上是逆向工程了蜂窝网络的运作方式,才实现了他们所实现的功能。很棒的电影,疯狂的故事。
最喜欢的产品
Lenny Rachitsky: 下一个问题,你最近有没有发现一个特别喜欢的新产品?可以是 app,也可以是某个实体产品。
Eli Schwartz: 不是什么新产品,而且我是那种会反复爱上它的类型。可能有点老套,但就是我的手机。我最近去了东南亚旅行,之前在那边住过一段时间,已经八年没回去了。用手机能做的事情太多了——旅行时用 Google Maps、Waze,点外卖,跨越多个国家,同一个 app 打车、支付。太方便了。八年前我在那里的时候,还得买 SIM 卡,手机基本不能用,也没法支付。这次体验完全不同了,我几乎不需要电脑。所以又重新爱上了我的手机。
还有一个……也不算新,但我非常喜欢,就是 Grammarly。我喜欢写作,所有东西都是自己写,不用 AI 写。Grammarly 让我成为一个更好的写作者。
Lenny Rachitsky: 你知道我最近对 Grammarly 有什么发现吗?我刚升级了他们的付费套餐。他们大概是向你推销付费方案最厉害的产品,因为它就一直在你面前晃。“嘿,我们有这么多建议,有这么多可以改进的地方。只要付一百多块钱,我们有这么多建议能让你的文章好得多。“就一直在你面前。他们太擅长这个了。他们把我拿下了。而且我也很高兴,毕竟如果你全职做这件事的话,那笔钱也不算多。
Eli Schwartz: 我在书上用 Grammarly 的时候特别尴尬,满屏都是蓝色和绿色的标记。我就想,“天哪,要是我早知道就好了。”
Lenny Rachitsky: 我的 newsletter 其实有一位非常厉害的文案编辑。每次她都能找出一百个可以改进的地方,即使我觉得已经完美了。有意思的是,她不采纳 Grammarly 建议的情况比我预想的要多。这一点挺有趣的。不过总之,Grammarly 确实很棒,我一直在用。
人生座右铭
Lenny Rachitsky: 还有两个问题。你有没有最喜欢的人生座右铭,经常自己使用、与朋友家人分享,或者在工作生活中觉得很受用的?
Eli Schwartz: 我不知道能不能浓缩成一句话,但我一直鼓励大家的是:真正地想得大、想得远。我刚跟 LinkedIn 开始合作做咨询,而我跟 LinkedIn 谈合作已经谈了六年了。所以永远不要放弃。你遇到一个人,提出一个想法,你永远不知道它会通向哪里。我的整个职业生涯都是如此——在不同的地方生活,不要只想着当下会怎样,而是在建立一段关系。你永远不知道什么事情会在哪里发生。所以要看大方向。
Lenny Rachitsky: 很喜欢。顺着这个思路,最后一个问题。好奇你这里会想到什么。你最引以为傲的 SEO 成果是什么?
最引以为傲的 SEO 成果
Eli Schwartz: 我不知道能不能说最引以为傲。但我真的特别喜欢我在 Tinder 做的事情,因为它把整个旅程串联在了一起。有些用户不知道 Tinder 能解决一个非常具体的问题,而它就在那里解决那个问题。而且这不是 Tinder 对自己的定位。他们把自己看作一个约会 app,但它其实是一个解决孤独感的问题,恰好回答了一个 Google 搜索。
结束语
Lenny Rachitsky: 这次太棒了。我觉得这会帮助很多人思考 SEO,尤其是当他们意识到情况正在变化时,他们现在有了这个资源,可以说”我明白了,这是我应该改变的,这是我应该做的。“我特别喜欢其中贯穿的主线——“其实没那么复杂,即使你以前从没做过,你也可以做到。“非常感谢你来这里,跟我们分享这么多智慧。最后两个问题。如果有人想联系你、跟你合作,在哪里可以找到你?顺便也推荐一下你的书。最后,听众怎样能帮到你?
Eli Schwartz: 你可以在 LinkedIn 上找到我,搜索 Eli Schwartz 就行。也一定要看看我的书,书名叫 Product-Led SEO。其实有一点关于个人品牌和个人排名的心得——搜索你自己名字的时候,排在哪里并不重要。如果 LinkedIn 排在你名字的第一位,那也挺好的,因为别人找到你了。很多时候,品牌和个人会非常在意自己的排名位置,但重要的是这个旅程。只要找到你了,就是找到你了,是不是第一并不重要。
这也说明了 SEO 不全是关于链接的。我相信搜索我自己的名字,大概率你的也是,我的排名比 LinkedIn 还高。还有我的书,如果你搜索 Product-Led SEO,我个人的网站——域名权重并不是最好的——排名比 Amazon 还高。这应该就能直接反驳那种”一切皆关乎链接”的说法。不是一切关乎 SEO 指标。真正重要的是正确的匹配。所以做正确的匹配,你就会出现在你该出现的位置。
大家最能帮到我的方式就是订阅我的 newsletter,给我写作的灵感和反馈。我真的很享受写作,也很享受听到大家的想法。我的 newsletter 在 Productledseo.com。
Lenny Rachitsky: 你所有东西都这么一致,我很喜欢。Productledseo。Eli,非常感谢你来。
Eli Schwartz: 谢谢邀请。
Lenny Rachitsky: 大家再见。
感谢大家的收听。如果你觉得这期节目有价值,可以在 Apple Podcasts、Spotify 或你最喜欢的播客 app 上订阅。也请考虑给我们评分或写评论,这对其他听众发现这个播客真的很有帮助。你可以在 lennyspodcast.com 找到所有往期节目或了解更多关于这个节目的信息。下期再见。
术语表
| 原文 | 中文 |
|---|---|
| Ahrefs | Ahrefs(SEO 分析工具) |
| AI Overviews | AI 概览(Google 的 AI 搜索摘要功能) |
| Airbnb | Airbnb(短租住宿平台) |
| algo update | 算法更新 |
| AOV | AOV(客单价,Average Order Value) |
| Ben Thompson | Ben Thompson(科技行业分析师,Stratechery 作者) |
| beta | beta(测试版) |
| Bing | Bing(微软搜索引擎) |
| bottoms-up forecast | 自下而上预测 |
| buyer journey | 购买旅程 |
| Capterra | Capterra(软件评测平台) |
| Casey Winters | Casey Winters(增长与产品顾问) |
| Cleveland Clinic | Cleveland Clinic(克利夫兰医学中心) |
| click-through rate | 点击率 |
| CMS | CMS(内容管理系统) |
| CNET | CNET(科技媒体网站) |
| CPM | CPM(千次展示费用,Cost Per Mille) |
| CRM | CRM(客户关系管理) |
| default agreements | 默认协议(搜索引擎与设备厂商/浏览器的预装合作协议) |
| DOJ | DOJ(美国司法部,Department of Justice) |
| DuckDuckGo | DuckDuckGo(注重隐私的搜索引擎) |
| editorial | 编辑型(内容) |
| Ethan Smith | Ethan Smith(SEO 顾问,Graphmate 创始人) |
| Expedia | Expedia(在线旅行预订平台) |
| Faire | Faire(批发电商平台) |
| fat head | 肥头(与”长尾”相对,指搜索量集中在少数热门关键词的分布形态) |
| Fiverr | Fiverr(自由职业者服务平台) |
| flywheel | 飞轮 |
| freemium | 免费增值(免费使用+付费升级的商业模式) |
| funnel | 漏斗(营销漏斗,指用户从发现到转化的旅程阶段) |
| G2 | G2(软件评测平台) |
| Gabriel Weinberg | Gabriel Weinberg(DuckDuckGo CEO) |
| Google Search Console | Google Search Console(Google 搜索管理工具) |
| Healthline | Healthline(医疗健康信息网站) |
| hockey stick | 曲棍球棍式增长(指先平缓后急剧上升的增长曲线) |
| inbound | inbound(入站流量/渠道) |
| index | 索引库(搜索引擎的网页数据库) |
| Jasper | Jasper(AI 写作工具) |
| JCPenney | JCPenney(美国百货零售商) |
| Kayak | Kayak(旅行搜索引擎) |
| keyword research tool | 关键词研究工具 |
| KPI | KPI(关键绩效指标,Key Performance Indicator) |
| lightning round | 闪电问答 |
| LLM | LLM(大语言模型,Large Language Model) |
| Luc Levesque | Luc Levesque(SEO 从业者,TripAdvisor 早期增长策略负责人) |
| Marriott Marquis | 万豪侯爵酒店 |
| Martin Lindstrom | Martin Lindstrom(品牌顾问与作家) |
| mid-funnel | 漏斗中部 |
| Mixpanel | Mixpanel(产品数据分析工具) |
| MQL | MQL(营销合格线索,Marketing Qualified Lead) |
| Neva | Neva(前 Google 高管创立的搜索引擎公司) |
| No Opinions | No Opinions(Noah Smith 的 newsletter 名称) |
| Noah Smith | Noah Smith(经济学家、时事评论作者) |
| Nordstrom | Nordstrom(美国高端百货零售商) |
| OpenAI | OpenAI(AI 研究公司) |
| organic traffic | 自然流量(非付费搜索带来的访问量) |
| page speed | 页面速度 |
| Perplexity | Perplexity(AI 搜索引擎) |
| persona | 用户画像 |
| PM | PM(产品经理,Product Manager) |
| PRD | PRD(产品需求文档,Product Requirements Document) |
| product-led SEO | 产品驱动型 SEO |
| programmatic SEO | 程序化 SEO |
| Quora | Quora(问答社区平台) |
| Red Ventures | Red Ventures(数字营销与媒体公司) |
| rev share | 收入分成 |
| SaaS | SaaS(软件即服务) |
| SEC | SEC(美国证券交易委员会,Securities and Exchange Commission) |
| Semrush | Semrush(SEO 与数字营销分析工具) |
| SEO | SEO(搜索引擎优化) |
| SEO audit | SEO 审计 |
| Similarweb | Similarweb(网站流量分析工具) |
| Simon Sinek | Simon Sinek(领导力与激励领域作家) |
| SOC 2 | SOC 2(服务组织控制报告,一种安全合规认证) |
| structured data | 结构化数据 |
| SurveyMonkey | SurveyMonkey(在线调查平台) |
| TAM | TAM(总可达市场规模,Total Addressable Market) |
| top of funnel | 漏斗顶部 |
| top-down | 自上而下 |
| TripAdvisor | TripAdvisor(旅行点评与预订平台) |
| Udi Milo | Udi Milo(Tinder 增长负责人) |
| UGC | UGC(用户生成内容,User-Generated Content) |
| unstructured data | 非结构化数据 |
| Upwork | Upwork(自由职业者服务平台) |
| Vanta | Vanta(自动化合规平台) |
| Wade Foster | Wade Foster(Zapier 联合创始人) |
| Waldorf Astoria | 华尔道夫酒店 |
| WebMD | WebMD(医疗健康信息网站) |
| Writer | Writer(AI 写作平台) |
| Yuriy Timen | Yuriy Timen(增长顾问) |
| Zillow | Zillow(房地产信息与估值平台) |
此文档由 AI 分片翻译(translate_long_document)