Zigging vs. zagging(逆势而行):HubSpot 如何打造一家 300 亿美元的公司 | Dharmesh Shah(联合创始人/CTO)
Zigging vs. zagging: How HubSpot built a $30B company | Dharmesh Shah (co-founder/CTO)
Dharmesh Shah: Some of the best startup advice I’ve heard is startups should focus on one thing and be really, really exceptionally world-class at that one thing. And one of our early zigs is we are going to do exactly the opposite of that.
Some Fun Facts
Lenny: You have no direct reports and I don’t believe you’ve ever had direct reports at HubSpot?
Dharmesh Shah: I could become passively okay at management with some training, with some coaching. I don’t want to spend any years of my life becoming passively okay at something.
Speech Prep and Humor Methodology
Lenny: What was that process like to define the culture?
Dharmesh Shah: My co-founder and I were having one of our founders meetings and he said, “Oh, Dharmesh, I hear this culture thing is really important. By the way, can you go do that?” I’m like, “Okay. Brian, of all the people in all the company, is like, I am the worst possible person.” It’s not that I don’t like people, I just don’t like being around them a whole lot.
Using Humor in Presentations
Lenny: Something that’s really unique and interesting about you is you’re obsessed with comedy and keynote prep.
Dharmesh Shah: It comes down to this metric that stand-up comedians use called LPM laughs per minute. I have custom software that I’ve written that will say, “Okay, here are the points at which the audience laugh.”
The Specifics of LPM
Lenny: Today my guest is Dharmesh Shah. Dharmesh is the co-founder and CTO of HubSpot and also one of the most fascinating and first principled thinkers I’ve ever met. In our conversation, we cover a lot of ground. Dharmesh’s hilarious and ingenious approach to putting together a talk, including measuring laughs per minute, his biggest lessons from being a public company exec for over 10 years now, especially while being a startup guy at heart, how he approached creating and scaling the culture of HubSpot, which you’ll find both hilarious and inspiring.
Why founders and product teams are all fighting the second law of thermodynamics, how to zig, while everyone else is zagging, how and why Dharmesh leans into his strengths, including never having a single direct report during his 18 years of running HubSpot and so much more. This episode is so fun and will expand your mind in many ways. With that, I bring you Dharmesh Shah after a short word from our sponsors.
And 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 it helps the podcast tremendously.
Once you’re ready, simply embed the dashboard or report into your application with a tiny code snippet. The best part, your end users can use Explo’s AI features for their own report and dashboard generation, eliminating customer data requests for your support team. Build and embed a fully white-labeled analytics experience in days. Try it for free at xplo.co/lenny. That’s E-X-P-L-O.co/ lenny.
Dharmesh, thank you so much for being here and welcome to the podcast.
The SoloWare Concept
Dharmesh Shah: Thanks for having me, Lenny. It’s an honor.
Lenny: It’s my honor, and thank you for joining me. You are a wildly fascinating human and so I thought it’d be fun to start with just a bunch of fun facts that I found about you online. And what I’m thinking is I’ll just walk through them. Tell me if they’re true as I walk through them, and then this is going to lead to a bunch of different topics that I want to talk about. How does that sound?
Choosing Not to Manage
Dharmesh Shah: Sounds great.
Building the Company Your Way
Lenny: Okay, here we go. So one, you have no direct reports and I don’t believe you’ve ever had direct reports at HubSpot.
Dharmesh Shah: That is correct. 7,000 plus employees? Exactly zero direct reports from time, T equals zero.
Advice on Going Public
Lenny: Okay, we’re going to talk about that. You also don’t do one-on-one meetings as a result. You don’t have reports to have one-on-one meetings with?
Dharmesh Shah: That’s correct.
Sustaining Passion and Motivation
Lenny: You built many side projects while at HubSpot, including a product called Wordplay, which at one point made $90,000 per month and had 16 million users.
All Employees as Insiders
Dharmesh Shah: That is correct.
Contrarian Thinking and First Principles
Lenny: You also bought chat.com for $10 million and then you sold it two months later for more money than you bought it for?
High-Conviction Bet: Focusing on SMBs
Dharmesh Shah: It was actually 15 plus million dollars, but I had kind of stated eight figures, but… well, yes.
Lenny: Ooh. Is this breaking news?
From No Titles to Classic Titles
Dharmesh Shah: It’s breaking news, yeah.
Uniform Early-Stage Salaries
Lenny: Wow. So that’s not what you bought it for?
Designing for Simplicity
Dharmesh Shah: Yes.
Entropy and Complexity
Lenny: Okay, and you sold it still for an undisclosed amount, is that right?
Fighting for Simplicity
Dharmesh Shah: Undisclosed amount, yeah. And more than that, I’ll just say that.
Lenny: Yeah, there was a profit, and I know you gave some money to charity and there’s some promises you made where people commented on your LinkedIn post and you gave more money away, so it was awesome. Okay. Also, you’re a billionaire?
Reverse Gravity in the SMB Space
Dharmesh Shah: That is once again true. It’s still true.
High-Conviction, Low-Consensus Bets
Lenny: Great. Also, you were born in a village in India that had no paved streets, no traffic lights, no hospitals.
The “Mile Wide, Inch Deep” Strategy
Dharmesh Shah: That is correct.
When to Go Wide, Not Deep
Lenny: Okay. Before HubSpot, you founded two companies, but I think the more important fact there is you promised your wife you would not start another company before you started HubSpot and then you ended up and went on and started HubSpot.
Dharmesh Shah: Yes. I had promised her that I would not do start… I had to hang up the proverbial entrepreneurial hat. Best laid plans and all that. I met my co-founder in grad school, which is that I was going to… The plan was, “Oh, I’ll go back to grad school and go find myself.” And then the plan was to go eventually teach, not do another startup. But then I met my co-founder there and one thing led to another as these things go.
The Flash Tags Concept
Lenny: I hope she’s forgiven you.
Dharmesh Shah: I do too.
The Decision-Making Mechanism
Lenny: Okay. Final fact. So Kipp, is that how you pronounce his name? Kipp, your CMO?
The Calorie Principle of Decisions
Dharmesh Shah: Mm-hmm.
Lenny: Okay, great. So I asked him what to ask you, and he tells me something that’s really unique and interesting about you is you’re obsessed with copywriting and comedy and keynote prep. That you have a very unique approach to preparing for a talk and how you think about humor and slide design and story arc.
The Default Stance: Saying No
Dharmesh Shah: That’s very true. Public speaking is not something that comes naturally to me. So we’re going to start there. We can talk a little bit about the copywriting, other things that are sort of related. And one of the… and I try to not do things that I’m not good at and don’t enjoy. One of the rare exceptions that has snuck through is the need for me to get on stages as founder when we have our annual HubSpot event, our conference, which at the time when we did our first one, it was, like, 150 people at the Marriott. And then it’s steadily grown since then, and so now it’ll be like, 10,000 people in the live audience and hundreds of thousands online.
But as I came to this realization that I was not going to be able to talk my way out of it, and I saw where this was headed, I’m like, “Okay, this is something I’m going to have to actually learn, because I can’t delegate it, can’t talk my way out of it.” And so I’m a big believer in this talent versus skill. So we all understand talent. It’s like if someone has a talent for music, someone has a talent for athletics, whatever it happens to be. And that’s true. That exists, but the way I think about it is talent basically controls the slope of the curve, but most things are actually acquirable skills. So for instance, if you have a talent for music, let’s say, you might learn music faster than someone else and your ceiling may be higher, but does not mean that other people can’t learn music, right?
It’s like an acquirable skill. That’s lesson number one. And lesson number two is that there is a process, and this is the engineer in me of just kind of functional decomposition of a problem to say, ” Okay, in order for me to be a pretty good public speaker, what are the underlying sub-skills required in order to accomplish that? Okay, skill number one, you have to be able to stand up on stage and not pass out.” That’s kind the bare minimum. And so what I did over the years is to say, “Okay, I’m going to functionally decompose this skill of public speaking in front of,” what I call high-stakes speaking. I’ve got to actually hold the audience’s attention. That’s part of success. So it doesn’t matter what the message is, nothing else will matter if I lose their attention, which is increasingly easy to do now in kind of short attention span society.
And every year I will pick a different part of the public speaking skill set and say, “Oh, I want to learn about slide design and visual expression,” Not the actual design. I’m not a designer. But then the one thing that I came up that has had the most impact is humor. Then you break that skill down just like the engineer would do. It’s like, okay, well there’s humor writing, there’s humor, execution on stage, whatever. There’s standup comedy. There’s like, okay, so how do you put all these pieces together? And along that journey, I came up with this… and I’ve always had this, I’m a very data-driven, quant based person. It’s like, “Okay, well, how will I know whether something is actually funny enough or not, whether I’m losing the attention or not?” And so in preparation for one of these high stakes keynotes that I do every year, I will have my talk, I will do practice runs.
I will have increasingly larger actual live audiences that I will practice it in front of. I will record those talks, everything so far. Nothing unusual yet. I will have the talks transcribed. I have custom software that I’ve written that will say, “Okay, here are the points at which the audience laughed.” Actually audibly laughed. That’s the only way it counts. And this actually… and it comes down to this metric that standup comedians use called LPM laughs per minute. And there’s actually a benchmark that you can say, “Oh, well…” And standup comedians have a very, very high LPM. Business talks have a very, very low LPM. The most popular Ted talks there’s a high strong correlation between the most popular Ted talks and high. So I have a goal, look, I’m going to have a minimum I want to improve. And so I have software that will calculate the laughs per minute.
And so then you learn these little kind of tools of the tricks of the trades. It’s like, okay, if you’re solving literally for that ratio of laughs per minute, there are two ways to improve it. One is to add more laughs. Two is to decrease the number of words between laughs, right? So one of the things a software measures, it gives me a visual map. It’s like, here’s the entire talk broken down with little square per minute of talk, or 30 seconds, and it’s like, here’s what they like. It’s like, “Oh, here I went one minute and 17 seconds and nobody laughed.” So I have to do one of two things. Shorten that segment if there’s nothing funny that can be said without being overt about it, or I have to say, “Okay, I’m going to find something funny to try to inject into this and kind of break it up.”
Anyway. And that’s a long way of saying if I can get up on a public stage and learn that particular skill, anyone can learn just about anything. It just comes down to practice and measurement and just getting incrementally better over time.
Framework for Evaluating Ideas
Lenny: Wow. I feel like I’ve just learned so much about you and how you think just from that question and how you approach this problem. On this humor piece, I’m curious, any lessons to share on how to be funny?
Four-Dimensional Startup Idea Framework
Dharmesh Shah: One of the kind of tricks of the trade in humor, and you learned this particularly in stand-up comedy, but just any comedy. And I just hadn’t thought of it this way, is that… and this is very tactical, but it works and it’s a very easy thing to apply. So let’s say you’re looking at… and most of us, we’re not stand-up comedians. We’re not going to tell jokes in a classic sense.
And even stand-up comedians don’t tell jokes. They actually have stories and they happen to have a punchline or two. So, two tactical pieces of advice. One is when you’re telling the story, whatever the funny bit is, those have to literally be the last words of that particular segment. So once you deliver it, then you have to stop talking. And the reason you have to stop talking is the audience needs about a half a second to react, and then they want the permission to laugh. And if you’re continuing to say words, it’s going to make them feel awkward and it’s going to be less funny. So even if moving a phrase in the sentence that you’re about to say, it’s like, “Well, that sounds kind weird to put those set of words at the end of the thing.” Do it anyway even though it’s awkward because it will actually work better.
That’s kind of tip number one. Tip number two on LPM generally is that having stories such that you can spend the time setting up the story, but then have multiple funny bits, multiple punchlines, because you’ve already made the investment. I’ve said 75 words in order to set up this context, say something funny that’s amusing, that’s whatever. Say some more things. You already have the… and you’ll see the stand-up comedians do this all the time. And then say something else that’s funny because a laugh is a laugh. Doesn’t really matter that it was still the same story. So leverage the investment you’ve already made in establishing context and scene, and then just squeeze more humor into the end of it in punctuated fashion.
Defining Culture: The Culture Code
Lenny: I feel like the entire podcast could just be this one thread, but I need to move on to other topics.
One quick question. So you said ideal laughs for a minute is two to three-ish? Is that what you said?
Culture as a Product
Dharmesh Shah: Yeah, two plus is hard for a business talk because stand-up comedians, the reason they can do it is that’s all they’re really solving for is LPM. In business talks, you actually have another agenda other than just being entertaining. You actually have a message. You’re trying to promote something. Promote an idea, promote a product. So anyway, in my mind, for most normal circumstance, even if you can get above a one, 1.2, 25, you’re in the top decile in terms of top, generally given by non-professionals.
Bugs in Your Culture
Lenny: Wow, this is incredible. So after every talk, you run this transcript and audio through your program to see how you did.
Core Values: Federal vs. State Laws
Dharmesh Shah: Yes, yes.
Lenny: And are you ever going to publish and release this program? Because that sounds really cool.
Culture Self-Fulfillment vs. Transparency
Dharmesh Shah: I have this notion of what I call SoloWare. And SoloWare is exactly what it sounds like. It’s software built for exactly one person, and in my case, that one person is me. And so I’ve been doing this thing for 30 years right now. I just build things that I would myself find useful. And the nice thing about Solo, one of the many nice things about it is that the UI is only for one person. You don’t have to do a whole lot of testing because it only needs to work for one person. And the most important thing is that because it’s only for one person, if it stops bringing utility, you can just turn it off. You can just stop using it. Versus if you have users, even 10 users, and you put it out there and then people would be disappointed if you took it down for whatever reason, that’s kind of hard.
So anyway, that’s the calculus I go through, “Is this useful enough to enough people where it’s worth the calories of making it non-solo ware, making it even micro-ware?” I just made that up, but yeah.
The Self-Fulfilling Vision Statement
Lenny: Wow. Okay. Let’s see how the YouTube comments react and how many people would want this and slash pay for it.
Let me go in a slightly different direction. We talked about how you have no direct reports, you’ve never had any direct reports? I know that’s somewhat related to a piece of advice you often share, which is to lean into your strengths and to not do things that you’re not amazing yet. So can you just talk about why this is the case? Why don’t you have any reports? And then I think this idea of just strengths and the importance focusing on strengths.
Dharmesh Shah: Yeah. I’ll tell you the quick story as to how this came about. My co-founder and I had the founding meeting for HubSpot. This is the week where we’ve decided we’re going to do this, and then we have this list of questions or topics for the founders discussion. One of them, and I think I’ve published a list, these are the questions all co-founders should ask each other in that kind of early period. One of the ones it’s like, okay, well who’s going to be CEO? This was an easy one because I had been CEO twice before and had figured out I’m not that good at it. I actually suck at that. And one of the reasons I suck at that is that part of the CEO role is being good at managing people. It’s not just about leading and having a vision or whatever. Actually have to do the management as a craft. And so, that part I already had in my head.
I think Brian, my co-founder, already had in his head that with… That’s how this… that was a very quick conversation. And in the moment, as soon as I had that win behind me, it’s like, “Okay, well I don’t want to be CEO.” Brian wanted to be CEO because he’d never done it before. I’m like, “okay, that was easy.” And then I said, “Oh, and by the way, I don’t want to have any direct reports.” It just popped into my head as an idea. I had not pre-planned it or anything, and I sprung it on him.
He’s like, “Okay, well…” And the reason is because I have learned that I suck at management. I’m a reasonably smart person. I think I could become passively okay at management with some training, with some coaching or whatever. I don’t want to spend any years of my life becoming passively okay at something. I would rather take those same calories and take the things that I’m good at that I actually enjoy. And those things are highly correlated. You tend to be good at the things you enjoy. You tend to enjoy the things you’re good at. And Brian was like, “Yeah, sure.” And then I’m like, “No, no, Brian, you don’t understand.” Because then I was like, “There’s going to be a time in the company.” He was like, “Oh, our VP of engineering just quit. Just be interim with something or whatever and have this team report to you until we find..” It’s like, “We’re going to have those things and you’re going to have to promise to me we’re not going to do that. We’re going to…” It’s like I am not going to have direct reports.
And we had that kind of heart to heart conversation, and it was one of the best decisions I’ve made both for myself and for HubSpot. Had I not made that decision, I’m a startup guy, I don’t know that I would’ve survived at a company at HubSpot scale. And right now I can honestly say I’m having a better time at HubSpot now at 7,000 people than I was having at 70 people. And the reason is, I get all the upside of scale, which is I can make big bets. So we can have long-term things and plans, whatever, and plan for global domination without the downsides of scale, which is, “Oh, you’re having to manage all these people and all the kind of mechanics.” And that’s hard. It’s a hard skill to have.
I appreciate other people that have it. I don’t. And so I get all the upside with very little of the downside of scale, and that’s what kind of keeps me engaged and energized and happy at HubSpot. So I think it’s been both productive for HubSpot and good for me.
Thoughts on AI
Lenny: I love that this is an example of how you can build a company the way that you want to build it. You don’t have to do it the way everybody else has done. I imagine there’s a lot of challenges to this too, but I think it’s inspiring to just like, “Hey, I’m going to design a company the way I want to build it.”
AI as Cognitive Scaling
Dharmesh Shah: Yeah, it’s amazing what you can get away with in terms of shaping the universe to your liking. I think people automatically assume that things have to go a certain way, especially founders and first time founders.
It’s like, you should at least try it. It’s like, “Okay, well I want to do it…” First of, I’ll give you another tactical example. Both my co-founder and I are night people, not morning people. Just happens… I didn’t pick him because of that. We learned that later. And so we made it just from the early days of HubSpot in the early years at least, we’re like, “No meetings before 11 A.M. Period.” That was it. And then we adjusted it, I think maybe three or four years and we’re like, “All right, we could have meetings before 11:00 AM. You just can’t invite one of the co-founders to it. So you’re welcome to meet amongst yourselves if you choose to do so.” But anyway, yeah, it’s little things.
Lenny: That’s an amazing rule. I have a personal rule where I have no meetings before 3:00 P.M. because I don’t work anywhere. I can more so create my schedule, but the idea is I create this deep work time in the beginnings of the day.
A Historic Opportunity for Founders
Dharmesh Shah: I love that. Yeah.
Lenny: Okay. So you talked about you’re a startup guy. At the same time HubSpot is approaching its 10 years of being a public company and I think 18 years since founding.
Imperative to Declarative Paradigm Shift
Dharmesh Shah: Yup.
Lenny: I’m curious what lessons you’ve taken from that experience being someone that’s clearly startuppy person, being an exec at a very fast growing and large company. And I’m thinking from the perspective of a founder who’s maybe thinking about starting a company or starting company, what advice you might have to share for them? And also for people in the exact role today?
Dharmesh Shah: A couple of things. One, since we are publicly traded down, this doesn’t get talked about enough. I’m going to go and get this out there and use this as a platform to get this message out, which is I think too many founders, once they get to a… They are very apprehensive of going public and going through the IPO process and being a publicly traded company because they think… I mean, that’s like the beginning of the end. “It’s going to change everything and my life is going to suck.”
And I think they’re over indexing on what they think it’s like to be a publicly traded company. In my own personal experience, and Brian, my co-founder would attest to this. And yes, there’s certainly a tax or things we have to do now as a publicly traded company that we didn’t have to do as a privately traded company. But there are actually benefits. For instance, if you kind of venture back in private, let’s say, you’re sort of at a little bit the whim of the market and how VCs perceive your category and all these things or whatever. And you get, every now and then, when you’re out raising a round of capital, you sort of get marked to market. Now here’s sort of what the valuation is based on what someone’s willing to kind of fund the company at. One of the nicest things, and this is underappreciated, is that we know right now what the market values HubSpot at. I can tell you that any business day of the week. And there’s a niceness to that. And that’s the other thing I tell folks within the company is that the valuation, which is we’re in a relatively efficient market. The public stock market is a relatively efficient by the economic definition of the term, in efficient market. So, the valuation will oscillate around the value.
So, if you focus on creating values, here’s what we’re actually building, valuation will sometimes be higher than you deserve, sometimes will be lower than you deserve, but over the fullness of time, those two things will move in lockstep, right? That’s been demonstrated. That’s what happens. That’s almost a definition of an efficient market is that information moves essentially, and then the valuation catches up with the actual value.
But here’s the other social reason why I think founders should go public, which is… So in HubSpot, and I hadn’t thought about this so I’m not going to take credit for being this generous, magnanimous person, is that in our first pre-public years, I think we were eight years old when we went public, we had created roughly a billion dollars in market cap, give or take. We hadn’t been a unicorn. So it was slightly below when we went public. So HubSpot was never a unicorn because that’s private privately held company, over a billion.
In the subsequent years, we went from a billion dollars in market cap to 30 billion, which is roughly what it is now. That first billion of market cap creation, a very small number of people got to participate in. The rest of the $29 billion in market cap, everyone, every public investor got a chance to participate in. And so I like the idea of, okay, well you believe in your company. There’s lots of people that believed in HubSpot. It’s nice to let them kind of participate. Our customers, our partners, our well-wishers, and they can sort of now have the upside. And you don’t get that in the private markets because it’s a very closed… you can’t just buy shares in every private company you like. I wish that existed, but it doesn’t. So, it’s a way to let the market at large participate in your growth earlier.
Lenny: What about from the perspective of just staying excited and motivated and finding things to take on, is there any lessons there for founders that are, I don’t know, at a larger company and just, “Okay, this is going to help me stay excited about what I’m doing”?
Dharmesh Shah: Yeah, I think this is… and Bezos exemplifies this, I think the best, which is, just because you’re a public company does not mandate that you do certain things. All that’s necessary is the kind of transparency that says, “Here’s what the company is doing, here’s what we’re about, here’s what we’re solving.” And he had this with his very first annual shareholders letter, which he obviously writes every year, which is, “This is what Amazon is about, we’re solving for the long term or whatever.” And he didn’t say this in a snarky way. If you reread the letter, but if you read between the lines, it’s like, if you don’t agree with this approach, you should not buy Amazon shares. It’s like, it’s a completely optional thing. “You don’t have to buy it, but this is the way we’re going to run the business.”
And he stayed true to that. And you can continue to do that. I’ll tell you one quick story. This is once again, the anti-dissuade founders that might be reluctant to go public. So one of the core values at HubSpot that’s been part of the culture from the early early days, literally days, is transparency. And so, one of our mechanisms by which we implement transparency is we’ve had this thing that says all information within HubSpot is equally shared with everyone in the company, period.
Everything. So our balance sheet, what we raised capital at, whether we’re going to be able to meet payroll in six weeks, all those things, everything was transparent all the time. And with two exceptions. One is if it was illegal for us to share, let’s say if we were looking at some acquisition or whatever and we had non-disclosure or the information was ours, not completely ours to share. And one example of that was salaries. So, we feel that salaries are co-owned both by the company and by that individual. And it’s not up to us to share someone’s salary if they don’t want to do that. Okay. All right.
So, we had that transparency and then as we go public, we have this meeting with our investment bankers and our lawyers. This is in the IPO prep process, and they sit down and as part of this all day meeting, one of the questions is like, “Okay, so who are your designated insiders?”
So who are your designated insiders? Brian nor I had ever taken the company public before, we’re like, “We don’t know what that is.” It’s like, “Oh, it’s this group of people that will have access to all the financials and will know.” And we’re like, “Oh, okay, so how long is that list?” And they’re like, “Five or six.” And so then Brian says, “Seven? Seven?” And they’re like, “Yeah, you can do seven.” And I say, “Eight? Could you do eight?” And so we discovered through that mathematical induction process, is like, “If it’s true for N, is it true for N plus one?” Is that there’s no actual legal limit to how many insiders a company can have.
And so we ended up doing, literally the day we went public, is we designated every single employee to be a designated insider. So that allowed us to maintain the transparency thing because there was no rule that said you could only have 5, 6, 10, and there was hundreds of people. And we did not make it optional. We didn’t say, “Oh,” because there are downsides to being an insider. There’s windows within which you can’t trade, there’s stuff. But we’re like, “Okay, well, we believe in transparency. Everyone’s going to have access to it, and it’s not an opt in kind of thing. You’re like, “This is it. We’re all in it together."" And we still do this to this day, even at 7,000 people, everybody’s an insider, so we can still share all the financials with everyone all the time.
Lenny: You have many contrarian opinions is what I’m getting from this conversation already. And the way you’re pushing the envelope on, “Can we have 10, can we have 1,000?” I’m curious if there’s anything else there of just contrarian ways of running a company. I know there’s probably many, but what comes to mind?
Dharmesh Shah: So one of the phrases we like to use in our industry overall is this notion of first principles. I don’t know who popularized it, but it’s used a lot. I’m not going to say something snarky, but I think most of the time first principles is used, it’s actually used incorrectly. Because the way first principles is supposed to work, as Elon would probably describe it, is that it’s not first principles like first principles of what we believe, these are our core first principles in the company, it’s the first principles of the universe. It’s not what you believe is true, it’s what do we all collectively know to be true to the best of our knowledge? It’s science. That’s physics. Like the first and second law of thermodynamics, those are first principles. All the layers on top of that are things that you think are true, are your assumptions or the decisions you made, and that’s fine. You may have made some decisions.
For instance, transparency is not a first principle, it’s a founding principle. It’s not everyone should do this, it’s a completely core thing. And so yes, I can be a contrarian, but you have to limit the dimensions. So the number of things that you’re a contrarian on has to be greater than zero, but not too high. So what I think of is high conviction, low consensus bets. So high conviction is, “We really, really believe this,” and low consensus is that most other people don’t. And there are other people have phrased this differently. I think Peter Thiel actually says it more simply and elegantly, which is, “You need to be right about something that other people think you’re wrong about for a very long time.” And you just have to be right.
So one of the early high-conviction decisions we made at HubSpot was that we were going to focus on SMB as our target market. And I’ll make a case both for SMB and for having a high conviction bet. It doesn’t have to be your target market, but something. Once you make that bet, because it’s low consensus, it’s going to be hard and necessary for you to have the conviction because everyone disagrees, including your board, including your investors, including potential investors. So we have that for literally all 18 years of HubSpot’s history. We have always had the… All during the IPO roadshow, it’s like, “Yes, we get that you’ve on SMB, but so what’s the path to the enterprise?” No, you don’t understand. This is not a go-to market strategy. This is not like, “Oh, we’re going to conquer SMB first,” this is not Bowling Pin, this is not Geoffrey Moore, this is a, “We are here for SMB.” That’s what we’re doing.
And so we maintain that conviction for a very long time, and that helps you make other bets that are secondary. But you don’t want to reinvent everything. So I’ll give you an example of something that we had in the early years of HubSpot that we kind of forwent. So in the early years of HubSpot, we’re like, “Okay, we believe in a flat organization.” One of the manifestations of that flat organization is we had no titles in the company. Nobody had a title. So even though you had business cards, whatever, we had no titles. If you introduce yourself in a meeting, and we had this up to hundreds of people, when you introduce yourself like, “Oh, I work in product, I work in engineering,” whatever, it’s not like, “I’m director of X or VP of Y,” because that, as a thing, didn’t exist.
Then as we scaled, we came to the realization that there’s actual value that people ascribe to titles. And the reason they ascribe value to titles is because there are rumors that there’s life beyond HubSpot. I couldn’t believe it. “What? There’s a life out there outside of HubSpot?” It’s like, ” Yeah, looks like.” And then when we go to the Thanksgiving dinner or whatever, and I talk to my aunt, I need some shorthand. It’s like, “Oh yeah, I got promoted to director. I got…” Just to show some progression because it’s a signaling device.
And in this life beyond HubSpot and outside of HubSpot, it’s like, other people are also going to want to know where I was in… It’s like, “Okay,” and this was the winning argument. I love this argument. Because we had a daylong meeting on talking about this one topic. And the winning argument was, “Dharmesh and Brian, I recognize why we don’t have titles. I recognize this.” And so they made the case of, “Yeah, there’s value. And by virtue of you not giving us titles, that means you’re having to compensate us in an economic sense in other ways. So you’re losing out because it’s like, is that value high enough to not do that?” That’s a strong case.
And then we had it down to three choices. We’re like, “Okay, well,” we could have chosen to just stay firm with still no titles. We could have said, “Okay, we’re going to have classic titles.” Then I had the option on the table of, “Make up your own title.” You can be grand poobah of X, and that can be your… And then we ended up making the decision to go with classic titles because that actually served the need of, there’s a benchmark, there’s a standard, otherwise it doesn’t mean anything. So having made-up titles is pretty much the equivalent of having no titles because they lack any meaning anyway. But that’s one of those things that, okay, you iterate, you change things as they go.
Lenny: Yeah. I think there’s so much to this, “Let’s just try things differently. Okay, that didn’t work. Okay. It’s fine. We’re just going to do what everyone else is doing here.” And then often you discover a thing that, “Okay, this is actually better.”
Dharmesh Shah: Yeah.
Lenny: Just to understand, when people have no titles, so someone say as a product manager, do you call them a product manager, or just like, “I’m an employee at HubSpot, I have literally no title”?
Dharmesh Shah: At the time, it was no title. And now we have classic titles, so we have levels of things and-
Lenny: Yeah. But back then it’s like, “You’re just here-”
Dharmesh Shah: No title. Literally no title.
Lenny: ”… we don’t know exactly.”
Dharmesh Shah: No title.
Lenny: I saw somewhere that you paid everyone $5,000 monthly as their salary at the beginning of HubSpot, including yourselves for a long time.
Dharmesh Shah: Yes.
Lenny: Is that right? Okay.
Dharmesh Shah: Yep.
Lenny: Another fun fact. How long did that last, by the way?
Dharmesh Shah: I’m trying to think back. It’s a long time ago. I want to say probably the first year, year and a half.
Lenny: Amazing.
Dharmesh Shah: One of the things that, and this is one of the core… And not everything that we… Obviously not everything was right, but not everything is applicable to other companies. But there are some things that I think are, and one of the things that I think are applicable is just this need to solve for simplicity. It’s like the simpler you can make things in the early years, the better off you generally are. And this is going to sound weird, so how we wound up with transparency is not because of some moral, it’s like, “Oh, everyone deserved to have this,” and this is the truth, so it’s not some moral position. It was a, when we hired the first employee, and both Brian and I [inaudible 00:33:54], when we hired the first employee, we had to make the decision, it’s like, “Oh, so what files do we give them access to and what…” And so Brian is just like, “All of it. Nothing to hide. Why would we do that?”
And so then we thought out… And then the output of that was like, “Okay, well, A, binary decisions are much easier than non-binary decisions.” It’s like, “Okay, everything or nothing is much, much simpler.” And then it’s like, “Okay, well, if it works for Employee Number Three in the company, why wouldn’t it work at Four? Why wouldn’t it work at Five? Why wouldn’t it work at 50?” And we just kept doing it because it’s like, “Okay, well, until it breaks, we’ll just keep doing that.” And this applies to so many different aspects of HubSpot, but as we try to start with the simplest possible thing that might work, and then, if necessary, add the complexity, add the things or whatever. Even our original, this is actually the tactical thing that everyone should take away, especially early-stage startups, very first HubSpot office has exactly four tables because aspirations for growth, because we were in a co-working space, so, we lived in WeWork back in the day, here in Cambridge, Massachusetts.
And so we had four chairs, and the only distinguishing characteristic of the four chairs, there were two chairs that were by the window, two chairs, not by the window. Okay, great. Hire that first employee. Other decision we have to make is like, “Okay, well where do people sit?” And so this one is like, “Well, it feels unfair to us that we would just get the window seats or whatever.” It’s like, “Okay, we believe in a flat organization, why?” And so what we did is we said, “Oh, we’ll just do a lottery.” And it’s not like, “Oh, whoever pulls the thing gets the window seat.” Whoever pulls gets to choose from the remaining seats, because different people ascribe different utility to different seats. It’s like maybe you don’t want a window seat. Maybe you like the darker corner, whatever it is.
And so we did that with employee number one. Then we did that with employee number two, and we did it every time we hired a new employee, we would do a seat shuffle. We’re going to do the lottery, we’re going to go down the sequence, and of all the seats, that was the algorithm. And our investors were like, “Oh, that’s cute and that’s funny. It’s not going to work at 15, 20 people. It’s not going to work at 50.” It worked at 25 worked at 50, worked at 100, worked at 200. We updated the algorithm, we did some local optimizations, which we said, “Oh, there are some groups that should not be next to other groups because engineers like quiet time and salespeople don’t, and so we’re going to try to optimize a little bit.” So it’s not completely random anymore. Now it’s like within this part of… Well, we still would do it. Then we said, “Okay, we’re not going to do it for every hire, we’re going to do it every quarter.” And that stood, by the way, for hundreds of people, that mechanism still worked. It’s like it’s… Yeah.
Lenny: Oh my God.
Dharmesh Shah: So the moral of the story is, that one decision of simplifying, can you imagine the amount of politics we avoided? In corporate America, the amount of calories spent is like, “Well, your office is bigger than my office. Your desk is one inch higher or one…” Like, “No.” It’s like all of that just went away by virtue of that one decision. And we don’t even know, it’s incalculable, honestly, how much grief that saved us. And then there’s all other side benefits to it in terms of people getting to know each other better, better chemistry, there’s lots of good upside, but just the simplicity is almost always a better answer.
Lenny: I love that everything often comes back to some algorithm you’ve developed or this N-plus-one approach of, “Let’s just keep going until something goes wrong.” This is amazing. I’m excited for many more of these hilarious, amazing examples of how you did things differently. You’ve been talking about simplicity. The opposite extreme is, you could say, entropy, and someone told me that you have this concept that every leader in every business is fighting the second law of thermodynamics, which I believe is entropy.
Dharmesh Shah: Yeah. Well, the actual law is within a closed system, entropy increases over time.
Lenny: Okay.
Dharmesh Shah: Yeah.
Lenny: Okay.
Dharmesh Shah: We don’t have to get into the [inaudible 00:37:50] of this.
Lenny: No, I love this.
Dharmesh Shah: I’m not a physics guy.
Lenny: No, we’re getting to the first principles of everything. Talk about that. What does that mean when you… And what does that look like when you’re running a business?
Dharmesh Shah: Okay, so in lay person’s terms, the second law of thermodynamics, I’m paraphrasing here, is that, “Over time, unless you intervene, everything goes to crap.” I’m paraphrasing. Which is, essentially, the amount of disorder and randomness in a system is going to increase over time. You see this like when you break an egg, the molecules don’t come together and form an egg. That doesn’t happen. It’s always the other way around becomes disorderly and more random. That’s the second law of thermodynamics. And this happens in companies as well, and it happens in code, almost every level of abstraction, you will see some variation of this. And so the way I think about this is that in the early stages of a company, you’re essentially fighting to survive. Just trying not to die. That’s the thing. Okay, priority one, don’t die. In the second phase of a company, you’re trying not to stagnate. But it’s like you still want to be able to drive growth and be able to do things like, “Okay, we managed not to die, because stagnation is essentially death, so we’re still trying not to die.”
But then the third stage is you’re fighting complexity. And that’s the thing that you will crumble under your own weight over time, absent some external intervention. And you see this happen all the time, and it shows up and manifests in different ways, which is, everything gets more complicated. You need more layers of management, more headcount, more this, more that, more everything. Everything becomes harder. Margins go… It’s like just bad things start to happen, you become slower. And that’s the eventual… It’s a slower death, because you’re at more scale now, but it’s still eventual death. Complexity does kill companies. Maybe not as quickly as other things, but much more reliably than other things.
And so this is why it’s never too early to plant the seeds of simplicity, put them in there, make that part of the culture of the organization. And we have a thing that are part of our culture and guiding principles around what I call “fight for simplicity.” It’s literally those three words. And the message we’re trying to convey is, “Simplicity is worth fighting for.” That’s thing number one, it’s important, but the other one is that it requires fighting for. It does not happen, and it will be a fight, because the universe is working against you. And it will take calories to fight for that simplicity, because everything, even well-intentioned people, will introduce complexity because that’s the natural way of the world. We want more tiers in our pricing, we want these knobs and dials in the product, we want these more… It’s like all of it tends towards the second law and entropy increases. So anyway.
Lenny: I love this. The algorithm for seating is a good example of this. Is there any other examples, either in the product or strategy, where you pushed for simplicity and that ended up being right?
Dharmesh Shah: Yeah. So on the product side, in the early years, and Brian gets credit for the actual implementation of this. So we had a relatively broad product, and we can talk about that, pros and cons of that even in the early years, but we were solving for simplicity, and we got this from Apple, we can talk a little bit more about genesis of this, but we had a rule in the HubSpot product as the product grew in those early years that every time you added what we thought of as a knob or dial, called a feature, you had to take one out somewhere else. That’s a net amount of… And this is a very coarse measurement. It’s like, “Okay, well, not every radio button, checkbox, drop-down, whatever menu item that you put in your nav is necessarily equivalent, but it’s better than nothing. It’s better than having no constraints.
And so once again, this goes to the binary thing, it’s like it just at least forces you to think about it, versus the… The other mistake I think people make in product all the time is that we measure the cost of a feature based on the, usually, or even a new product, based on the cost of implementation. That’s the first-order thinking, it’s like, “Oh, it’s going to take six months to develop, it’s going to be Y engineers and Z designers or whatever.” Second-order thinking is thinking through the maintenance of that feature. It’s like, “Oh, it’s not just the first version that goes out, it’s like now we have this code base and we have to support and improve or whatever.” I get that. The third-order thinking, which I think is the most nuanced, and it turns out to be the most important, is the other costs that complexity adds.
Okay, we’ll come back to this. So when you go from product number one to product number two, it’s like, “Okay, product number two is going to cost us this much to develop, it’s going to have this risk associated with whether it’s successful or not, all these things. So there’s going to be this carrying cost for product number two.” What companies don’t think through is that that is… And the maintenance of that, it’s like, “Oh, we’re going to need a team to maintain that new product.” No, what actually happens is that now when you go from product one to two, you have added dimensional complexity to your business.
What I mean by that, it’s not an incremental increase like, “Oh, we went from one to two,” it’s like, now every decision you make has to be made through the lens of, “Now we have two products. We just hired an engineer, do they work on product number one or product number two? We’re going to launch a marketing campaign. Do we spend five minutes talking about product number one and two minutes talking about product number two? How do we do anything?” Every chart you look at in terms of the growth, revenue per whatever, it’s all of it. Now, every chart that you ever had now has to be sliced by product one and product two in order to really capture that precision.
And so now you have this new dimensional complexity that you just hadn’t planned on. And this applies, once again, every level of abstraction. So everything you do in your business should factor in the long-term cost of that complexity. And it should be worth it. Of course, you need, and I can make the case that you need to build product number two and product number N, N plus one over time, but you should be mindful about the cost of that complexity. [inaudible 00:43:39].
Lenny: Is there something you’ve done to help operationalize that? Because I imagine everyone’s like, “Yes, this is great. Don’t do too many things. Simplify.” Hard to do when you’re like, “Oh, we got to drive growth. We have this awesome idea, we have this opportunity, we have a competitor.” I know that this is part of the culture, so maybe that’s the answer, but how do you keep people to actually this principle?
Dharmesh Shah: So it’s really hard, I’ll say that. And things that have made it easier, it’s a little bit idiosyncratic to HubSpot, so part of what has made it easier for us to keep things simple is the early constraints we impose. And we didn’t know this at the time. So because we’re building for SMB, we have a freemium product. All right, well, when you’re doing that, there is literally a limit to how much complexity you can add, because we don’t have an army of people that are going to spend in 18 months implementing something. We have to be able to support a product that’s free. That’s hard to do. It has to be simple enough that someone can get value from this thing. And so there’s these self-imposed things.
So the lesson I would carry away is come up with systematic ways and mechanisms, as Amazon would call them, of putting guardrails and constraints in to the degree that you can. Because you can put words on a page and try to infuse it in the culture, it’s like, “Oh, we believe in simplicity, we fight for this,” and you can have meetings and you can get up in all-hands meetings and reinforce that, which I encourage you to do, but a really, even a reasonably well done system will outbeat any other mechanism that you can try to put it [inaudible 00:45:14]. Those things deteriorate over time. It’s hard to scale them. So I’m a big believer in systems and imposed constraints versus-
Lenny: I can’t help but talk about this SMB piece, because I do a lot of angel investing, and classically, companies that focus on small businesses are very challenging, because getting to small businesses at scale is very hard. They’re often more old school, not early adopters. You sell one and that’s just one small customer, and they’re not going to upgrade and roll it out to this large group, and they’re not going to increase spend very quickly. For founders that are exploring SMB or trying to win something SMB, any advice other than just, maybe the advice is, “Probably don’t do it, it’s very hard”?
Dharmesh Shah: No, my advice is absolutely do it.
Lenny: Amazing. Great.
Dharmesh Shah: A, because it’s hard, and we’ll talk about that, but B, okay, so I’ll tell you this quick story in terms of how we chose SMB. And so my co-founder and I met in grad school, and he had done software, he was on the sales and marketing side, I was on the product engineering side, didn’t know each other before grad school, and we met there. But one of the things, as we were canoodling on possible ideas of working together on the startup, we said, “Oh,” we had done both an enterprise software before, the opposite of opposite of SMB. And there are definitely challenges with enterprise software. One of the biggest one is that life as an enterprise software company sucks as a startup. Sales cycles are super long, feedback loops are super long, you have to have sale, there’s all this kind of… You have revenue concentration, which means you don’t completely own your product roadmap anymore because whoever’s writing the biggest check will ask for these three things and it’s really hard to say no, especially as a startup, all these things that go along with that.
It’s like, “Okay, we sort of know that game.” At the other end of the spectrum, you have consumer startups. It’s like, “Okay.” The problem with consumer startups is they have very bi-modal outcomes. It’s like either you can be super successful, the next Meta, the next Google, the next whatever, or it’s going to be to zero. It is very rare that you have, “Oh, you made a few million bucks or whatever.” Because either worked, because this is a massive market. The case for SMB is you have almost the best of both worlds. You have the nicety of enterprise, which is, “Oh, I can solve a product and people pay me money for it. I don’t have to be advertising subsidized, I don’t have this bi-modal outcome. I can build an incrementally more valuable business that’s measurable.”
So we’re “Oh,” we went from a 100K in ARR to a million ARR, to 2 million ARR because we’re a business software company. But it has the benefit of consumer because there’s millions of them out there to sell to, and there is no revenue concentration. You do control your roadmap. You have very short feedback loops. You can try pretty much anything that you want because there’s… Like, “Oh, we have 50 customers. If we do this and screw it up, it’s like there’s 5 million more.” Don’t intentionally screw it up, but you can experiment, you can take on risks. And by the way, when HubSpot started SMB, it was literally a hundred times harder to succeed in SMB. And we could not raise capital. It was so hard because the common argument was there literally is exactly one company in the history of software that’s made a global brand, billions of dollars of market that was into it. No one else had ever created SMB-focused software company before.
And we’re like, “Yeah.” It’s like, “We know.” But the nice thing about it is once you figure the physics of it out, because it’s so hard to do, you end up, and this has played out, now there’s lots more proof points. There’s Shopify, there’s a bunch of companies that have succeeded in SMB. But one of the things about SMB is that in the software market… By the way, all I know is software. That’s the scope of my experience, is that in software, there’s what I call reverse gravity. That over time, the market will always pull you up. Because what ends up happening is that smart founders and smart management teams will look at the numbers and say, “Oh, well, as it turns out, our bigger customers stay with us longer, pay us more, often have higher NP, all these things.”
And so if you allow yourself to be led by that, which seems like a reasonable, smart thing to do, you will get pulled up in the market. And because of this reverse gravity, almost every company winds up being, every successful software company, ends up being an enterprise software company over the full list of time. That’s what you see. The byproduct of that reverse gravity is that every software company ends up being an enterprise software company, so you end up competing with literally everyone. In the SMB world, you’re only competing with the people stupid enough to stay focused on SMB, or try to do that. It’s a harder thing to accomplish, but once you figure out the physics of it and make it work, it’s a much more sustainable, much more fun model. So I encourage, particularly startups, and maybe you fight reverse gravity for some period of time, and maybe… But at least start there, try to make it work, until it gets really, really, really painful, and then try some more. It’s like there’s just so much value in SMB. It’s that idea.
Lenny: Okay, so this theme has come up a bunch already in our conversation. I asked Chris Miller, previous podcast guest, what you’re amazing at and what I should talk to you about. And he said that you’re very good at zigging when other people are zagging, and zagging when other people are zigging. And I think that’s already come across in our conversation. I guess, is there anything else along those lines that you think might be helpful to people, in terms of how you think about that, or examples of that in action?
Dharmesh Shah: So this comes down to high-conviction, low-consensus bets. And so we had like three things that we were zigging… And by the way, it’s not just me. My co-founder and I both have that zig versus zag. And we ask ourselves this to this day, it’s like, “Okay, well we understand this is the way the world moves, this is the way it’s normally done, but have we considered this completely? And I’m not saying we should do it, but have we considered this alternate path?” So I would encourage you, at any stage of the company, you don’t necessarily have to do the zig versus zag, but you should at least know what the zig would’ve been and have talked it through. So that’s thing number one.
Some of the best startup advice I’ve heard and I’ve ever given is, “Startups should focus on one thing and be really, really exceptionally world-class better than anyone else at that one thing.” And one of our early zigs is, “We are going to do exactly the opposite of that.” So from year one, HubSpot decided to build an SEO tool, web analytics, blogging tool, content management, all of it. And every one of those categories that we were building a product in, literally this is year one, had great products with great companies behind them. And so it’s like, “Okay, well, why would you do that?” And-
And so it’s like, okay, well why would you do that? And the reason is that the one thing we wanted to be good at was solving for the actual customer problem that existed, and the customer problem that existed was not a dearth of tools, lots of great SEO blogging, everything, all the tools existed, but SMB specifically did not have the wherewithal to put all those pieces together. So you and I, it’s like, “Oh, I can throw a website up. I can put Google Analytics on it. I can put a Wufoo Typeform format. I can do all these things. I can wire doc. No big deal. Why is this so hard?” But for most people, most SMBs, that’s a science project, and so we said, okay, in order for us to solve the actual customer problem, we have to solve the actual customer problem, even if it’s uncomfortable, even though it’s contrarian and every instinct and every advice we’ve heard or given says we should not do this. We should not go that broad.
But once we make that decision so that the other lesson learned, it’s like, and we said this in the early years and it’s gotten better since, let’s say you do this, and this is the all-in-one approach, we’re going to have this broad-based product. It’s like, “Okay, we’re going to be many miles wide and only so inches deep.” One of these, and this goes back to our very systematic thinking, is like, “Okay, we’re going to measure each of those individual product categories that we’re now playing in. Are we in the top three in the market in that category? If the answer is yes, that means we invested too much in that category.” We should not be in the top three because our value proposition is not that we are the top three blogging tools or one of the top three web analytics tools. We have the top three, whatever, our value proposition is that everything works so well together that that being in the top three and having those right features or whatever doesn’t matter.
Doesn’t matter as much as the all-in-one. And that means we over-index on one individual category by virtue. That was one of our heuristics for being able to tell that we were over-indexed on one particular aspect of it.
Lenny: That is hilarious. Rarely do I hear we’re in the top three and we’re doing the wrong thing. We are too good at this.
Dharmesh Shah: And by the way, and we changed that over time in the startup world, and now I would argue that most of the categories we play in, we are, as the surveys would tell you that we’re in the top three in each of those categories now, but at the time when we are very, very resource constrained, we were trying to be disciplined and focused.
Lenny: We’re just racking up all these very contrarian approaches to company building. I imagine people listening to this might feel like, “Okay, cool, I’m just going to build a bunch of stuff. This is great. I worked for HubSpot.” Do you have any heuristics for when it makes sense to go wide and not deep that you share with founders? When do you think it makes sense versus now you should actually follow the common advice here?
Dharmesh Shah: Yeah, so it comes down to what problem you’re solving. So I think one of the mistakes I think founders make is that, especially product oriented founders, is that we fall in love with the solution instead of falling in love with the actual problem, and you need to fall in love with the problem and putting that constraint on yourself forces you to understand and at least label and define what problem you’re solving because oftentimes we can describe our product or whatever, but we struggle sometimes as founders to describe the actual problem. So if you can make the case, which HubSpot did, it’s like, okay, it’s not like we immediately jump to that like, “Oh, we have this epiphany. We’re going to build the all-in-one, do all the things kind of thing.” We’re like, “Oh, in actually talking to customers”, we considered doing it like, “Oh, we’ll just do this marketing piece.”
We’ll just do this or whatever. It’s like, no, that’s not their issue, their issues it’s not that these things don’t exist, so if you, as a result of talking to customers and understanding the problem deeply, you come to the conclusion that you need to do more than just one thing, but then you should force yourself to have a discipline that we did, which is, here’s why we’re going to go a little bit broad, and here’s how we’re going to make sure that that broadness doesn’t kill us. You have to have the self-imposed constraint. You can’t have the best of both worlds. You can’t say, ” We’re going to be all in one, and oh, by the way, we’re going to be the best.” No, you can’t be the best. You can say, we solve this particular problem the best, but we’re not going to have the best product along any of those dimensions.
Lenny: Amazing. I love this advice.
Imagine a place where you can find all your potential customers and get your message in front of them in a cost-efficient way. If you’re a B-to-B business, that place exists, and it’s called LinkedIn. LinkedIn Ads allows you to build the right relationships, drive results, and reach your customers in a respectful environment. Two of my portfolio companies, Webflow and Census, are LinkedIn success stories. Census had a 10 x increase in pipeline with the LinkedIn startup team for Webflow after ramping up on LinkedIn in Q4, they had the highest marketing source revenue quarter to date. With LinkedIn ads you’ll have direct access to and can build relationships with decision makers, including 950 million members, 180 million senior execs, and over 10 million C-level executives. You’ll be able to drive results with targeting and measurement tools built specifically for B-to-B.
In tech, LinkedIn generated two to five x higher return on ad spend than any other social media platforms. Audiences on LinkedIn have two times the buying power of the average web audience, and you’ll work with a partner who respects the B-to-B world you operate in. Make B-to-B marketing everything it can be and get $100 credit on your next campaign. Just go to linkedin.com/pod Lenny to claim your credit. That’s linkedin.com/ pod Lenny. Terms and conditions apply.
Okay, I’m going to go in a totally different direction for now and see where this takes us. I want to talk about flash tags.
Dharmesh Shah: Sure.
Lenny: So this is something that you came up with that came from a problem of you’re founder, people come to you for feedback, you give them some random thought, and they treat that as gospel and they go do it and spending months fixing something that you’re like, oh, it’s just a thought. Talk about flash tags.
Dharmesh Shah: And this is exactly the problem. It’s not just, once again, most things exist at every level of abstraction. So anyone that manages people, anyone that leaves has this exact same issue. It’s the megaphone issue. Someone will pass you in the hall when we had offices in halls, they’ll ask you a question, and as it turns out, this is particularly true of founders. I have an opinion on everything, literally everything. So if you ask me a question I will have, it won’t be thought out, it won’t be right, but I will have one, right? And so the challenge is people will take that and over-index on what was an opinion, and because it’s inefficient to qualify every time someone asks you something, every time you put an email out there and say, “Oh, by the way”, this is just my opinion, you can put IMHO or something like that every now and then, but then people say, “Oh, that’s just politeness”, or something like that, right?
Okay, so here’s what flash tags are literally, and I live most of my life through email, but I do this in other media as well. It’s like the hashtag that we all know, but it’s a discrete set of them, an escalating set of what I call the dying on the hill spectrum. So here’s how it goes. The bottom hashtag is hashtag FYI, and the definition of the FYI is, “I came across this interesting blog post”, or “I found this news, or whatever, just letting you know, no response expected”, fine, go on with your life. Next level up is hashtag suggestion. “I have this thought that came into my head because I listened to the episode on Lenny’s podcast with Sam from Microsoft or whatever, and this feels like it’d be a good thing to explore, but hashtag suggestion means I still don’t expect a response.”
“You don’t have to do the thing I’m asking, but it is something that I would do if I were you, or at least I would consider if I were you.” Okay, next level up is hashtag recommendation. Hashtag recommendation is, “I’ve thought about this a lot. I’ve done some research, I’ve done some digging, I’ve done some soul-searching. I would do this.” It’s still not a, “You have to do this, but if you decide not to do this, I would appreciate a response in terms of why you don’t want to do this, what you learned in the process of exploring this particular thing, other alternatives or something.” So it’s like a response expected, but still not a mandate. And then the last one is hashtag plea. “I beg of you, we don’t have mandates at HubSpot. I have thought about this so much to my core, in my soul, I believe this is the thing we should be doing. I’m going to plea with you to please just do this.”
Still not a mandate, by the way, and so now it’s commonly accepted. It’s on the Wiki, and I use this multiple times every day. Other people at HubSpot use it. It is remarkably effective because self-descriptive hashtags are searchable, but you can go back, and it’s like, “Oh, I want to know the last five hashtag recommendations that I made, how those turn out?” There’s just so much goodness that comes out of it. It’s not hard to understand or describe.
Lenny: It’s really interesting that none of them are just do it. You say that there’s no concept of there’s no mandates, essentially. Why is that? And do you feel like people just treat plea as, “Okay, I actually just really have to do this?”
Dharmesh Shah: So we’ve always had, as part of early manifestation of HubSpot culture, we believe in autonomy. Lots of companies believe in autonomy, and we try to carry that idea as far as we can, and for the most part we do that and that, and we hashtag plea as part of that. This is like we trust our people. We try to hire the best. We expect them to learn, and we still want them to have the discretion to be able to make hard decisions, and we call them DRIs. We didn’t invent that term directly responsible individual. And we’ll talk more about, we should talk about more about decisions and how decisions get made but the reason I do hashtag plea is that it’s a soft way of a mandate without being a mandate. It’s like, okay, and by the way, the number of times that happens is I can count on the fingers of one hand and still have fingers left over.
That almost never gets to that because either I talk myself out of it. It’s like, okay, and this goes to the, is this a hill I’m willing to die on? Right? It’s like, okay, I get it. I may disagree. Is it the end of the world? Is it really going to do that much harm? And is the cost of imposing that founder card go do this? It does happen just not that often.
Lenny: Awesome. And we’ll link to that. There’s a post you wrote about how to use these.
Dharmesh Shah: It’s flashtag.org If you want a quick way just to get to that blog post. I believe in domain names and shortcuts because I believe in simplicity.
Lenny: You mentioned decision making. There’s something interesting there. Talk about how you make decisions at HubSpot.
Dharmesh Shah: Like most organizations, HubSpot historically has made decisions poorly. It is just all a spectrum. I think decisions that were the hardest things because it’s hard to strike the balance, and we’ve evolved it over. I think we’re much better than we were in our earlier years. The things we get wrong, most people get wrong, and so I’ll tell you the things we get right now is that, so we’re very data oriented. We like to make data informed decisions, but data doesn’t make decisions, people do. So that’s like thing number one. It’s like you have lots of geeky founding companies or whatever. It’s like, “Oh, we have a chart that says this, so we should go do that”. That doesn’t work, doesn’t work outside, data’s awesome, and we also have one position that we designate to make the decision, and that is not necessarily the executive over that team.
It’s like, okay, well, and they literally own that. And we didn’t invent this idea either. This existed long before HubSpot, but it actually works. It’s like you have to decide who’s going to make a decision, and that is almost the number one decision to make, and it’s like, okay, don’t pick people you don’t trust to make the decision, and so that’s kind of thing number two. And the biggest mistake, and we used to make this much more than we make now, is, and Amazon has stated this really well, theirs is called disagree and commit. We have a slight twist on, I’ll tell you why we twist it a little bit. Ours is debate, decide, unite. That’s our kind of phrase that we use in HubSpot culture, but the idea here, the core premise of it is that once you make a decision, getting alignment around that decision, because there will always be at any amount of scale, people that disagree with that decision because they were debating the other side, they would of picked option A and we ended up picking option B.
But it’s extremely important that even when a decision gets made and we decided to do something that’s contrary to the thing that you believe and would’ve debated for and did debate for, you lost on that particular one, fine. Hard to do, but so important, so valuable, and I’ll tell you why we tweaked the Amazon one. So the disagree and commit sounds a little bit harsher than we like, which is out. It’s fine for you, disagree, but then F and commit, right? It’s like, okay, I get that. I’m biased. So the debate is, okay, so let’s have open debate. Let’s get all the options on the table. Let’s actually make the decision, and then it’s unite around that decision, not just, it’s like, okay, let’s come together around that, and that sounds more Hubspotty so anyway.
Lenny: It is a lot more friendly. I was thinking that as you were talking, debate, decide, unite versus disagree and commit, and then there’s have backbone in that one.
Dharmesh Shah: The other thing I’ll tell you, this is partly HubSpot thing with partly a me thing, and this will not surprise you at all, is I like to think systematically around decisions, and so I’ll give you an example. The question I ask myself when faced with a decision or a thing that I’m trying to do is, “Okay, if I could write Python code of Python function to come up with an answer to this, what would be variables or the coefficients”, okay, let’s say you’re not, if you were making Excel spreadsheet to make this decision, what would the columns be? All right, you don’t have to decide the weights, but at least decide the factors. Then the next step is like, okay, here are the factors that I think matter in the making of this decision. I’m going to decide between these seven options. Here are the things I’m going to look at.
You would do this for picking a school to go to, or it works and you don’t have to take it to the nth degree, but the exercise of going through it’s like, “Okay, well step one, if I can just identify the factors that I think should impact this decision, that’s an 80% win right there.” Not because it’s just in head. If the team works collectively to say, “Okay, here are all the factors that I think should contribute to the making of this decision.” Then number two is, okay, we can’t assign exact weights, so we can’t write the Excel formula just yet or the Python code, but can we at least stack-rate the factors? This factor is more important than that factor. Now you get a closer sense of approximation. You’re still not going to use the spreadsheet to make the decision, but the human making a decision has much better information they had than pre-going through that exercise.
One of the mistakes people make around decisions, and Amazon has written about this, I’ll give you my take on it, is that the calories you spend on a decision should be proportional to the consequences of that decision, pure and simple. He calls them one-way doors versus two-way doors, which is a binary heuristic, which is, I love the elegance and simplicity of that. I’m taking it one level further, which is, it literally should be proportional. It’s like, okay, if you have a really, really big decision that’s going to be expensive to change, really, really, but a lot of calories making that decision, you may still get it wrong, but it’s worth it. And it’s not just around decisions, around activities and bests and things that you’re going to do is that the calories you spend should be proportional to the outcomes and what you’re seeking and the value of it.
And that’s one of the things I try to apply to my life. It’s like, okay, my geeky nature wants me to go nine miles deep and start writing code and do this thing or whatever, but I found that that one simple thing is like, okay, is this a first-order approximation where I can just identify the factors or it can, and the other thing that helps by the way is, and we’ll talk about this. So one is my default position on most things, and I don’t mean this in a negative way, is, no. It’s like, “Oh, shiny object. Should I go do this?” “No.” And then I have to force myself like, “Okay, I need reasons to say yes.” I have to force myself because it’s easy to say I love new things, I love new people, I love all of that. So I have a, sorry, but no post on my blog that life-changing for me.
And so the idea there is that this goes back to other first principles is time is with a few constants. So every time you say yes to something, by definition you’re saying no to something else, and so this goes back to the product heuristic we were using in the earlier years of our class. Like, oh, when you put something in, you have to take something out because we want to keep the complexity roughly constant if we can. Same thing, so it’s okay to say yes to things, but then you have to force yourself, based on the calories that you’re committing to say, “Okay, well, what am I going to take out of my schedule, out of my life, in order for me to be able to make room for this thing?” Because as it turns out time, I have not been able to, despite my best attempts, bend the laws of space and time. So how do I make room for this new thing that I want to say yes to?
Lenny: I love the way your mind works. I’m excited to talk about culture and how this connects and the way you think about culture and through all these frameworks and lenses. Before we do that, you hinted at this idea of how you pick what to work on. And I know you have kind of an interesting framework for how to decide what ideas are good and what ideas are worth investing in. I think both as a founder and a product leader, how do you think about deciding what is worth investing in?
Dharmesh Shah: So imagine the Excel spreadsheet again. So let’s say you’re a value and idea for a startup, for a new product, a new feature, whatever happens to be, I think this applies. So the first thing you should look at is, and I measure things on a scale of zero to 10 because I like quantifiable things. I like to be able to multiply X by Y. I don’t like letter grades, pet peeve of mine, but anyway, so put a number on it. It’s like, okay, if this thing were successful, so let’s say we’re just picking a new startup idea. I’m considering starting this new business. If it were successful, what could it be? What’s the magnet to the outcome? However you decide, it could be impact, it could be revenue, it could be market cap, it could be portfolio, whatever it is, it doesn’t matter. I mean, it matters, but not relevant for this discussion.
So that’s the potential. Thing number two is probability of success. And now I’m going to pause here and tell you the most common mistake people make is they think through the probability first without actually forcing themselves to think through the potential, and here’s what happens is if they apply a filter that says, “Oh, I only have a one in 10 chance to pull this off, there’s only a 10% success rate or whatever”, not going to do that. Well, if it’s a one in 10%, one in 10 chance at 10 billion, it might be worth it. It might be maybe not. Sometimes it’s based on stage of life and things like that in terms of risks you can afford, but for instance, so by the way, and when I say probability, that’s the biggie side. I mean, you can do it multiple ways. You can say, “Oh, here’s the X percent chance you can do in statistics as an event entry.”
It’s like, oh, here’s path A, there’s a Y percent chance that happens, and here’s what the outcome would be, Z percent chance or X. Anyway, here are the popular things, and you come up with what’s called the expected value. Anyway, so let’s say we just look at two branches like, “Oh, there’s a fifty-fifty chance that I’m going to make 10 million.” It won’t tickle one branch just to keep this shorter. So the expected value of that is $5 million. That’s literally the expected value. It’s like, oh, you got a fifty-fifty chance at a 10 million outcome, and let’s say you have a 10% chance at a hundred million dollar outcome or 500, whatever it is. So mathematically, it’s like, okay, well, you should at least consider that other one, even though the chances are lower because the expected value is higher. Okay? So that’s thing number one. Anyway, we’ll get through the rest.
So start with the potential outcome, then look at probability not the other way around because then you’re going to apply that mental filter and throw out ideas that may not have been worth throwing out. The third thing is what I think of is either passion or proximity. Do you care about this actual thing that you can actually bring to bear? And then the fourth is, since I need a P word, it’s like prowess. Do you have some assets? Something that makes you uniquely positioned? Either if you’re building a new product, you have existing lines of code that you can reuse, you have a market that you already have access to, there’s something about this particular idea that you have an unfair advantage pursuing, and that impacts your probability of success, and this is the other reason. You need to look at all those things, and once again, no one thing is the deciding factor.
You’re not going to just pick the thing that has the high expected value to take the potential times, whatever. It’s like you have to sort of look at all of them and look at your own situation, but it helps to think through that, to have the discipline to say, “Oh, how big could it be? What are the chances? How much do I care about solving this problem? Is this something I can work on for two years, 10 years, 20 years?” And finally, “Why me? Why me? Why my company? Why would we succeed at this? Why it’s our chance of success, even low, higher than the rest of the world?” Like why?
Lenny: Okay, so I’m already picturing the spreadsheet here. For every idea you have, you basically have a column, potential probability of success.
Dharmesh Shah: Yes.
Lenny: Proximity, you said? Just like how…
Dharmesh Shah: Particular passion.
Lenny: Okay, passion.
Dharmesh Shah: Yeah. I started with passion, and I have an issue with passion as a word because it’s ambiguous, because we talk within startup circles. It’s like, “Oh, pursue your passion.” Sure, but it’s like, you know what? Most of the companies that exist today that are even successful ones with awesome founders, they didn’t necessarily pursue their passion. It’s like they maybe are baffled in their scenario, particular problem, or a market segment or something like that, but it’s like it’s okay to become passionate about something as you dig into it or whatever. That’s fine too.
Lenny: Yeah, one of my favorite startups that I’ve invested in, it’s called Zip. It’s a procurement platform. The founders were not passionate about procurement when they got into the space, they just saw a huge opportunity. They thought they could build a much better product, and they got passionate about it once they started working on it and finding products.
Dharmesh Shah: This is the thing. So yes, if you can find your true calling and this is what you were meant to do in the world, and you have the opportunity to do it, great. That’s a tiny, small fraction of people. It doesn’t happen. You may not even know what your passion be, especially if you’re a first time founder. It’s like, okay, well, I haven’t seen the world. I don’t know.
Lenny: Okay. So just for people, if they’re creating a spreadsheet for all their ideas, for startups, it’s potential, probability of success, passion, prowess.
Dharmesh Shah: Yes.
Lenny: Okay. Amazing. Okay, let’s talk about culture. I know you spent a lot of your time on helping build HubSpot’s culture, create the original culture. You created this epic culture code deck. I think it’s 128 pages. So maybe just to dive into it, what was that process like to define the culture of a company and maintain the culture over time?
Dharmesh Shah: I’ll say awful. No, the beginning of it. Okay, so I’ll give you the very quick story because my co-founder and I, having one of our founders meetings or founders dinners as we do, and he had been meeting with this CEO group, and he said, “Oh, Dharmesh, I hear this culture thing is really important.” By the way, the word culture had never been mentioned prior to this conversation, the halls of HubSpot ever. And this was how many years in? This is, I want to say three or four years in. I can check the records because by the way, even Brian and I, even back when people work on offices, most of our communication was over email and Async. We’ve always been big believers in that written form, but anyway, so I can tell you whether you know where cultural was used or not, and it was not.
And he’s like, “Oh, I’ve heard”, because talking to us, “Culture is super important. Dharmesh, can you go do that?” I’m like, “Okay, Brian, all the people in all the companies, I am literally the bottom decile, if not the worst possible person. It’s not that I don’t like people, I just don’t like being around them a whole lot, and from what little I know of culture, it sort of involves people. I don’t know, but it seems like I’d be the wrong person.” But anyway, I’m a team player. I’m like, okay, it does feel like it would be important, and so I took on the task and I didn’t really know because when he said, “Dharmesh, can you go do that?” What does that mean? And so when I took it to mean, it’s like, okay, well culture actually already exists. We have a culture. It’s like whatever it is, and it’s working pretty well, companies doing well, people seem happy.
And so I said, “Okay, well, what I’m really trying to do is kind of describe the culture that’s there.” So it’s not creating culture, it’s articulating the culture, so we know what it is, and so then this is the side of me showing up again. So the question I asked myself is I literally asked myself this question, which is, “If I could write a Python function that would measure the probability of success of a new person coming into HubSpot, what would the factors be?” What are the things that if I could measure, if I knew I had a true function that says, “Oh, on a scale of zero to 10…”
If I knew, if I had a truth function that says, “Oh, on a scale of 0 to 10, here’s how they rank on this, and this, and this, and this, and this, and this,” I could approximate how successful and how good they’re going to be for HubSpot. That’s my exercise.
Okay. So I did an internal survey, and I’ll tell you why this was such an awful experience in the beginning, because this is also a good story. So once I took on this task, I sent an email to the company. It’s like, “Hey, everyone. I’m just going to be digging into culture, so I’m going to send this survey out. I just want to get an understanding so I can…” Great. And the negativity and the visceral response to that one email was literally up until that point, it still ranks in top three, worst things I’ve ever…
People felt so negatively and so strongly, and I was not expecting it, right? And I’ll tell you, the one line that really just was gut-wrenching was like, “Dharmesh, here we’re talking about culture. Next, we’re going to have like posters up on the wall about excellence. HubSpot is not the company that I thought I joined.” Like oh. It is just like the deep part of me is like I just did not know how overloaded that idea of culture was, right? And so then when you dig into it, it’s like, oh, because so many companies had done it so poorly and just didn’t walk the walk or whatever, this is why. There was a reason why there was this kind of negative response.
So anyway. So I went back to my thing. It’s like, “I get it. So I’m going to try to describe what we have, and I’m going to write this truth function.” So I wrote the slide deck, which had exactly 16 slides, and that answered this one question, which is, if I could write a Python code, what are the factors, so what are the attributes that we would look for in humans. And the reason I came up with that particular question is that when I did the survey, I did the NPS survey, which everyone knows that you would send to customers. It’s like, “Oh, on a field of 0 to 10, how likely are you to recommend HubSpot as a place to work?” and the other NPS question is like, “Why did you give that answer?”
As it turns out, here’s the thing that I’ve learned which was troubling, everyone was happy, and the reason they were happy is because of the other people at HubSpot. That was the number one reason, which is both good and bad. It’s like okay, well, that’s awesome, but that doesn’t tell me anything, right? It’s like, okay, well, that’s a self-reinforcing kind of recursive loop. And so that’s when I was like, “Okay, well, what makes the people at HubSpot the people that are at HubSpot? Let’s come up with the attributes that define what it means to be one of these people that makes other people happy.”
And so we came up with the original set of things that were embodied in that slide deck, and I called that slide deck the Culture Code deck, and the reason I called it the Culture Code deck is not because it was a code of conduct, or a moral code, or, “This is how companies should be run.” It was because if I could write code to kind of define the culture of HubSpot, this is what the code would look like. And since that point, it’s like the question I keep asking myself is, if we could have an operating system that ran the company, just the geeky part of me, not that you could do that, but what would that look like, right?
There are very few original ideas I’ve had that are any good. One of them is this thing that culture is a product, period, and that every company builds two products, one is the product they build for their customers, and the other is a product they build for their team. That’s what culture is, it’s the product you build for your team. And let’s say we stipulate that that just happens to be true. Just humor me for a minute. If you assume that’s true, then a bunch of things kind of follow from that postulate, that hypothesis, which is, “Oh, well, just like we would never build a product without talking to customers, we would never build a culture without talking to the customers of the product,” which are the people. So we do a repeat of that original NPS survey we’ve been doing every quarter now forever to figure out what the NPS of that product is, our culture product.
The other thing is like, okay, well there’s no product person in the history of humankind that would’ve said, “Oh, I built this product. It’s awesome. I’m done.” No one would ever say that, and this is the number one mistake, I think, I’m biased, I think founders make, is that they say, “Oh, we have an awesome culture, and my job is to preserve the culture.” That is not the job because the needs of your customers change, the needs of the employees change, right? The culture exists to help the great people do great things for the company. That’s why the culture is there, and the things they need later as the company scales are different. And so just like you would never freeze the lines of code of a product, you would never freeze the lines of code of a culture. Culture should be iterated on just like you iterate on product.
And there’s so many other things that kind of fall. So just like in a product, you will get feedback from the customers like, “Oh, here are the bugs,” and we actually call them bugs. It’s like, “Oh, here are the bugs in the culture.” At the all-hands meeting, so we have this survey that comes back. Once again, transparency notion. Every single survey response, it’s anonymous in terms of who submitted it, we publish it, right? The quant score of the subjective thing, everything, then we will categorize it, and then at the next all-hands meeting, we call them all-minds… Anyway, a little-
Lenny: I love that, all-minds.
Dharmesh Shah: All-minds.
And it’s like, “Okay, so here are the things that we heard.” Just like we would do for a product meeting, it’s like, “Oh, the customers, here’s what we’re hearing back.” It’s like, “Here’s the bug that you’ve identified. Here’s the ones we’re going to fix.” And when we’re going to fix some, we’re going to commit like, “Oh, you said this was broken. Okay, we’re going to do something about that.” Here are the things that like, “Yeah, I know you don’t like it that way, but it works as design,” which we do in product all the time, right? And you should have the right to do that. It’s like, yep, some people are going to disagree, and there may be any number of reasons why we think that needs to stay the way, we’re not saying necessarily forever, but as it stands, we are not prioritizing this, we are not fixing that thing that you think needs fixing for whatever reason.
So the one lesson here is, very few things I’ve learned that I think are universally applicable, that one, I think, is right, and just so much greatness and goodness falls out of that, is if you think of the people as customers and if you think of the culture as product, it takes away this kind of abstract nature of working on culture and then takes away this thing that I heard this back in the early commentary from the HubSpot team, is like, oh, well, culture is not something you really build because that seems fake, that seems inauthentic that you just go, it’s like, “Oh, we’re going to craft this culture.” No, we craft in the same way. I’m not saying we’re going to impose it, like we’re going to do this often in isolation, but we’re going to work on it, and it’s something you actually do manifest, and it’s not just up to the founders.
Lenny: Wow, what an amazing story. You’re definitely getting laughs per minute for me, and I think there’s at least one laugh per minute so far in this podcast. That’s amazing. And I think this point you just made about your job is not to preserve the culture as a leader, I think that’s really profound, because I think most people don’t think that. They think their job is, “We need to maintain our culture as we grow.” It’s all fading away, our culture’s changing, but your point is that that’s good and that’s great.
Dharmesh Shah: Bringing it back to product, I think this is an important thing. This is the second order lesson learned after you start with the culture is product, is that one of the challenges almost everyone will deal with is that you have these things… And so by the way, just because culture is product does not mean there’s not a sense of having what some might label as core values, right? Because we have this in product as well. There are core values that the product believes in, like we believe in our product, that we’re going to hold constant through it, right? Like, “Oh, we’re building for SMB. Want it to be easy to use.” There’s constraints that we put on ourselves that are intentional that even though we iterate, we’re not going to iterate away from these things, kind of that high conviction bet thing again.
But not everything that’s part of your product is a core value, right? So not everything that was part of your culture in the early years. So transparency is one that has endured over a long time, right? It’s part of our core value, and we preserved it even though it was troublesome through the IPO. I shared that story. And then there are things that we changed that were part of like, oh, we believe this meritocracy thing in a flat organization, but the manifestation of that was a bud with lack of titles, right? And so we changed that. It’s like, “Okay, well, there are things…”
And here’s the other thing. It’s a mixed metaphor problem because I haven’t been able to come up with a good metaphor yet, but there’ll be things that you have that are kind of core… And the way I describe it right now, I’m just going to put it out there, I’m kicking myself for not coming up with a better metaphor, is like federal versus state law. So there are federal laws we have here in the United States that says, “Okay, here are the things, and these are the inalienable rights of all citizens of the United States. These are the things that are core.” But then we have state things, right? It’s like, “Oh, for this particular state on these particular aspects, they can optimize for their particular state.” It’s like, “Yes, we understand this is the corollary, but for our state, we’re going to do this.”
So we have something similar at HubSpot. We have a core set of things that are part of the culture, part of the thing that’s like, “Okay, here’s what we believe,” and then there are things that the individual groups, so our head of engineering could come back, it’s like, “Yeah, I know our overall, like the founder’s leaning is X, but for my particular group, that doesn’t really make sense. We’re going to do this.”
So I’ll give you an example. So one of the core things from the early years of HubSpot is like we don’t care about what hours you work, like from what time to what time. It’s like it doesn’t matter, like literally does not matter. We didn’t track it, we don’t care. But then over time, it’s like, “Well, we have a support organization,” and that organization, as it turns out, does matter to our customers, right? Like we need to have coverage and things like that. It’s like, “Okay, well, that sort of makes sense,” right?
But then you sort of have to, A, pick your battles, but then fight back. For instance, transparency, there have been times where individual org leader’s like, “Okay, Dharmesh, I know you believe in this transparency thing and Brian believes in this transparency thing, but it’s inefficient for someone that is a SDR in the sales organization to have access to the high-order financials. It’s a waste of time. It’s a distraction.” That was the argument. It’s like, “You may be right, but the binary decision, the fact that we value transparency so much, we’re going to…” Yes, maybe it is a distraction, fine. And the other one is like, “Oh, I don’t want my team to have it because I just don’t…” You know, we were hiring so fast and it’s like, once again, that’s a lot of trust to put into people or whatever. It’s like yeah, but then one of the other lessons is cultures over time become self-fulfilling prophecies.
So in the early setup, big lesson learned. So when I put the first version, the public-facing version of the Culture Code out there, and I got feedback from the team and like, “Dharmesh, you say this in the Culture Code deck, and it’s just not true.” It was not true enough for us to kind of state that as a thing. And so the thing that I did is in the deck, I have what I call these little liner notes. It’s like stuff going on. It’s like, “Oh, this is not really true yet, so it’s more of an aspiration. This is how we want it to be, but we’re not walking the walk yet on this thing.”
So I didn’t take it out because I wanted it to be an aspiration, and here’s the thing that I had not thought it through. I cannot take credit for it. What happened over the fullness of time is that those things that were aspirations became increasingly true by virtue of being in the deck because as people join, it’s like, “Oh, that’s something we aspire to, so I’m going to pretend that’s actually the culture of HubSpot,” which is exactly what we want. So it’s okay to say things that are aspirational. Just call them out. Don’t present things that are not true, but they’ll back off of the things that you would like to be true.
Lenny: Wow. There’s so many lessons there. It reminds me of Airbnb, actually. I was there when they created the original, there’s these six core values of Airbnb. One of them was simplify, it turns out. And then, I don’t know, some number of years later, the founders realized, “We’re not actually good at this,” and values should reflect who you are not aspire to be, and I like this middle ground you found of like, “I’m going to put in here, but I’ll note this isn’t necessarily who we are.” We didn’t do that. That would’ve been a smart move. So they removed that core value. They’re like, “We’re not actually this,” and going from six to four. There’s another one they removed. So I was really impressed because we’re like, “Okay, wait.” Because it comes up in meetings. They’re like, “Simplify. No, this is actually really complicated.”
Okay, so what I’m going to do, I have one more question about AI, and to make this work timing-wise, I’m going to cut the lightning round, so hope that’s okay.
Dharmesh Shah: Totally fine.
Lenny: Okay. I know AI is very top of mind for you.
Dharmesh Shah: By the way, this is my diabolical plan to have to skip the lightning round because that’s the thing I’m not good at, like I’m not [inaudible 01:31:05]. Anyway, you know I was dreading the most.
Lenny: Playing 4D chess here. You knew exactly the timing and you’re like, “I will take this long for all these answers so it’ll perfectly play out.”
So I know you’re spending a lot of time on AI. It’s on many people’s minds these days. You built this product called ChatSpot. Is it still called ChatSpot? Is that the product you changed the name for or that’s still ChatSpot?
Dharmesh Shah: It’s still ChatSpot.
Lenny: Still ChatSpot. And I know that you’ve handed that off now to a team actually that’s working on this.
Dharmesh Shah: Yep.
Lenny: It’s not like just the entire mesh project anymore.
Dharmesh Shah: Yep.
Lenny: So I don’t know exactly where to go with this, but just why is AI so top of mind? And where do you think people should be spending time? Where do you think things are heading? What comes to mind around AI?
Dharmesh Shah: Yeah. Obviously, I’m not alone in my excitement, but I will say this. So I’m about to hit the 30-year anniversary of my working in commercial software, right? So I’ve been at this for a long, long time, and that’s all I’ve ever known, that’s all I’ve ever done. And the last time I was this excited was when the web first came along in the mid to late-’90s, and I’m like, “Oh, this literally changed everything,” right? It’s like relevant to everything. It’s like literally everything is going to be impacted by this thing called the internet. And I was in my early twenties, right?
I’ve been here for mobile, I’ve been here for social, I’ve seen lots of things that people got really excited about, but nothing to me has hit the Richter scale to the degree that AI does, and the reason is, okay, so if you think about early computers, PCs, that kind of gave us compute at scale, which was awesome. The internet effectively gave us distribution at scale. It allowed us to take information, products, whatever, and get them out at scale because everything was now connected, where before it was not. What AI gives us now is cognition at scale, the ability to literally amplify the human brain, right? That’s never been done before. Look, not that we haven’t been trying, but now we have demonstrable evidence and products that actually work. We’ve used ChatGPT and products like it. And so it’s life-changing, right? And so for the product people out there, which is a lot of you, AI…
By the way, the internet unlocked a massive opportunity because, especially for startups, it’s like, “Oh, we can just reimagine the world with this new thing called the internet that was not possible a year ago and now is. What can we now unlock?” and lots of great companies were created. I think something similar will happen here.
And there’s a couple of things just tactically that I think are possible vectors that can satisfy my spreadsheet of if I were going to choose ideas, and these apply across industries, is that, so as software companies and product builders, most of us have always called our products like, “Oh, we build intuitive products.” We say that, right? As it turns out, it’s just not true because as a user of X product, whatever it happens to be, it’s like, okay, I’ve got this mental model, the thing I’m trying to accomplish, and then I have to translate that into a series of clicks, drags, touches, and swipes in order to accomplish the thing that I want in the software product that I’m using, and what varies is the degree to which that translation needs to occur.
But there is this impedance mismatch, right? There’s a translation that has to happen from what’s in my head, like, “I want to remove this background in Photoshop,” “I want to create this report in HubSpot.” And so the opportunity now is that we’re going from what was an imperative model, this is what engineers would call like a step-by-step, you give instructions to the computer like, “Do this, do this, do this, do this,” and then hopefully you get the output you’re looking for, to what engineers would call a declarative model. A declarative model is you describe the outcome you want, not the steps to get there, and the closest analogy that most people are familiar with is SQL, the querying language, right?
So before SQL, if you wanted to get data out of a system, you wrote code that says, “Okay, loop over all my customer records, then filter out the ones that have less than a million dollars in revenue, and then give me the result,” and SQL said, “No, just put a where clause on it. Describe the data you want, and what columns you want, and how you want to order it, and what filters you want to apply,” and it figures out if the database figures out how to get you the thing that you’re looking for.
Now we have the ability to do that for interfaces for software, right? So instead of saying, “Click here, drag here,” it’s like, “I want you to generate a report that’s all my customers in Europe that we signed up last quarter that were more than 100K in ARR.” And you just literally take the thing that’s in your head and express it in whatever medium you choose, text, voice, doesn’t matter, and the software now can actually… So now the technology exists to pull that off. Did not exist before, and now it exists.
I know it didn’t exist before because six and a half years ago, I built this product called GrowthBot. That was a chatbot for HubSpot software and business software generally for marketing and salespeople. It’s essentially what ChatSpot is now. Amazing product, GrowthBot, and it had just one teeny tiny small problem, which is it didn’t work. Really well-conceived, really brilliant idea. It just didn’t work and the technology wasn’t there. You just couldn’t do it with any degree of fidelity because AI hadn’t… And now you can. And that’s what caused me to go prove out my theory, is like I had a chip on my shoulder, particularly with that idea, right? But I think, and I’m not suggesting that web and mobile interfaces are going to go away, but there are a class of use cases across pretty much all software where we can make things easier for the human by not forcing that translation layer and just make it truly intuitive. Yeah.
So thank you for coming to my TED Talk or my [inaudible 01:37:01] podcast.
Lenny: So what’s great about watching you, because you kind of build in public with this AI stuff, that you’re doing it, like your approach to learning about this is building and doing. When you launch ChatSpot, you tweeted that it took you many, many late nights of work.
For people that want to understand what’s going on, get ahead of where things are going, build the skills to be relevant in the future, do you have any advice? Is it just build stuff and do it, or is there anything else you’d recommend for people that want to understand what the hell is going on?
Dharmesh Shah: Yeah, and this is very simple practical advice, and this is the same thing for engineers in software as well, is don’t try to go learn something because you think it’s worth learning. Find an actual problem that you care about and go try to solve it with that new thing, right? If you have something tangible that you’re trying to do, so for instance, going back to the talk around public speaking, it’s like, okay, if I had said, “Oh, I want to learn about humor and stand-up comedy,” I don’t think I would’ve accomplished a thing. But I had a goal in mind, which is like, “Hey, I need to be able to get up on a stage and a hold of a thing and I have this measurable thing.”
So whatever it is, whatever line of business you’re in, whatever you do as a professional, I guarantee you there are ways AI can help you, even with the tools that we have out there right now. So either use the tools, build tools on top of the APIs that exist, very easily accessible right now, very cheap to start. Yeah, just do it and iterate. And I’m a big believer also in learning in public as well. That’s why the Culture Code is out there. It’s like, okay, well, if you accept the fact that culture is a product, we’re going to learn about things much better, like instead of just getting feedback from the employees at HubSpot, why don’t we put it out to the world? Because the internet is exceptionally good at telling you why you’re an idiot, right? It’s like it’s one of its core strengths. So I would recommend, have a goal in mind, write about it. I’m a big believer in writing generally, but yeah, do it in public.
Lenny: I love that advice.
Is there anything else that you wanted to share, or touch on, or leave listeners with before we wrap up?
Dharmesh Shah: We covered a lot of territory.
Lenny: Yeah, we did. Yeah, we did.
Dharmesh Shah: I might do the quit while I’m ahead thing. I’ll leave it with this. So I have my definition of success I came up with years ago that I kind of still stand by and that HubSpot actually stands by, is success is making the people who believed in you look brilliant. That’s it. What I like about that is it’s not like an inward-facing thing. This is the kind of, like customer folks, like all the employees that join your company when you’re a founder that were kind of silly enough to, “Could I join you on this thing?” whatever, like customers that believed in the early product that completely sucked, investors that believed, like all of that, do what you can to like when they reflect back on it, it’s like, “Oh, boy was I smart to having believed in Lenny and what he’s doing.” So yeah, I think…
Lenny: That is really beautiful. Dharmesh, you are wonderful. Thank you so much for doing this.
Dharmesh Shah: This has been a lot of fun. I’m glad we were able to put it together.
Lenny: Same. We covered so much ground. We got through everything that I was hoping to get through and more. Where can folks find you online if they want to follow what you’re up to and maybe reach out if they have questions? And then how can listeners be useful to you?
Dharmesh Shah: Sure. So I’m Dharmesh, D-H-A- R-M-E-S-H, everywhere. I’m not only findable, I’m unavoidable on the internet as it turns out. And follow me on social media. Tell me where I’m being an idiot and what I can learn. Here’s the thing, leave a YouTube comment and tell me which Lenny episodes were your favorite so I can go back, because I’ve been binge-watching anyway. I’d love to know.
Lenny: And then speaking of that view, I know you have websites of your content, you have your YouTube channel. Where can folks find those things?
Dharmesh Shah: Yeah, if you just go to dharmesh.com, that’ll have links to the other places. I’m probably most often on LinkedIn as it turns out, which is a whole topic for a whole other day, but…
Lenny: Yeah. Awesome. All right. I’ll let you go. Dharmesh, thank you so much for being here.
Dharmesh Shah: Thank you. Bye, everyone.
Lenny: 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 |
|---|---|
| 4D chess | 四维象棋 |
| all-hands meeting | 全体会议 |
| ARR (Annual Recurring Revenue) | ARR(年度经常性收入) |
| bi-modal | 双峰 |
| Bowling Pin | Bowling Pin(保龄球策略,保留原文) |
| Brian | Brian(Dharmesh 的联合创始人,保留原文) |
| Cambridge, Massachusetts | 马萨诸塞州剑桥市 |
| carrying cost | 持有成本 |
| chat.com | chat.com(域名,保留原文) |
| ChatSpot | ChatSpot(产品名,保留原文) |
| Chris Miller | Chris Miller(人名,保留原文) |
| code of conduct | 行为准则 |
| core values | 核心价值观 |
| Culture Code | Culture Code(文化代码,HubSpot 的文化幻灯片,保留原文) |
| debate, decide, unite | debate, decide, unite(辩论、决定、团结) |
| declarative model | 声明式模型 |
| designated insiders | 指定内部人 |
| Dharmesh Shah | Dharmesh Shah(人名,保留原文) |
| dimensional complexity | 维度复杂性 |
| disagree and commit | disagree and commit(不认同但执行,Amazon 的决策原则) |
| DRI (Directly Responsible Individual) | DRI(直接负责人) |
| dying on the hill spectrum | 殉道光谱 |
| efficient market | 有效市场 |
| Elon | Elon(指 Elon Musk,保留原文写法) |
| entropy | 熵 |
| expected value | 期望值 |
| feedback loops | 反馈环路 |
| first principles | 第一性原理 |
| first-order thinking | 一阶思维 |
| flash tags | flash tags(Dharmesh 提出的分级标签系统,保留原文) |
| flat organization | 扁平化组织 |
| founding principle | 创始原则 |
| freemium | 免费增值 |
| functional decomposition | 功能分解 |
| Geoffrey Moore | Geoffrey Moore(人名,保留原文) |
| go-to market strategy | 市场进入策略 |
| GrowthBot | GrowthBot(产品名,保留原文) |
| guardrails | 护栏 |
| hashtag | hashtag(标签) |
| heuristic | 启发式方法 |
| high conviction, low consensus bets | 高信念、低共识押注 |
| high-stakes speaking | 高风险演讲 |
| HubSpot | HubSpot(公司名,保留原文) |
| impedance mismatch | 阻抗失配 |
| imperative model | 命令式模型 |
| inalienable rights | 不可剥夺的权利 |
| Kipp | Kipp(人名,保留原文) |
| learning in public | 公开学习 |
| Lenny | Lenny(播客主持人,保留原文) |
| lightning round | 闪电问答 |
| liner notes | liner notes(旁注,保留原文) |
| LPM (Laughs Per Minute) | LPM(每分钟笑声次数) |
| marked to market | 标记市值 |
| market cap | 市值 |
| mathematical induction | 数学归纳法 |
| megaphone issue | 扩音器效应 |
| mental model | 心智模型 |
| meritocracy | 精英制度 |
| MicroWare | MicroWare(为小规模用户群定制的软件,保留原文) |
| N+1 | N+1(逐次递增策略,保留原文) |
| NPS (Net Promoter Score) | NPS(净推荐值) |
| one-way door / two-way door | 单向门 / 双向门 |
| operating system | 操作系统 |
| opt in | opt-in(选择加入) |
| Peter Thiel | Peter Thiel(人名,保留原文) |
| prowess | 能力 |
| punchline | 包袱 |
| recursive loop | 递归循环 |
| revenue concentration | 收入集中度 |
| reverse gravity | 反向引力 |
| roadshow | 路演 |
| SDR (Sales Development Representative) | SDR(销售开发代表) |
| second law of thermodynamics | 热力学第二定律 |
| second-order thinking | 二阶思维 |
| SMB (Small and Medium Business) | SMB(中小企业) |
| SoloWare | SoloWare(为单人定制的软件,保留原文) |
| SQL | SQL(结构化查询语言,保留原文) |
| stack-rate | 排序 |
| third-order thinking | 三阶思维 |
| truth function | 真值函数 |
| Typeform | Typeform(产品名,保留原文) |
| unicorn | 独角兽 |
| VP of engineering | 工程 VP |
| where clause | where 子句 |
| Wordplay | Wordplay(产品名,保留原文) |
| Wufoo | Wufoo(产品名,保留原文) |
Reformatted by reformat_english.py
Zigging vs. zagging(逆势而行):HubSpot 如何打造一家 300 亿美元的公司 | Dharmesh Shah(联合创始人/CTO)
精彩预告
Dharmesh Shah: 我听过的一些最好的创业建议是:创业公司应该专注于一件事,并且在那一件事上做到真正世界级的出色。而我们早期的逆势之举之一,就是决定做完全相反的事。
Lenny: 你没有任何直接下属,而且我相信你在 HubSpot 从来就没有过直接下属?
Dharmesh Shah: 我可以通过一些培训和辅导,在管理方面达到勉强过得去的水平。但我不想把生命中任何一年花在成为”勉强过得去”的事情上。
Lenny: 定义企业文化的那个过程是怎样的?
Dharmesh Shah: 我和联合创始人正在开一次创始人会议,他说:“哦,Dharmesh,我听说这个文化的东西很重要。对了,你能去搞一下吗?“我说:“好吧。Brian,全公司上下所有人里,我大概是最不适合做这件事的人。“不是我不喜欢人,我只是不太喜欢和他们待在一起。
Lenny: 你身上有一点非常独特且有趣,就是你痴迷于喜剧和主题演讲的准备。
Dharmesh Shah: 这要归结到一个脱口秀演员使用的指标,叫做 LPM(Laughs Per Minute,每分钟笑声次数)。我自己写了专门的软件,它会告诉我:“好的,这里是观众笑的时间点。“
嘉宾介绍
Lenny: 今天的嘉宾是 Dharmesh Shah。Dharmesh 是 HubSpot 的联合创始人兼 CTO,也是我见过的最令人着迷、最善于从第一性原理思考的人之一。在这次对话中,我们涵盖了很多内容:Dharmesh 在准备演讲时那套既搞笑又天才的方法,包括测量每分钟笑声次数;他作为上市公司高管超过十年的最大心得,尤其是他骨子里仍然是一个创业者;他如何创建和扩展 HubSpot 的文化——你会发现这段故事既搞笑又令人振奋。为什么创始人和产品团队都在与热力学第二定律对抗,如何在其他人都在顺势而行时选择逆势而动,Dharmesh 如何以及为什么倾向于发挥自己的优势,包括在运营 HubSpot 的十八年中从未有过一个直接下属,以及更多内容。这一期非常有趣,会在很多方面拓宽你的思路。接下来,有请 Dharmesh Shah。
有趣的事实
Lenny: Dharmesh,非常感谢你来参加节目,欢迎来到播客。
Dharmesh Shah: 谢谢你的邀请,Lenny。荣幸之至。
Lenny: 这是我的荣幸,感谢你的参与。你是一个非常迷人的人,所以我想先从我在网上发现的关于你的一些趣事开始。我的想法是,我逐一列出来,一边说一边告诉我它们是否属实,然后这会引出我想要讨论的很多不同话题。怎么样?
Dharmesh Shah: 听起来很好。
Lenny: 好,我们开始。第一,你没有任何直接下属,而且我相信你在 HubSpot 从来就没有过直接下属。
Dharmesh Shah: 正确。七千多名员工,但从时间 T 等于零到现在,恰好零个直接下属。
Lenny: 好,这个我们待会儿再聊。因此你也不做一对一会议了——因为没有下属,就没有人跟你做一对一会议?
Dharmesh Shah: 正确。
Lenny: 你在 HubSpot 期间做了很多副项目,包括一个叫 Wordplay 的产品,一度每月收入九万美元,拥有一千六百万用户。
Dharmesh Shah: 正确。
Lenny: 你还花了一千万美元买了 chat.com,然后两个月后又以比你买入价更高的价格卖掉了?
Dharmesh Shah: 实际上是一千五百多万美元,不过我当时说的是八位数……嗯,是的。
Lenny: 哇哦,这是独家新闻吗?
Dharmesh Shah: 是独家新闻,没错。
Lenny: 哇。所以那不是你的买入价?
Dharmesh Shah: 对。
Lenny: 好吧,你仍然是以未公开的金额卖出的,对吗?
Dharmesh Shah: 未公开的金额,是的。而且不止于此,我只能说这么多。
Lenny: 是的,是盈利的。我知道你还给慈善机构捐了一些钱,还有你在 LinkedIn 帖子上做的承诺——有人评论,你就又多捐了一些,非常棒。好吧。还有,你是亿万富翁?
Dharmesh Shah: 这也是对的。目前还是。
Lenny: 好的。还有,你出生在印度的一个村庄,那里没有铺好的街道,没有红绿灯,没有医院。
Dharmesh Shah: 正确。
Lenny: 好的。在 HubSpot 之前,你创办过两家公司,但我觉得更重要的事实是——你向妻子保证过不会再创办新公司,然后你又跑去创办了 HubSpot。
Dharmesh Shah: 是的。我向她保证过我不会再创业……我本来要收起那顶创业者的帽子了。计划赶不上变化嘛。我在研究生院遇到了我的联合创始人,当时我的计划是——“哦,我要回研究生院找找自己”,然后最终去教书,不再做创业。但我在那里遇到了我的联合创始人,然后一切就这样顺其自然地发生了。
Lenny: 希望她已经原谅你了。
Dharmesh Shah: 我也希望如此。
Lenny: 好的。最后一个事实。Kipp,是这么发音的吗?Kipp,你们的市场总监?
Dharmesh Shah: 嗯嗯。
演讲准备与幽默方法论
Lenny: 好的。我问过他应该问你什么,他告诉我,你身上有一个非常独特且有趣的特质——你对文案写作、喜剧和主题演讲准备有着近乎痴迷的投入。你有一种非常独特的方式来准备一场演讲,以及你如何思考幽默、幻灯片设计和故事线。
Dharmesh Shah: 确实如此。公开演讲对我来说并不是与生俱来的能力。我们可以从这里开始聊起,之后再谈谈文案写作,以及其他一些相关的事情。我有一个原则——尽量不做自己不擅长也不享受的事情。但这个原则有一个罕见的例外:作为创始人,我必须登上舞台。HubSpot 每年都有年度活动,我们的年度大会。最开始办第一届的时候,大约只有 150 人,在万豪酒店。之后稳步增长,现在现场观众大约有一万人,线上还有数十万人观看。
但我逐渐意识到,这件事我既推不掉也躲不开。我看到了趋势走向,心想:“好吧,这件事我必须真正去学会,因为没法委派给别人,也搪塞不过去。” 所以我非常相信天赋与技能之间的区分。我们都理解天赋——比如有人在音乐上有天赋,有人在运动上有天赋,诸如此类。天赋确实存在。但我的看法是,天赋基本上决定的是曲线的斜率,而大多数能力实际上是可习得的技能。举例来说,如果你有音乐天赋,你学音乐可能比别人更快,天花板也可能更高,但这并不意味着别人学不了音乐,对吧?
这是第一课——它是一项可习得的技能。第二课是:存在一套方法论。这是我作为工程师的本能——对问题进行功能分解。我会问:“要成为一名相当不错的公开演讲者,需要具备哪些底层子技能?” 第一个技能:你得能站在台上不晕过去。这是最低要求。这些年来我一直在做的事情就是,对”公开演讲”这个技能进行功能分解——或者说我称之为”高风险演讲”:我必须真正抓住观众的注意力。这是成功的组成部分。不管你的信息是什么,如果失去了注意力,其他一切都不重要。而在当今这个注意力短暂的社会,失去注意力越来越容易了。
每年我会挑选公开演讲技能体系中的一个不同部分来学习。比如:“好,我想学幻灯片设计和视觉表达。” 不是真正的设计,我不是设计师。后来我发现影响最大的东西,是幽默。然后你就像工程师那样,把这个技能再分解下去:好,有幽默写作,有台上幽默执行,等等。还有单口喜剧。那怎么把这些要素组合起来?在这个过程中,我想到了一个东西——我一直是一个非常数据驱动、量化导向的人。我就想:“好,我怎么知道某个东西够不够好笑?我有没有在失去注意力?” 所以在准备这些每年一次的高风险主题演讲时,我会先把演讲内容准备好,然后进行排练。
我会找越来越多的真实观众来练习,每次都会录制下来。到目前为止这些都还算正常。我还会把演讲转录成文字,然后运行我自己写的定制软件,它会标注出观众笑了的时间点——必须是可听到的笑声才算数。这涉及到一个单口喜剧演员使用的指标,叫作 LPM(Laughs Per Minute,每分钟笑声次数)。实际上有一个基准可以参考——单口喜剧演员的 LPM 非常非常高,商业演讲的 LPM 非常非常低。而最受欢迎的 TED 演讲之间,存在很强的相关性——最受欢迎的 TED 演讲往往 LPM 也很高。所以我给自己设定了一个目标:必须达到一个最低标准,并且持续提升。我的软件会计算每分钟笑声次数。
然后你就会学到这些行业小窍门。好,如果你要优化这个笑声比例,有两条路径:一是增加笑声次数,二是减少两次笑声之间的词数。我的软件会生成一张可视化地图——把整场演讲按每分钟或每 30 秒分解成小方块,标注出观众喜欢的部分。你会看到:“哦,这里我讲了一分十七秒,没一个人笑。” 那我就得做两件事之一:如果这段内容没法在不显得刻意的情况下加入笑点,就缩短它;或者我必须想办法在中间注入一些有趣的东西,把它打散。
说了这么多,核心意思就是:如果我都能站上公开舞台,学会这项技能,那任何人都能学会几乎任何东西。关键就在于练习、衡量、持续地一点点进步。
如何在演讲中运用幽默
Lenny: 哇。我觉得仅仅从这个问题和你解决问题的方式中,我就已经对你和你的思维方式了解了很多。关于幽默这部分,我很好奇,有没有什么经验可以分享——怎样才能更有趣?
Dharmesh Shah: 幽默有一个行业技巧,在单口喜剧中尤为明显,但其实适用于所有喜剧形式。我之前从没这样想过——这个技巧非常实操,而且很容易应用。假设你看一个故事——我们大多数人不是单口喜剧演员,不会以经典意义上去”讲笑话”。
其实即使是单口喜剧演员也不讲笑话。他们实际上是在讲故事,碰巧其中有一两个包袱。所以,两条实操建议。第一:当你讲述一个故事时,不管有趣的点是什么,那个点必须恰好是那个段落的最后几个词。一旦你把它说出来,就必须停下来。原因是,观众大约需要半秒钟来反应,然后他们需要获得”可以笑了”的许可。如果你继续说话,会让他们感到尴尬,效果就会大打折扣。所以,即使你需要调整句子中的措辞位置——“嗯,把这几个词放在末尾听起来有点奇怪”——也还是这么做。哪怕感觉别扭,实际效果会更好。
这是第一条建议。第二条关于 LPM 的建议是:构建这样的故事——你可以花时间铺垫,但之后在其中设置多个笑点、多个包袱。因为你已经做了投入——我已经说了 75 个词来建立这个情境,说了一些好笑的、有趣的内容。然后继续说一些话,你已经有了上下文基础——你会看到单口喜剧演员一直在这样做。然后再抛出一个笑点。因为一次笑就是一次笑,不管它是不是还在同一个故事里。所以要利用你已经投入的、用于建立情境和场景的成本,然后以间隔的方式在后面挤出更多幽默。
Lenny: 我觉得整期播客光这一个话题就够了,但我得转到其他话题了。
LPM 的具体数值
Lenny: 快速问一下,你刚才说理想的每分钟笑声次数是两到三次左右?是这么说的吗?
Dharmesh Shah: 对,超过两次对于商务演讲来说很难,因为单口喜剧演员之所以能做到,是因为他们唯一要优化的就是 LPM。而商务演讲中,你实际上有另一个议程,不只是为了娱乐。你确实有一个要传达的信息。你在推广某种东西——推广一个想法,推广一个产品。所以在我看来,在大多数正常情况下,即使你能达到 1、1.2、1.25 以上,在非专业人士的演讲中你也已经是 top 10% 了。
Lenny: 哇,这太厉害了。所以每次演讲之后,你都会把文字稿和音频跑一遍你的程序,看看自己表现如何?
Dharmesh Shah: 是的,是的。
Lenny: 你打算公开发布这个程序吗?听起来真的很酷。
SoloWare 的概念
Dharmesh Shah: 我有一个概念,我称之为 SoloWare。SoloWare 顾名思义,就是专门为一个人构建的软件,在我的情况下,那个人就是我自己。这件事我已经做了 30 年了。我就是做一些自己觉得有用的东西。SoloWare 的一个好处——好处有很多——其中之一是 UI 只需要服务一个人。你不需要做大量测试,因为它只需要对一个人有效。最重要的是,因为只服务一个人,如果它不再有用了,你可以直接关掉它,直接停止使用。相比之下,如果你有用户,哪怕只有 10 个用户,你把它放出去了,然后因为某种原因要下线,人们会失望,这就比较麻烦了。
所以,这就是我的权衡逻辑:这个东西对足够多的人来说是否有足够的价值,值得我花精力把它从 SoloWare 变成非 SoloWare,甚至变成 MicroWare?MicroWare 这个词我刚编的,不过就是这样。
Lenny: 哇,好的。让我们看看 YouTube 评论区会有什么反应,看看有多少人想要这个,愿意为之付费。
不做管理者的选择
让我换个稍微不同的话题。我们之前聊到你没有任何直属下属,你从来没有任何直属下属?我知道这在某种程度上跟你经常分享的一条建议有关,就是发挥自己的优势,不要去做那些你还不擅长的事情。你能谈谈为什么会这样吗?为什么你没有直属下属?然后再聊聊这个关于优势的理念,以及专注优势的重要性。
Dharmesh Shah: 好。我来简单讲讲这件事的来龙去脉。我的联合创始人和我举行了 HubSpot 的创始会议。那是我们决定要做这件事的那一周,然后我们有一系列创始人的讨论问题或话题清单。其中有一个——我想我已经公开发布过那份清单了——是所有联合创始人在早期阶段应该互相问的问题。其中一个是:好吧,谁来当 CEO?这个问题很简单,因为我之前已经两次当过 CEO,而且我已经想明白了我不擅长这个。我其实很差。我差的原因之一是,CEO 角色的一部分是善于管理人。不仅仅是领导力和愿景之类的。你实际上必须把管理当作一门手艺来做。所以,这一点我已经心里有数了。
我想 Brian,我的联合创始人,心里也早有答案——所以那个对话非常快就结束了。在那一刻,一旦那个问题解决了,就像,“好,我不想当 CEO。” Brian 想当 CEO,因为他从来没做过。我心想,“好吧,这就简单了。” 然后我说,“哦,顺便说一下,我不想有任何直属下属。” 这个念头就是突然冒出来的,我事先没有计划过,直接就跟他提了。
他的反应是,“好吧……” 原因是我已经认识到我不擅长管理。我是一个还算聪明的人。我想经过一些训练、辅导之类的,我可以达到”勉强还行”的水平。但我不想花生命中任何一年的时间去变得”勉强还行”。我宁愿把同样的精力投入到那些我擅长而且真正享受的事情上。这两者是高度相关的。你倾向于擅长你享受的事情,也倾向于享受你擅长的事情。Brian 说,“好啊,没问题。” 然后我说,“不不不,Brian,你不明白。” 因为接下来我说的是,“公司发展到某个阶段会有这种情况。” 他说,“哦,我们的工程 VP 刚离职了。你先临时顶一下,让这个团队向你汇报,等我们找到人再说……” 我说,“这种事肯定会出现,你得向我保证我们不会这么做。我们……” 就是这样,我不会有直属下属。
我们进行了一次非常坦诚的对话,那是自己做过的最好的决定之一,无论对我自己还是对 HubSpot 来说。如果当初没有做这个决定,我是个创业型的人,我不确定自己能在 HubSpot 这样规模的公司里撑下来。而现在我可以坦诚地说,我在 HubSpot 现在的 7000 人规模下比当年 70 人的时候过得更开心。原因是,我获得了规模带来的所有好处——我可以下大赌注,我们可以有长期的事情和计划,为全球扩张做规划——同时不需要承担规模的代价,就是”哦,你得管理这么多人,处理各种琐碎事务”。那很难,那是一项很难掌握的技能。
我欣赏那些具备这项技能的人。我不具备。所以我获得了规模的几乎所有好处,却几乎没有规模的代价,这就是让我在 HubSpot 保持投入、充满活力、感到开心的原因。所以我认为这个决定对 HubSpot 的业绩有利,对我个人也有益。
按自己的方式构建公司
Lenny: 我很喜欢这个例子——它说明了你可以按自己想要的方式来构建公司,不必照搬别人的做法。我猜这样做也会有很多挑战,但我觉得这很鼓舞人心,就是”嘿,我要按自己的方式来设计这家公司”。
Dharmesh Shah: 是的,在按照自己的喜好塑造这个世界方面,你能做到的事情之多是令人惊讶的。我觉得人们会自动默认事情必须按照某种方式进行,尤其是创始人和第一次创业的人。
你至少应该试试。比如,“好吧,我想这样做……” 我再给你一个实操层面的例子。我的联合创始人和我都是夜猫子,不是早起的人。碰巧如此——我当初选他不是因为这一点,是后来才发现的。所以从 HubSpot 的早期,至少在前几年,我们定了规矩:“上午 11 点之前不开会。就这样。” 没得商量。后来大约三、四年后我们调整了一下,说,“好吧,11 点之前可以开会,只是不能邀请任何一位联合创始人参加。你们如果愿意,可以自己开会。” 不过就是这样,一些小事。
Lenny: 这条规矩太棒了。我有一条个人规矩:下午 3 点之前不开会,因为我不在任何地方上班,可以更自由地安排自己的日程。但核心理念是一样的——在一天的开始创造深度工作的时间。
Dharmesh Shah: 我很喜欢这个。是的。
Lenny: 好的。你刚才说你是创业型的人,但与此同时 HubSpot 上市已经快 10 年了,从创立算起大概 18 年了。
Dharmesh Shah: 没错。
关于上市的建议
Lenny: 我很好奇,你作为一个明显具有创业气质的人,同时在一家快速成长的大型公司担任高管,你从这段经历中学到了什么教训?我的意思是,对于那些正在考虑创业的创始人,你有什么建议可以分享?对于那些正处于同样角色的人呢?
Dharmesh Shah: 几点想法。第一,既然我们已经是上市公司了,这一点其实没被谈得够多。我想借这个机会把这个信息传出去——我认为太多创始人在达到某个阶段后,对上市、走 IPO 流程、成为公众公司这件事非常抗拒,因为他们觉得……那就像是末日的开始。“一切都会变,我的日子会很难过。”
我认为他们过度放大了对公众公司生活的想象。以我个人的经验,我的联合创始人 Brian 也可以作证——当然,作为上市公司,我们确实有一些作为私有公司时不需要承担的成本和事务。但实际上也有好处。比如,作为私有公司时,你在某种程度上受制于市场环境和 VC 对你所在品类的看法等等。只有在每次融资的时候,你才会被”标记市值”——也就是根据有人愿意出资的金额来确定估值。而作为上市公司,最美妙的一点——这一点被低估了——就是我们随时知道市场对 HubSpot 的估值是多少。任何一个交易日我都能告诉你。这有一种踏实感。这也是我对公司内部说的另一个关于估值的事情:我们处于一个相对有效的市场中。公开股票市场,按经济学的定义来说,是一个相对有效的市场。所以估值会围绕价值波动。
如果你专注于创造价值——也就是我们真正在构建的东西——估值有时会高于你应得的,有时会低于你应得的,但长期来看,这两样东西会同步移动,对吧?这已经被证明了。这就是事实。这几乎是有效市场的定义——信息本质上在流动,然后估值会追上实际价值。
但还有一个社会层面的原因,我认为创始人应该上市。HubSpot 在上市前的那些年里——我觉得我们上市时已经成立了八年——我们创造了大约十亿美元的市值,上下浮动一些。我们当时还不是独角兽,上市时略低于那个门槛。所以 HubSpot 从来不是独角兽,因为独角兽是指估值超过十亿的私有公司。
在随后的这些年里,我们的市值从十亿美元增长到了大约三百亿美元,差不多就是现在的水平。那最初十亿美元的市值创造,只有极少数人得以参与。而其余的两百九十亿美元市值增长,每一个公开市场的投资者都有机会参与。所以我喜欢这个理念——你相信你的公司,很多人相信 HubSpot,让他们也能参与进来是件好事。我们的客户、合作伙伴、支持者们,现在都能分享这个上涨的收益。这在私有市场里是做不到的,因为那是一个非常封闭的体系——你不能随便买你喜欢的私有公司的股份。我希望这能实现,但目前不行。所以上市是一种让更广泛的市场更早参与你成长的方式。
保持热情与动力
Lenny: 那从保持兴奋和动力、找到新挑战的角度来看呢?对于那些在大公司的创始人,有没有什么经验——“好吧,这能帮助我保持对正在做的事情的热情”?
Dharmesh Shah: 有的,我觉得这件事……Bezos 在这方面是最好的典范——成为上市公司并不意味着你必须做某些特定的事情。你唯一需要做的就是保持透明:“这是公司在做的事,这是我们的方向,这是我们要解决的问题。“他在第一封年度股东信中就做了这件事——他每年都会写——“这就是 Amazon 的本质,我们着眼于长期”之类的。他不是以一种傲慢的方式说的。如果你重读那封信,细读言外之意,它传递的是:如果你不认同这种方式,你不应该买 Amazon 的股票。这完全是可选的。“你不必买,但我们会这样经营这家公司。”
他始终如一地践行了这一点。你也可以继续这样做。我给你讲个小故事。这又是一个用来打消那些对上市犹豫的创始人顾虑的例子。HubSpot 的核心价值观之一是透明,这是从公司成立最初期——真的是最初那几天——就是文化的一部分。我们践行透明的一个机制是:HubSpot 内部的所有信息对公司里的每一个人同等开放,没有例外。
所有东西。我们的资产负债表,我们融资的估值,我们六周后能不能发得出工资——所有这些,所有信息始终透明。只有两个例外。一是如果法律不允许我们分享,比如我们正在考察某个收购标的,签了保密协议,或者信息不完全属于我们。另一个例子是薪资。我们认为薪资是由公司和个人共同拥有的。如果本人不愿意,我们没有权力公开某人的薪资。好的。
我们一直保持着这种透明度,然后我们准备上市了,和投资银行家、律师开会。这是 IPO 准备过程中的一次全天会议,其中一个问题是:“好吧,你们指定的内部人是谁?”
Brian 和我之前都没带公司上过市,我们说:“我们不知道那是什么。“他们说:“哦,就是那群能够接触到所有财务数据、了解内情的人。“我们说:“哦,好的,那这个名单有多长?“他们说:“五六个。“然后 Brian 说:“七个?能七个吗?“他们说:“可以,七个没问题。“我又说:“八个呢?能八个吗?“于是我们通过数学归纳法发现了——“如果 N 成立,那 N+1 是否也成立?“——原来法律上对一家公司可以有多少个内部人是没有实际上限的。
全员指定内部人
Dharmesh Shah: 所以我们最终做的,就是在我们上市的那一天,把每一个员工都指定为指定内部人。这让我们得以维持透明度的做法,因为没有规定说你只能有五个、六个或十个内部人,而我们有几百号人。我们没有把它做成可选的。我们没有说”哦,你可以自己选”,因为当内部人是有代价的。有些窗口期你不能交易,诸如此类。但我们的态度是:“好吧,我们相信透明度。每个人都能接触到所有信息,而且这不是一个可以 opt-in 的事情。就是这样,我们大家一起承担。“我们至今仍在这样做,即使已经有七千人了,每个人都是内部人,所以我们可以一直跟所有人分享所有财务数据。
逆向思维与第一性原理
Lenny: 你有很多逆向思维的观点,这是我从这次对话中已经感受到的。你不断推进边界——“我们能设十个吗?能设一千个吗?“我很好奇,在经营公司方面还有没有其他逆向的做法?我知道可能有很多,但你脑海中浮现的是什么?
Dharmesh Shah: 我们这个行业很喜欢用的一个词就是”第一性原理”这个概念。我不知道是谁推广开的,但确实被用得很多。我不想说什么刻薄的话,但我觉得大多数时候人们使用第一性原理,其实都用错了。因为第一性原理应该怎么运作——正如 Elon 可能会描述的那样——不是说”我们相信的第一性原则”,不是”这些是公司的核心第一性原则”,而是宇宙的第一性原理。不是你以为真的东西,而是我们所有人根据目前所知共同确认真实的东西。那是科学,是物理学。比如热力学第一定律和第二定律,那些才是第一性原理。在它之上的所有层,都是你以为真的东西,是你的假设或你做的决策,这没问题,你可能确实做了一些决策。
比如,透明度不是一个第一性原理,它是一个创始原则。不是说每个人都应该这样做,但它是我们公司完全核心的东西。所以我确实可以做一个逆向思维者,但你必须限制维度。你在多少个事情上持逆向观点,这个数量必须大于零,但不能太高。我称之为”高信念、低共识”的押注。高信念就是”我们真的、真的相信这个”,低共识就是大多数其他人不这么认为。也有人用不同的方式表述过这个概念。我觉得 Peter Thiel 说得更简单、更优雅:“你需要在一件别人认为你错了的事情上是对的,并且持续很长时间。“你只需要是对的。
高信念押注:专注 SMB
Dharmesh Shah: 我们早期在 HubSpot 做的一个高信念决策,就是将 SMB 作为我们的目标市场。我会同时说明选择 SMB 的理由,以及做一个高信念押注的理由——你的高信念押注不一定是目标市场,但总得有某个东西。一旦你做出这个押注,因为它是低共识的,所以你很难,也必须要有信念,因为所有人都不同意你,包括你的董事会、你的投资人、潜在投资人。在 HubSpot 整整十八年的历史中,我们一直如此。整个 IPO 路演期间都是这样,人们问:“是的,我们知道你们专注于 SMB,但通往企业市场的路径是什么?“不,你不明白。这不是一个市场进入策略。不是说”我们先攻克 SMB”,不是 Bowling Pin 策略,不是 Geoffrey Moore 的那一套,而是”我们就是为 SMB 而来的”。这就是我们要做的事。
所以我们在很长一段时间里保持了这种信念,而这有助于你做出其他次要的押注。但你不想把所有东西都重新发明一遍。我举一个我们早期在 HubSpot 做过但后来放弃的事情的例子。HubSpot 早期,我们说:“好吧,我们相信扁平化组织。“扁平化组织的一个体现是,公司里没有人有头衔。没有人有头衔。即使你有名片之类的,我们也没有头衔。如果你在会议上做自我介绍——当时我们已经几百人了——你介绍自己就是”我在做产品""我在做工程”之类的,而不是”我是某某总监""某某 VP”,因为头衔这个东西在公司里根本不存在。
从无头衔到经典头衔
Dharmesh Shah: 后来随着规模扩大,我们意识到人们对头衔确实赋予了实际价值。他们之所以赋予头衔价值,是因为有传言说 HubSpot 之外还有生活。我简直不敢相信。“什么?HubSpot 之外还有生活?“嗯,看起来是的。然后你去参加感恩节晚餐之类的场合,跟你阿姨聊天,你需要某种简短的说法。比如:“哦对,我升到总监了。我升到……”只是需要展示一些进展,因为它是一种信号机制。
而且在 HubSpot 之外的生活中——HubSpot 之外——其他人也会想知道我处于什么位置。好了,这是最终说服我们的论点。我特别喜欢这个论点。因为我们花了一整天开会讨论这一个话题。最终说服我们的论点是:“Dharmesh、Brian,我理解为什么我们不要头衔。我理解。“然后他们陈述了理由:“是的,头衔有价值。而你们因为不给我们头衔,就意味着你们必须在经济上用其他方式补偿我们。所以你们其实是亏了——那个价值高到不值得这样做吗?“这是一个很有说服力的论点。
然后我们把选择缩小到三个。我们可以说:“好吧,“我们可以选择坚持不要头衔。也可以说:“好吧,我们用经典头衔。“然后桌上还有一个选项:“自己编一个头衔。“你可以称自己为某某大统领,那就是你的头衔。最终我们决定采用经典头衔,因为这确实满足了那个需求——有一个基准、一个标准,否则它没有任何意义。自编头衔基本上等同于没有头衔,因为它们本身就不具备任何意义。但这就是那种你会不断迭代、随着发展而调整的事情之一。
Lenny: 是的。我觉得这里面有很多这种”让我们试试不同的做法。好的,没成功。没关系。我们就在这件事上跟大家一样。“然后往往你会发现一些东西,“好吧,这个确实更好。”
Dharmesh Shah: 对。
Lenny: 想确认一下,当大家没有头衔的时候,比如一个人是产品经理,你们叫他产品经理吗?还是就是”我是 HubSpot 的员工,我没有任何头衔”?
Dharmesh Shah: 当时就是没有头衔。现在我们有经典头衔了,所以有各种层级——
Lenny: 对,但那时候就是”你就是在这里——”
Dharmesh Shah: 没有头衔。真的没有头衔。
Lenny: ”……我们也不确定具体叫什么。”
Dharmesh Shah: 没有头衔。
创业初期统一薪资
Lenny: 我在某处看到,HubSpot 初期你给每个人每月五千美元的薪资,包括你们自己,而且持续了很长时间。
Dharmesh Shah: 是的。
Lenny: 这个对吗?好的。
Dharmesh Shah: 对。
Lenny: 又一个有趣的事实。顺便问一下,这持续了多久?
Dharmesh Shah: 我回忆一下,那是很久以前的事了。我想大概是第一年,一年半左右。
Lenny: 太厉害了。
为简单而设计
Dharmesh Shah: 有件事,这是我们的核心……当然不是我们所做的每件事都是对的,也不是每件事都适用于其他公司。但有些东西我觉得是通用的,其中之一就是”为简单而设计”的需求。在早期岁月里,你越能把事情简化,通常就越好。这听起来可能有点奇怪——我们之所以走向透明,并非出于什么道德立场,不是说”每个人都应该拥有这些”,也不是什么追求真理的道德诉求。而是这样的:当我们招聘第一位员工时——Brian 和我两人,当我们招聘第一位员工时,我们不得不做一个决定:“那我们给他们访问哪些文件?“然后 Brian 就说:“全部。没什么好藏的。为什么要搞那个?”
于是我们顺着这个思路想下去……得出的结论是:“好吧,第一,二元决策比非二元决策简单得多。“要么全部开放,要么全部不开放,这比中间态简单太多。然后我们想:“好,如果这对第三号员工有效,为什么对第四号不行?为什么对第五号不行?为什么对第五十号不行?“我们就一直这么做下去,因为——“好,在它崩掉之前,我们就继续这么做。“这个原则适用于 HubSpot 的方方面面。我们总是从可能奏效的最简单方案开始,然后如果必要,再逐步加上复杂性、加上其他东西。
甚至我们最早的办公室——这是每个人都可以带走的一个实操经验,尤其是早期创业公司——第一个 HubSpot 办公室恰好只有四张桌子,因为我们对增长有野心,而且当时在联合办公空间里。我们当时住在马萨诸塞州剑桥市的 WeWork 里。所以我们就四把椅子,唯一能区分它们的特征是:两把靠窗,两把不靠窗。好,招了第一个员工。接下来要做另一个决定:“好,那大家坐哪儿?“我们觉得,如果我们自己直接占了靠窗的位置,那不太公平。我们信奉扁平化组织,对吧?所以我们说:“那就抽签吧。“而且不是”谁抽中谁就坐靠窗位”,而是谁抽中就先从剩余的位子里选——因为不同的人对不同座位有不同的效用偏好。也许你并不想要靠窗的位子,也许你就喜欢暗一点的角落,什么情况都有。
于是第一个员工时我们这么做了,第二个员工也这么做了。每招一个新员工,我们就来一次座位重排:抽签,按顺序选位,这就是我们的算法。我们的投资人当时说:“哦,挺可爱的,也挺有趣。到十五、二十人的时候就不行了。到五十人的时候就更不可能了。“结果二十五人时行得通,五十人行得通,一百人行得通,二百人也行得通。我们对算法做了升级,做了一些局部优化——比如,有些组不应该挨着其他组,因为工程师需要安静,而销售人员不需要,所以我们做了一点优化。不再完全随机了。变成了:在这个区域内……但我们还是会抽。后来我们又调整:“不再每次招人都重排了,改为每季度一次。“这个机制,顺便说一句,一直撑到了几百人的规模,依然运作良好。
Lenny: 天哪。
Dharmesh Shah: 所以这个故事的意义在于:简化,就这么一个决定——你能想象我们避免了多少公司政治吗?在职场中,人们消耗大量精力在”你的办公室比我的大”、“你的桌子高了一英寸”这种事上。而这一切,仅凭那一个决定就全部消失了。我们甚至不知道——说实话这是无法计算的——它为我们省去了多少内耗。除此之外还有很多附带好处:人们彼此认识得更深,团队化学反应更好,很多正面的效果。但仅就”简单”本身而言,它几乎总是更好的答案。
熵与复杂性
Lenny: 我很喜欢的一点是,所有事情最终都追溯到某个你开发的算法,或者那种”N+1”式的方法——“就一直走下去,直到出问题为止”。太精彩了。我期待更多这类既有趣又精彩的、你另辟蹊径做事情的例子。你一直在谈简单性。它的对立面,可以说是熵。有人告诉我你有一个观点:每个领导者、每家企业都在对抗热力学第二定律,也就是熵。
Dharmesh Shah: 对。这条定律的原话是:在一个封闭系统内,熵会随时间增加。
Lenny: 好的。
Dharmesh Shah: 对。
Lenny: 好的。
Dharmesh Shah: 我们不必深究其中的物理学细节。
Lenny: 不,我很喜欢这个。
Dharmesh Shah: 我不是学物理的。
Lenny: 不,我们正在触及一切事物的第一性原理。继续说。你说的这个——在经营企业时它意味着什么?具体是什么样的?
Dharmesh Shah: 好,用通俗的话来说,热力学第二定律——我这里是在转述——就是:“随着时间推移,除非你主动干预,否则一切都会走向崩溃。“我在转述。本质上,一个系统中无序和随机性的总量会随时间增加。你见过这种现象:你打碎一个鸡蛋,分子不会自动聚合重新变回一个鸡蛋。这不会发生。它永远朝反方向走——变得更加无序、更加随机。这就是热力学第二定律。
这在公司中同样会发生,在代码中也同样会发生,几乎在每一个抽象层面上,你都能看到某种变体。所以我对此的思考方式是:在公司早期阶段,你本质上是在为生存而战。就是努力不死。这是第一要务:别死。在公司的第二阶段,你在努力不陷入停滞。你仍然希望能驱动增长,能做事情——“好,我们活下来了,但停滞本质上就是死亡,所以我们仍然在努力不死。”
然后第三阶段,你在对抗复杂性。这就是——如果没有某种外部干预,你最终会在自身的重压之下崩塌。你会看到这种情况不断发生,它以不同方式显现:一切都变得更复杂。你需要更多的管理层级,更多的人头,更多的这个、更多的那个,一切都变得更难。利润率……总之坏事开始发生,你变得更慢。这是一种更慢的死亡——因为你现在规模更大了——但它终究还是死亡。复杂性确实会杀死公司。也许不像其他因素那么快,但比其他因素更可靠。
为简化而战
Dharmesh Shah: 所以这就是为什么尽早种下简化的种子永远不嫌早,把它们植入进去,让这成为组织文化的一部分。我们的文化和指导原则中有一个东西,围绕着我所说的”为简化而战”。就是这三个字。我们想传达的信息是,“简化值得为之战斗。” 这是第一点,它很重要;但另一点是,它需要你为之战斗。简化不会自动发生,而且这将是一场战斗,因为整个宇宙都在和你作对。你需要消耗能量来捍卫简化,因为一切——即使是出于善意的人——都会引入复杂性,因为这是世界的自然法则。我们想要更多的定价层级,我们想要产品中有这些旋钮和刻度盘,我们想要更多这些……所有这一切都在趋向热力学第二定律,熵在增加。总之就是这样。
Lenny: 我很喜欢这个。排座位的算法就是一个很好的例子。还有没有其他例子,无论是在产品还是战略层面,你推动了简化,最终证明是对的?
Dharmesh Shah: 有的。在产品方面,早年的时候——Brian 做了实际的落地执行,功劳归他。我们有一个相对较宽的产品线,这个我们可以以后再聊,即使在早期这样做也有利有弊。但我们一直在追求简化,这个理念我们从 Apple 那里学来的,稍后可以多聊聊它的起源。但我们在 HubSpot 产品中有一条规则,在那些年产品不断成长的过程中,每次你添加一个我们所谓的旋钮或刻度盘——也就是功能——你就必须在其他地方去掉一个。这是一个净值的概念……当然这是一个非常粗糙的衡量方式。就好比,“好吧,并不是你导航栏里放进去的每个单选按钮、复选框、下拉菜单项都一定是等价的。” 但这总比没有好。总比没有任何约束要好。
再说一次,这又回到了二值思维——它至少迫使你去思考这个问题,而不是……我认为人们在产品中经常犯的另一个错误是,我们衡量一个功能的成本——或者甚至是一个新产品——通常是基于实现成本。这是一阶思维,就像,“哦,开发要花六个月,需要 Y 个工程师和 Z 个设计师之类的。” 二阶思维是考虑那个功能的维护成本。就像,“哦,不只是发布第一个版本那么简单,现在我们有了这个代码库,需要去支持和持续改进。” 我理解这一点。而三阶思维,我认为是最微妙的,事实证明也是最重要的,是复杂性带来的其他成本。
好,我们回过头来说这个。所以当你从产品一号走到产品二号时,就像,“好的,产品二号要花这么多钱来开发,它有成功与否的风险,还有所有这些事情。所以产品二号会有这样一个持有成本。” 公司没有想清楚的是……以及它的维护——“哦,我们需要一个团队来维护那个新产品。” 不,实际发生的是,当你从产品一变成产品二时,你给你的业务增加了维度复杂性。
我的意思是,这不是一个增量式的增长——“哦,我们从一变成二了”——而是现在你做的每一个决策都要透过”我们现在有两个产品”这个视角来审视。我们刚招了一个工程师,他们做产品一号还是产品二号?我们要搞一个营销活动,产品一号花五分钟,产品二号花两分钟?我们怎么做任何事情?你看的每一张关于增长的图表——每 X 的收入等等——全部都是。现在你曾经拥有的每一张图表都要按产品一和产品二来切分,才能真正捕捉到那种精确度。
所以现在你有了这种你完全没有预料到的新的维度复杂性。而且同样,这适用于每一个抽象层级。所以你在业务中做的每一件事都应该把复杂性的长期成本纳入考量。而且它应该是值得的。当然,你需要——而且我可以论证你需要构建产品二号以及产品 N,N+1,随着时间推移——但你应该对复杂性的成本保持警觉。
Lenny: 有没有什么具体的做法来帮助把这个落地?因为我猜每个人都会说,“是的,这很好。别做太多事。简化。” 但当你面对”我们得驱动增长,我们有这个绝佳的想法,我们有机会,我们有竞争对手”的时候就很难做到了。我知道这是文化的一部分,所以也许这就是答案,但你如何让人们真正践行这个原则?
Dharmesh Shah: 确实很难,我先说这一点。让事情变得更容易的做法,有一点是 HubSpot 自身的特殊性,所以让我们保持简单的部分——部分原因在于我们早期给自己施加的约束。我们当时并不知道这一点。因为我们在做 SMB,我们有免费增值产品。好,当你做这个的时候,你能添加的复杂性是有实实在在的上限的,因为我们没有一支大军花 18 个月去部署实施。我们必须能够支撑一个免费的产品。这很难做到。它必须足够简单,让用户能从中获得价值。所以就有这些自我施加的限制。
所以我会总结出的经验是:建立系统性的方式和方法——用 Amazon 的话说就是”机制”——尽可能多地设置护栏和约束。因为你可以把文字写在纸上,试图将其注入文化——“哦,我们崇尚简化,我们为此而战”——你可以开会,可以在全员大会上站起来反复强调,我鼓励你这样做。但一个真正——哪怕只是做得还不错的系统——都会胜过你试图采用的任何其他机制。那些东西会随时间退化。很难规模化。所以我非常信仰系统和施加的约束,而不是——
Lenny: 我忍不住想聊聊 SMB 这个话题,因为我做了很多天使投资,经典的观点是,专注于小企业的公司非常具有挑战性,因为大规模触达小企业非常困难。他们往往比较传统,不是早期采用者。你卖出去一个,那也只是一个小客户,他们不会升级并向一大群人推广,他们的支出也不会快速增长。对于那些正在探索 SMB 或者试图在 SMB 领域获胜的创始人,除了”可能别做这个,它非常难”之外,你有什么建议吗?
Dharmesh Shah: 不,我的建议是绝对要做。
Lenny: 太好了。漂亮。
Dharmesh Shah: A,正因为它难——这个我们等下会聊到。但 B,好吧,我给你讲个小故事,关于我们当初怎么选择了 SMB。我的联合创始人和我是在研究生院认识的,他之前做软件,偏销售和市场方向,我是产品和工程方向。我们在研究生院之前互不认识,是在那里相遇的。当我们在琢磨可能一起做的创业点子时,我们说——我们之前都做过企业级软件,就是 SMB 的完全反面。企业级软件确实有它的问题。其中最大的一个是,做企业级软件公司的初创阶段非常痛苦。销售周期超长,反馈环路超长,你得有销售团队,还有各种各样的……你会面临收入集中度问题,意味着你不再完全拥有自己的产品路线图,因为开最大支票的那个人会要求你做这三件事,而你真的很难说不,尤其是作为一家初创公司,所有这些问题都随之而来。
我们心想,“好吧,这个游戏我们大概了解。“光谱的另一端是消费级创业公司。消费级创业公司的问题在于结果呈现非常明显的双峰分布——要么超级成功,成为下一个 Meta、下一个 Google、下一个什么什么,要么直接归零。很少会出现”哦,你赚了几百万”这种情况。因为如果做成了,面对的是一个巨大的市场。选择 SMB 的理由在于,你几乎可以兼得两者的好处。你有企业级软件的好处——“哦,我可以解决一个产品问题,人们会为此付钱给我。我不需要靠广告补贴,不会面临双峰结局。我可以逐步构建一个可衡量的、越来越有价值的业务。“我们就这样从 10 万 ARR 涨到 100 万 ARR,再到 200 万 ARR,因为我们是一家商业软件公司。但它又有消费级的好处,因为市场上有数以百万计的客户可以卖给他们,而且没有收入集中度。你确实能掌控自己的路线图。反馈环路非常短。你几乎可以尝试任何想尝试的东西——“哦,我们有 50 个客户,如果做这件事搞砸了,外面还有 500 万个。“不是说要故意搞砸,但你可以实验,可以承担风险。顺便说一句,当 HubSpot 开始做 SMB 的时候,在 SMB 领域取得成功的难度是现在的整整一百倍。我们根本融不到资。太难了,因为当时常见的论调是——在软件史上,真正在 SMB 领域建立全球品牌、做到数十亿美元市值的,确实只有一家公司。之前从未有人创建过专注于 SMB 的软件公司。
SMB 领域的反向引力
我们说,“是的,我们知道。“但好处在于,一旦你搞清楚它的物理规律——因为它确实很难——你最终会……而且这一点已经得到了验证,现在有更多的成功案例了。Shopify,还有一大批公司在 SMB 领域取得了成功。关于 SMB 有一点是,在软件市场……顺便说一句,我只懂软件,我的经验范围就是这些——在软件领域,存在一种我称之为反向引力(reverse gravity)的现象。随着时间的推移,市场总是会把你往上拉。因为最终会发生的是,聪明的创始人和管理团队会看数据,然后说,“哦,事实证明,我们更大的客户留得更久,付得更多,通常 NPS 也更高,诸如此类。”
所以如果你让自己被这些数据牵引——这看起来也是一件合理的、聪明的事——你会在市场中不断被往上拉。由于这种反向引力,几乎每一家成功的软件公司,在足够长的时间跨度里最终都会变成企业级软件公司。这就是你看到的现实。反向引力的副产品就是,每家软件公司最终都成为企业级软件公司,所以你最终会与所有人竞争。而在 SMB 领域,你只与那些蠢到一直专注于 SMB 的人竞争,或者尝试这样做的人。这是一件更难完成的事,但一旦你搞清楚它的物理规律并让它运转起来,这是一个更加可持续、更加有趣的模式。所以我特别鼓励初创公司这样做——也许你可以对抗一段时间反向引力,也许……但至少从 SMB 起步,尝试让它跑通,直到它变得真的、真的、真的很痛苦,然后再多坚持一会儿。SMB 里面有太多的价值了。就是这个道理。
逆向而行:高信念、低共识押注
Lenny: 好的,这个主题在我们对话中已经出现好几次了。我问过 Chris Miller——之前的播客嘉宾——你擅长什么、我应该和你聊些什么。他说你非常擅长在别人往右的时候往左,在别人往左的时候往右。我觉得这一点在我们对话中已经体现出来了。我想问,沿着这个思路,关于你怎么思考这件事,或者实际操作的例子,还有什么你认为可能会对人们有帮助的?
Dharmesh Shah: 这就回到了高信念、低共识押注。我们大概有三件事是逆向而行的……顺便说一下,不只是我。我和我的联合创始人都有这种”往左还是往右”的倾向。我们至今还会问自己这个问题——“好吧,我们理解世界是这样运转的,这是通常的做法,但我们有没有彻底考虑过?我不是说我们应该这么做,但我们有没有考虑过这条替代路径?“所以我鼓励你,在公司发展的任何阶段,你不一定要做逆向而行的事,但你至少应该知道正向的选择是什么,并且认真讨论过。这是第一点。
我听过也给出过的一些最好的创业建议是:“初创公司应该专注于一件事,在这件事上做到世界级的、超越所有人的卓越。“而我们早期的逆向之举是——“我们要做截然相反的事。“从第一年开始,HubSpot 就决定同时构建 SEO 工具、网页分析、博客工具、内容管理,全部都做。我们构建产品的每一个品类——注意这是第一年——背后都已经有优秀的产品和优秀的公司。所以问题来了——“好吧,那你为什么要这么做?”
原因是,我们想要做好的那”一件事”,是解决客户实际存在的问题。而客户实际存在的问题并不是工具匮乏——有大量优秀的 SEO、博客等工具,所有工具都存在——但 SMB 特别缺乏把这些碎片拼凑起来的能力。对你我来说,“哦,我可以搭一个网站,装上 Google Analytics,放一个 Wufoo、Typeform 表单,我可以做所有这些事,我可以把它接上文档。没什么大不了的。这有什么难的?“但对大多数人、大多数 SMB 来说,这是一个科研项目。所以我们说,为了真正解决客户的实际问题,我们必须真正解决客户的实际问题——即使这让人不舒服,即使这是逆向共识的,即使我们的每一个直觉和每一条我们听过或给过的建议都在说我们不应该这么做。我们不应该铺那么宽。
“英里宽、英寸深”的策略
Dharmesh Shah: 但一旦我们做出这个决定——另一个经验教训就是——我们在早期就说过这句话,后来情况有所好转——假设你采取了这种做法,就是”一体化”方案,我们要做一个覆盖面很广的产品。那就好比——“好吧,我们要做到英里宽,但只有几英寸深。“这又回到了我们非常系统化的思维方式——“好吧,我们要逐一衡量我们现在进入的每一个产品品类。我们在那个品类的市场里排前三吗?如果答案是排,那就说明我们在那个品类上投入太多了。“我们不应该排前三,因为我们的价值主张不是我们是排名前三的博客工具,或者排名前三的网页分析工具之一。我们排名前三什么的并不重要,我们的价值主张是所有东西协同运作得如此之好,以至于排前三、拥有那些所谓”正确”的功能根本不重要。
一体化比单品排名重要得多。这意味着我们在任何单一品类上都不应该过度投入。这就是我们用来判断自己在某个具体方向上是否过度投入的一条启发式规则。
Lenny: 这太有意思了。我很少听到有人说”我们排前三,说明我们做错了”。我们在这方面太强了。
Dharmesh Shah: 顺便说一句,后来情况随着时间发生了变化。在创业阶段,资源非常紧张,而如今我认为在我们参与的大多数品类里,各种调研都会告诉你我们在每个品类都排前三了。但在当时资源极度匮乏的情况下,我们是在努力保持纪律和专注。
什么时候应该做宽而不是做深
Lenny: 我们一路攒下了这么多逆向而为的公司建设方法。我想听众中可能有人会觉得——“好吧,太酷了,我就去构建一堆东西,挺好的。我在 HubSpot 干过。“你有没有什么启发式规则可以分享给创始人,判断什么时候应该做宽而不是做深?你觉得什么时候这样做是合理的,什么时候应该遵循常见的建议?
Dharmesh Shah: 关键在于你在解决什么问题。我认为创始人常犯的一个错误——尤其是产品导向型创始人——是我们爱上了解决方案,而不是爱上真正的问题。你需要爱上问题。给自己施加这个约束,会逼着你去理解、至少去标注和定义你到底在解决什么问题,因为很多时候我们能描述自己的产品,但作为创始人,我们有时却很难描述真正的问题。所以如果你能论证——HubSpot 当年就做到了这一点——并不是我们一开始就灵光一现——“哦,我们要做一个一体化、什么都能做的产品。“而是在跟客户交流的过程中,我们曾考虑过——“哦,我们就做这块营销功能就好了。”
我们就做这个或者那个。但不对,客户的痛点不在这里。他们的问题不是这些东西不存在。所以,如果你通过跟客户交流、深入理解问题,得出结论需要做的不仅仅是一件事,那你就应该像我们一样给自己施加一个纪律约束——这是我们为什么要稍微做宽的原因,这是我们如何确保这种宽度不会毁掉我们的方式。你必须有自我施加的约束。你不可能两全其美。你不能说——“我们要做一体化,顺便还要做到最好。“不,你做不到最好。你可以说,我们把这个特定问题解决得最好,但我们不会在任何单一维度上拥有最好的产品。
Lenny: 太棒了,我很喜欢这个建议。
Flash Tags
Lenny: 好,我现在要换一个完全不同的方向,看看能聊到哪里。我想谈谈 flash tags。
Dharmesh Shah: 好的。
Lenny: 这是你提出来的一个东西,起源于一个问题——你是创始人,人们来找你要反馈,你随口说了一个想法,他们就把那当成金科玉律,花好几个月去执行一件事,而你觉得——“哦,那只是个想法而已。“来谈谈 flash tags 吧。
Dharmesh Shah: 这正是问题所在。再一次说明,大多数问题在每个抽象层面都存在。任何管理别人的人,任何领导者,都会遇到同样的问题。这是”扩音器效应”。有人在走廊里碰到你——我们以前还有办公室、还有走廊的时候——他们会问你一个问题。事实证明,这对创始人来说尤其如此——我对所有事情都有观点,真的是所有事情。所以你问我一个问题,我会给一个观点,这个观点不一定经过深思熟虑,也不一定正确,但我就是会有一个,对吧?挑战在于,人们会把那当作一个观点并且过度放大,而每次有人问你什么,每次你发一封邮件,都要去限定一番——“哦,顺便说一下,这只是我的个人意见”,这效率太低了。你可以在邮件里偶尔加个 IMHO 之类的,但人们会说——“哦,那只是客套”,或者之类的话,对吧?
好吧,所以 flash tags 到底是什么——我生活中大部分时间通过邮件处理事务,但我在其他媒介上也这么做。它就像我们都知道的 hashtag,但它是一个有限的、递进式的集合,我称之为”殉道光谱”——就是你在多大程度上愿意为这件事拼命。它是这样的:最底层的标签是 #FYI,#FYI 的定义是——“我看到一篇有趣的博客文章”或者”我发现了这条新闻之类的,只是告诉你一声,不需要回复。“好吧,继续你的生活。往上一级是 #suggestion。“我脑子里冒出了一个想法,因为听了 Lenny 播客上 Microsoft 的 Sam 那一期之类的,感觉这可能是一个值得探索的方向。但 #suggestion 的意思是,我仍然不期待你回复。”
Dharmesh Shah: “你不必照做我说的这件事,但如果我是你,我会这么做,或者至少我会考虑一下。“好吧,再往上一级是 #recommendation。#recommendation 的意思是——“我对此深思熟虑过。我做了一些研究,做了一些挖掘,做了一些灵魂拷问。我会这么做。“这仍然不是——“你必须这么做”,但如果你决定不做这件事,我希望你能回复我,告诉我你为什么不想做,你在探索这件事的过程中学到了什么,有没有其他替代方案之类的。所以这是一种期待回复,但仍然不是强制要求。最后一个级别是 #plea。“我恳求你。HubSpot 没有强制指令。我从内心深处、从灵魂深处反复思考过这件事,我相信这就是我们应该做的事。我恳求你,请就照做吧。”
顺便说一下,这仍然不是强制指令。现在这套体系已经被普遍接受了。它写在公司的 Wiki 上,我每天都会使用好多次。HubSpot 的其他人也在用。它的效果非常显著,因为自描述性的 hashtag 是可搜索的——你可以回过头去看,“哦,我想知道我最近提出的五个 #recommendation 结果怎么样了?“它带来了很多好处。而且它不难理解,也不难描述。
Lenny: 很有意思的一点是,这里面没有一个层级是”直接做就对了”。你说基本上没有强制指令这个概念。为什么会这样?你觉得大家会不会把 #plea 当成——“好吧,这个我真的必须照做了”?
Dharmesh Shah: 作为 HubSpot 文化的早期体现之一,我们一直相信自主性。很多公司都相信自主性,我们试图把这个理念推到尽可能远的地步。在大多数情况下我们做到了,#plea 也是其中的一部分。这就是我们信任我们的员工。我们努力招募最优秀的人才,期望他们不断学习,我们仍然希望他们拥有做出艰难决定的自主裁量权。我们称之为 DRI。这个术语不是我们发明的——Directly Responsible Individual,直接负责人。我们后面应该多聊聊决策以及决策是如何做出的。我之所以用 #plea,是因为它是一种软性的强制,而不是真正的强制。就像——好吧,顺便说一句,这种情况发生的次数,一只手都数得过来,手指头还有剩余。
那几乎从来不会走到那一步,因为要么我自己先说服了自己放弃了。就像——好吧,这就回到了那个问题:这是我愿意为之殉道的那座山吗?对吧?就像——好吧,我理解了。我可能不同意。但这是世界末日吗?它真的会造成那么大的伤害吗?而且打出那张创始人牌——“去做这件事”——的代价又是什么?它确实会发生,只是不常发生。
Lenny: 太棒了。我们会附上链接,你写过一篇关于如何使用这些标签的文章。
Dharmesh Shah: 是 flashtag.org。如果你想要一个快捷方式直接跳到那篇博客文章的话。我相信域名和快捷方式,因为我相信简洁。
决策机制
Lenny: 你提到了决策。这里面有些有趣的东西。聊聊你们在 HubSpot 是怎么做决策的吧。
Dharmesh Shah: 和大多数组织一样,HubSpot 历史上在决策方面做得很差。这完全是一个光谱。我认为决策一直是最难的事情,因为很难找到平衡点,我们也在不断演进。我觉得我们现在比早年好多了。我们犯的错误,大多数人也会犯。我来告诉你我们现在做对的事情。首先,我们非常注重数据。我们喜欢做出数据驱动的决策,但做决策的是人,不是数据。这是第一件事。很多由极客创建的公司会说——“哦,我们有一张图表显示这个,所以我们应该去做那个。“这行不通。数据很棒,但我们还会指定一个职位来负责做决策,而且那不一定是该团队的主管。
就是——好吧,由他们来真正承担这个决策的所有权。这个想法也不是我们发明的,它在 HubSpot 之前就存在很久了,但它确实有效。你必须决定由谁来做一个决策,这几乎本身就是头号决策。就像——好吧,不要选你不信任的人去做决策。这是第二件事。而我们曾经犯的最大的错误——过去比现在犯得多得多——是……Amazon 对此有过很好的表述,他们叫做 disagree and commit。我们做了一个小小的改动,我来告诉你为什么我们稍作了调整。我们的是 debate, decide, unite——辩论、决定、团结。这是我们在 HubSpot 文化中使用的短语,但核心理念是:一旦做出决策,就要围绕这个决策达成一致。因为无论组织规模多大,总会有人不同意那个决策——他们之前在辩论中站在另一边,他们本来会选择方案 A,而最终我们选了方案 B。
但极其重要的是,即使一个决策被做出,而我们决定要做的事情与你所相信的、曾为之辩论的事情相违背,你在这件事上输了——没关系。这很难做到,但如此重要、如此有价值。我来告诉你为什么我们对 Amazon 的版本做了微调。disagree and commit 听起来比我们喜欢的要更生硬一些。就是——你可以不认同,没问题,但然后你就得 F 开头地老老实实执行,对吧?我觉得这有点……好吧,我有偏见。所以我们的”辩论”是——好吧,让我们公开辩论,把所有选项摆到桌面上,切实做出决定,然后围绕这个决策”团结”起来,而不仅仅是——就像——让我们在这个决策上凝聚在一起。这听起来更 HubSpot 一些。总之。
Lenny: 确实友好得多。你说话的时候我就在想,debate, decide, unite 对比 disagree and commit,还有 Amazon 那条领导力原则里”要有骨气”什么的。
Dharmesh Shah: 另一件事我告诉你,这有一部分是 HubSpot 的做法,有一部分是我个人的做法,而且这一点也不会让你惊讶——我喜欢用系统化的方式来思考决策。我给你举个例子。当我面对一个决策或一件我想做的事情时,我问自己的问题是——“好吧,如果我能写一段 Python 代码、一个 Python 函数来得出这个问题的答案,变量或系数会是什么?“好吧,假设你不用编程——如果你要用 Excel 电子表格来做这个决策,列会是什么?你不必确定权重,但至少确定因素。下一步就是——好吧,以下是我认为在做出这个决策时重要的因素。我要在这七个选项之间做决定。这些是我要看的东西。
你可以用这个方法来选学校,它确实管用。你不必把它推到极致,但走一遍这个过程本身就很有价值。就像——“好吧,第一步,如果我能把我认为应该影响这个决策的因素识别出来,那已经是 80% 的胜利了。“不是因为把它从脑子里倒出来而已。如果整个团队共同协作说——“好吧,以下是我认为应该影响这个决策的所有因素”,那就更好了。然后第二步是——好吧,我们无法分配精确的权重,所以我们还没法写出 Excel 公式或 Python 代码,但我们能不能至少给这些因素排个序?这个因素比那个因素更重要。现在你就有了一个更接近的近似感知。你仍然不会用电子表格来做最终决策,但做决策的人比没经过这个练习之前拥有了好得多的信息基础。
决策的卡路里原则
Dharmesh Shah: 人们在决策上常犯的一个错误——Amazon 也写过这个话题,我来说说我的看法——就是你在某个决策上消耗的”卡路里”应该与这个决策的后果成正比,就这么简单。他把决策分为单向门和双向门,这是一个二元的启发式方法,我很喜欢它的优雅和简洁。我把它再往前推了一步——字面上就应该成比例。就像,好吧,如果你有一个非常大、非常大的决策,改变起来代价极高,那就应该在这个决策上投入大量卡路里。你可能仍然会做错,但这是值得的。而且这不仅适用于决策,也适用于活动、押注以及你要做的各种事情——你投入的卡路里应该与结果、你追求的目标以及它的价值成正比。
默认立场:说不
这也是我试图应用到自己生活中的事情之一。就像,好吧,我的极客本性想让我钻到九英里深、开始写代码、做这个做那个,但我发现那一个简单的问题就够了——好吧,这是不是一个一阶近似,我只需要识别出那些因素就行了?另外顺便说一句,还有一件事很有帮助,我们后面会聊到。第一点,我在大多数事情上的默认立场——我不是以消极的方式说这个——是”不”。就像,“哦,闪亮的新东西。我应该去做吗?""不。“然后我得强迫自己——“好吧,我需要找到说好的理由。“我得强迫自己,因为我很容易说我喜欢新东西、喜欢新认识的人、喜欢所有这些。所以我在博客上写过一篇文章,标题就是”Sorry, But No”,那篇文章对我改变很大。
这个想法可以追溯到其他的第一性原理——时间是少数几个常量之一。所以每次你对一件事说好,定义上你就是在对另一件事说不。这又回到了我们早年做产品时用的一个启发式原则——当你要加一个东西进去的时候,必须拿出一个东西来,因为我们想尽量保持复杂度大致恒定。同样的道理——对事情说好没问题,但之后你得根据你投入的卡路里强迫自己说——“好吧,我要从日程中、从生活中拿掉什么,才能为这件事腾出空间?“因为事实证明,尽管我尽了最大努力,我还是没法弯曲时空法则。那我怎么为这个我想说好的新东西腾出空间呢?
评估创意的框架
Lenny: 我太喜欢你思维方式了。我很期待接下来聊文化——聊这一切如何与文化相连,以及你如何通过这些框架和视角来思考文化。不过在此之前,你刚才暗示了你如何选择做什么。我知道你有一个挺有意思的框架,用来判断哪些创意是好的、哪些值得投入。我想无论作为创始人还是产品领导者,你是怎么思考什么值得投入的?
Dharmesh Shah: 再想象一下那个 Excel 电子表格。假设你在评估一个创意——创业想法、新产品、新功能,什么都行,我觉得这个方法都适用。你应该看的第一件事是——我喜欢用零到十的量表来衡量事物,因为我喜欢可量化的东西。我喜欢能把 X 乘以 Y,不喜欢字母评级,这是我的一个小癖好。总之,给它一个数字。就像,好吧,如果这件事成功了——假设我们只是在选一个新的创业想法,我在考虑创办一个新的业务——如果它成功了,它能变成什么?这个结果的量级有多大?不管你怎么定义——可以是影响力,可以是收入,可以是市值,可以是投资组合,什么都行。这不重要——我是说,它重要,但对这个讨论来说不相关。
那就是潜力。第二个因素是成功的概率。现在我要在这里停下来告诉你,人们最常犯的错误是先想概率,而没有先强迫自己把潜力想清楚。结果就是——他们会用一个过滤器说:“哦,我只有十分之一的机会做成这件事,只有 10% 的成功率什么的”,那就不做了。但如果是十分之一的机会拿到一百亿呢,那也许值得。也许不值得。有时候这取决于你的人生阶段和你能承受的风险,但比如说——顺便说一句,当我提到概率的时候,这是大头。你可以用多种方式来计算。你可以说——“好的,这里有 X% 的概率”,你可以用统计学做事件分析——就像,好的,这是路径 A,有 Y% 的概率会发生,结果会是这个;Z% 的概率或者 X。总之,你列出各种可能的情况,得出所谓的期望值(expected value)。总之,假设我们只看两个分支——“哦,我有五五开的概率赚到一千万。“我们只看一个分支来简化。那么期望值就是五百万。这就是期望值——你有五五开的概率获得一千万的结果。假设你有 10% 的概率获得一亿的结果,或者五亿,不管多少。所以在数学上,好吧,你至少应该考虑另一个选项,即使概率更低,因为期望值更高。好?这是第一个要点。总之,我们继续讲剩下的。
所以先看潜在结果,再看概率,不要反过来——因为反过来你就会用心理过滤把一些本来不该扔掉的创意扔掉了。第三件事是我所说的热情(passion)或接近度(proximity)——你是否关心这件事,能否真正投入其中?第四个因素——因为我需要一个 P 开头的词——是能力(prowess)。你是否拥有某些资产?某种让你处于独特位置的东西?比如你在做一个新产品,你有可以复用的现有代码,你有已经可以触达的市场,这个特定创意有某种你在追求它时拥有不公平优势的地方——而这会影响你成功的概率,这也是你需要把这些都考虑进去的另一个原因。再说一次,没有任何单一因素是决定性的。你不会只选那个期望值最高的——潜力乘以概率什么的。而是你得把这些都看一看,结合自己的情况。但有这个思考过程是有帮助的,有这样的纪律性说——“它能有多大?概率是多少?我有多在乎解决这个这个问题?这是我能做两年、十年、二十年的事吗?“最后——“为什么是我?为什么是我的公司?我们凭什么在这件事上成功?为什么我们成功的概率,即使低,也比其他人高?“为什么?
Lenny: 好的,我已经在脑海里看到这个电子表格了。对于你的每个创意,你基本上都有一个列——潜力、成功概率。
Dharmesh Shah: 对。
Lenny: 接近度,你说的是?就是怎么……
Dharmesh Shah: 确切地说是热情。
Lenny: 好的,热情。
创业想法评估的四维框架
Dharmesh Shah: 对。我最开始用的是”热情”(passion),但我对”热情”这个词有意见,因为它有歧义,因为在创业圈里我们总在谈论它。就像,“哦,追随你的热情。“当然,但是你知道吗?今天存在的绝大多数公司,哪怕是那些拥有优秀创始人的成功公司,他们并不一定是追随了自己的热情。他们可能在某个具体问题、某个市场细分领域中碰巧投入进去了,但没关系——在你深入某个领域的过程中对它产生热情,这完全是可以的。
Lenny: 对,我投资过的一家最喜欢的创业公司叫 Zip,是一个采购平台。创始人在进入这个领域时对采购毫无热情,他们只是看到了一个巨大的机会。他们认为自己可以做出一个更好的产品,然后在实际动手做、找到产品的过程中,他们就逐渐对它产生热情了。
Dharmesh Shah: 就是这么回事。所以,是的,如果你能找到自己真正的使命,找到你在这个世界上注定要做的事,而且你有机会去做,那当然很好。但这只占极小的一部分人。大多数情况下不会这样。你甚至可能不知道自己的热情在哪里,尤其是如果你是第一次创业的人。就好比说,好吧,我还没见过这个世界,我不知道。
Lenny: 好的。所以如果大家在为创业想法做电子表格的话,评估维度就是——潜力、成功概率、热情、能力。
Dharmesh Shah: 对。
定义公司文化:Culture Code 的诞生
Lenny: 好的,太棒了。好,我们来聊聊文化。我知道你在构建 HubSpot 的文化、创建最初的文化体系上花了很多时间。你做了一份很了不起的 Culture Code 幻灯片,我记得是 128 页。也许我们可以深入聊聊,定义一家公司的文化并长期维护这个文化,这个过程是怎样的?
Dharmesh Shah: 我会说,一开始很糟糕。不,是整个起步阶段都很糟糕。好吧,我给你讲一下非常简短的来龙去脉。我和我的联合创始人正在进行我们例行的创始人会议或创始人晚餐,他之前参加了一个 CEO 小组,然后他说,“哦,Dharmesh,我听说这个文化的东西非常重要。“顺便说一句,在此之前,“文化”这个词在 HubSpot 的走廊里从未被提起过。而当时已经是公司成立多少年了?大概是三四年之后了。我可以查记录确认,因为即使是 Brian 和我,即使在人们还在办公室上班的年代,我们大部分沟通也是通过邮件和异步进行的。我们一直都是书面沟通的忠实拥趸。总之,我可以告诉你”文化”这个词有没有被使用过——答案是从来没有过。
他说,“哦,我听说”——因为他在跟那些人交流——“文化超级重要。Dharmesh,你能去做这件事吗?“我说,“好吧 Brian,所有公司里的所有人当中,我绝对是倒数那一档的,如果不是最不合适的人选的话。不是我不喜欢人,我只是不太喜欢和人待在一起。而据我对’文化’的有限了解,它好像需要和人打交道。我不知道,但看起来我不该是做这件事的人。“但不管怎样,我还是个有团队精神的人。我想,好吧,这确实看起来很重要,于是我就接下了这个任务。但我真的不太知道该怎么做,因为当他说”Dharmesh,你能去做这件事吗?“——这到底是什么意思?我把它理解为:文化其实已经存在了。我们已经有了一种文化。不管它是什么样的,它运转得还不错,公司发展得挺好,员工看起来也开心。
所以我说,“好吧,我真正想做的其实是描述现有的文化。“所以这不是创造文化,而是把文化 articulating 出来,让我们知道它是什么。然后,我那个极客的一面又出现了。我问自己一个问题——我真的就是问了这个问题——“如果我能写一个 Python 函数来衡量一个新人加入 HubSpot 后成功的概率,这个函数的变量会是什么?“如果我能测量这些因素,如果我有一个真实的函数说,“哦,在 0 到 10 的量表上……”
如果我有一个真值函数说,“哦,在 0 到 10 的量表上,他们在这个、这个、这个、这个、这个、这个上的得分是多少”,我就能近似地预测他们在 HubSpot 会有多成功、对 HubSpot 有多大贡献。这就是我做的练习。
于是我做了一次内部调查,我来告诉你为什么一开始这段经历如此糟糕,因为这也是一个很好的故事。我接下这个任务后,给全公司发了一封邮件:“大家好,我要开始研究文化了,我会发一份调查问卷,想了解一下情况以便……”很好。那封邮件引发的负面反应和情绪之强烈,直到今天仍排在我经历过的最糟糕事情的前三名。
人们的反应如此负面、如此强烈,而我完全没有预料到。我告诉你,有一条回复真的让人心碎——“Dharmesh,我们现在在谈文化,下一步是不是要在墙上贴什么’卓越’的海报了?HubSpot 不是我以为我加入的那家公司了。“那一刻我内心深处非常受挫,我完全不知道”文化”这个概念承载了多少额外含义。然后你深入研究就会发现,哦,原来如此多的公司把这件事做得如此之差,根本没有践行自己说的东西,所以人们才会有这种负面反应。
所以,我回到自己的思路上。我想,“我明白了。那我就来描述我们已有的东西,来写这个真值函数。“于是我写了那套幻灯片,正好 16 页,回答的就是这一个问题:如果我能写一段 Python 代码,变量是什么,也就是我们要在人身上寻找哪些属性。我之所以想到这个特定的问题,是因为在做调查时,我做了 NPS 调查——大家都知道那个发给客户的 NPS 问卷——“在 0 到 10 的量表上,你有多大可能推荐 HubSpot 作为一个工作场所?“另一个 NPS 问题是”你为什么给出这个评分?”
结果,我发现了一件令人不安的事:大家都很开心,而他们开心的原因是因为 HubSpot 的其他人。这是排名第一的原因,这既是好事也是坏事。好比说,好吧,这很棒,但这没有告诉我任何有用的信息。它是一个自我强化的递归循环。于是我就在想,“好吧,那是什么让 HubSpot 的人成为了 HubSpot 的人?让我们来找出那些属性,定义一下’成为这些让别人开心的人中的一员’意味着什么。”
于是我们总结出了最初的那套属性,体现在那套幻灯片中。我把那套幻灯片叫做 Culture Code deck,之所以叫 Culture Code,不是因为它是一套行为准则,或者道德规范,或者”公司应该这样运营”——而是因为如果我能写代码来定义 HubSpot 的文化,这些代码就会长这个样子。从那以后,我一直在问自己的问题是:如果我们能有一个运行这家公司的操作系统——这只是我极客的那一面在想,不是说真的能做到——但那个操作系统会是什么样子?
文化即产品
Dharmesh Shah: 我很少有真正好的原创想法。其中一个就是:文化是一种产品,句号。每家公司都在打造两个产品——一个是为客户打造的产品,另一个是为团队打造的产品。文化就是后者,是你为团队打造的产品。假设我们先承认这一点就是对的,先姑且听我说一分钟。如果你假设这是对的,那么从这个假设出发会推导出很多事情——“哦,既然如此,就像我们绝不会在不跟客户沟通的情况下打造产品一样,我们也不应该在不跟这个产品的客户——也就是员工——沟通的情况下打造文化。“于是我们重新做了最初的那份 NPS 调查,从那以后每个季度都持续在做,以测量这个文化产品的 NPS。
另一件事是,好吧,人类历史上从来没有哪个产品经理会说,“哦,我做完了这个产品,它棒极了,我收工了。“没有人会这么说。而我认为——我有偏见——这是创始人犯的头号错误:他们说,“哦,我们有一个很棒的文化,我的工作就是守护这个文化。“那不是你的工作,因为客户的需求会变,员工的需求也会变,对吧?文化的存在是为了帮助优秀的人为公司做出伟大的事情。这就是文化存在的理由。而随着公司规模的扩大,他们后来需要的东西是不一样的。所以,就像你绝不会冻结一个产品的代码行一样,你也绝不会冻结一个文化的代码行。文化应该像迭代产品一样被迭代。
文化中的 bug
还有很多其他类似的事情。就像在产品中,你会从客户那里得到反馈——“哦,这里是 bug”——我们实际上也叫它们 bug。就是,“哦,文化里也有 bug。“在全体会议上——我们会收回那些调查结果。再说一次,这是透明度的理念。每一份调查回复,在提交者身份上是匿名的,但我们会全部公开。量化评分也好,主观内容也好,全部公开。然后我们会做分类,在下一次全体会议上——我们叫它 all-minds……总之,一点——
Lenny: 我喜欢这个,all-minds(全脑会议)。
Dharmesh Shah: All-minds。
然后就是,“好的,以下是我们听到的。“就像我们在产品会议上做的那样——“客户反馈了这些。“然后,“这是你发现的 bug,这是我们打算修复的。“当我们说要修复一些问题时,我们会做出承诺——“你说这个坏了,好的,我们会处理。“还有一些是这样的——“是的,我知道你不喜欢它现在的样子,但它就是按设计运行的”——这在产品中我们一直在做,对吧?而且你应该有权这样做。就是,没错,有些人会不同意,可能有各种原因让我们认为它需要保持现状——我们不是说永远不变——但就目前而言,我们不优先处理这个,我们不打算修复你认为需要修复的那个东西,不管出于什么原因。
所以这里的教训是,我学到的少数几条我认为普遍适用的东西,这一条肯定是其中之一。从”文化即产品”这个观念中能衍生出如此多的美好与善意——如果你把员工当作客户,把文化当作产品,它就消除了”做文化”这件事的抽象感,也消除了我在早期从 HubSpot 团队那里听到的一种说法——“哦,文化不是你真的可以去’打造’的东西,因为那显得很假,很不真实,就好像你要——“不是的,我们打造文化的方式跟打造产品一样。我不是说我们要强加于人,不是说要在孤立中做这件事,但我们会持续地去做它,它确实是你所呈现出来的东西,而且它不仅仅取决于创始人。
Lenny: 哇,多么精彩的故事。你确实让我达到了 LPM 的标准,我觉得这期播客到目前为止至少每分钟有一个笑声。太棒了。而且我觉得你刚才提出的这个观点——作为领导者,你的工作不是守护文化——我认为非常深刻,因为我觉得大多数人不是这样想的。他们认为自己的工作是”我们需要在成长过程中维持我们的文化”——“一切都在消退,我们的文化在改变”——但你的观点是,这其实是好事,这是好事。
核心价值与联邦/州法的类比
Dharmesh Shah: 回到产品这个比喻,我觉得这是一个重要的事情。这是在你接受”文化即产品”之后的二阶教训:几乎每个人都会面对的一个挑战是,你会拥有这些东西……顺便说一下,文化是产品,并不意味着不存在某些人可能称之为核心价值观的东西,对吧?因为我们在产品中也有这种情况。产品有自己信奉的核心价值——比如我们相信我们的产品是为 SMB 打造的,要好用——我们给自己施加了这些有意的约束,即使我们不断迭代,也不会迭代掉这些东西,这又回到了那个高信念押注的概念。
但并不是产品中的每一样东西都是核心价值,对吧?所以也不是你早年文化中的每一样东西都是核心价值。比如透明度就是一个长期存续的东西,对吧?它是我们核心价值观的一部分,即使经历了 IPO 的麻烦时期,我们也坚持了下来。我之前分享过那个故事。然后也有一些我们改变的东西——比如,我们相信精英制度,相信扁平化组织,但它的具体表现形式是没有头衔,对吧?于是我们改了那个。就是,“好的,有些东西……”
还有一件事。这是一个混合比喻的问题,因为我还没想到一个更好的比喻——你会有一些属于核心的东西……我现在的描述方式,我就直接说出来了,我很懊恼自己没想到一个更好的比喻——就是联邦法与州法的类比。在美国,我们有联邦法,规定”好的,这些是所有公民不可剥夺的权利,这些是核心的东西。“但然后我们还有各州的东西,对吧?就是,“哦,对于这个特定的州,在这些特定方面,他们可以针对自己的州做优化。“就是,“是的,我们理解这是总的规则,但对于我们这个州,我们要这样做。”
我们在 HubSpot 也有类似的东西。我们有一套核心的东西属于文化的一部分,就是,“好的,这是我们的信念,“然后各个小组可以有自己的东西——比如我们的工程负责人可以说,“是的,我知道总体的、创始人的倾向是 X,但对于我这个特定的小组,那不太说得通。我们要这样做。”
举一个例子。HubSpot 早年有一个核心理念——我们不在乎你在什么时间工作,从几点到几点。就是完全不在乎。我们不追踪,也不在意。但随着时间推移,“嗯,我们有一个客户支持部门,“而那个部门,事实证明工作时间对我们的客户是重要的,对吧?我们需要有覆盖之类的。就是,“好的,这倒也说得通,“对吧?
文化的自我实现与透明度的取舍
Dharmesh Shah: 但你还是要选择你的战场,选定之后就要坚决反击。比如透明度,曾经有各个组织的负责人对我说,“好吧 Dharmesh,我知道你和 Brian 都信奉透明度,但让销售团队里的 SDR 也能看到高层财务数据,这很低效。浪费时间,也会分散注意力。“这就是他们的论点。我的回应是,“你可能是对的,但这是一个二元决策——我们对透明度的重视程度决定了我们要……”对,也许它确实会分散注意力,没关系。还有人说,“我不想让我的团队看到这些,因为我不想……”你知道,当时我们招聘速度很快,有人会说这等于把大量信任交给了员工之类的。我的态度是,没错,但这又引出了另一条经验——文化随着时间推移会变成自我实现的预言。
所以早年有一个很大的教训。当我把第一版公开对外的 Culture Code 发布出去之后,团队给了反馈:“Dharmesh,你在 Culture Code 幻灯片里说的这些东西,并不是真的。“它当时还没真到可以作为一个正式声明的程度。于是我在幻灯片里加了我所说的” liner notes”(旁注),就是一些补充说明。比如,“这个目前还不是真的,更多是一个愿景。这是我们希望它成为的样子,但我们在这一点上还没有真正践行。“
愿景声明的自我实现效应
我没有把它删掉,因为我希望它作为一个愿景存在。而接下来发生的事情是我当初没有预见到的,我也不能居功。随着时间的推移,那些原本是愿景的东西变得越来越真实,恰恰是因为它们被写在了幻灯片里——因为当新员工加入时,他们会看到,“哦,这是我们向往的东西,那我就把它当作 HubSpot 真实的文化来对待,“而这正是我们想要的。所以说,表达愿景是完全可以的,只要坦诚标注。不要把不真实的东西当作既成事实来呈现,但可以把你希望成真的东西写出来。
Lenny: 哇,这里面有太多经验了。这其实让我想起了 Airbnb。他们制定最初的六个核心价值观时我就在场,其中一个是”简化”(simplify)。然后过了若干年,创始人意识到,“我们其实并不擅长这个,“而价值观应该反映你是什么样的人,而不是你向往成为什么样的人。我很欣赏你找到的这种中间路线——“我把它放进去,但会注明这不一定代表我们的现状。“我们当时没有这么做,那本来是个聪明的做法。所以他们把那个核心价值观去掉了,说”我们其实不是这样的”,从六个变成了四个。还有一个也被移除了。我印象很深,因为在会议上会有人提起,“简化。不对,这件事其实真的很复杂。“
关于 AI 的思考
Lenny: 好,接下来我要做的是,关于 AI 我还有一个问题,为了保证时间,我打算把闪电问答环节砍掉,希望你可以接受。
Dharmesh Shah: 完全没问题。
Lenny: 好的。我知道 AI 是你现在非常关注的事情。
Dharmesh Shah: 顺便说一句,这正是我的邪恶计划——让你不得不跳过闪电问答,因为那是我最不擅长的东西,我真的不擅长……总之,你知道那是我最害怕的环节。
Lenny: 你在下四维象棋呢。你精确计算了每个回答的时间,“我要把每个回答都拉到足够长,这样刚好可以跳过闪电问答。”
Lenny: 我知道你在 AI 上花了很多时间。现在很多人都在关注这个领域。你做了一个叫 ChatSpot 的产品。它现在还叫 ChatSpot 吗?是你改了名字的那个产品,还是仍然叫 ChatSpot?
Dharmesh Shah: 仍然叫 ChatSpot。
Lenny: 仍然叫 ChatSpot。而且我知道你已经把它交给了一个专门的团队在做了。
Dharmesh Shah: 对。
Lenny: 不再只是你一个人的项目了。
Dharmesh Shah: 对。
Lenny: 我不太确定该从哪个角度切入,但就是,为什么 AI 如此让你牵挂?你觉得人们应该把时间花在哪里?你觉得事情正在朝什么方向发展?关于 AI,你脑海里浮现的是什么?
AI 是认知规模化的开端
Dharmesh Shah: 显然,不只是我一个人在兴奋,但我想说的是——我马上就要迎来从事商业软件工作的 30 周年了。我在这个领域干了很长很长时间,这是我所知道的一切,也是我一直在做的一切。而上一次我这么兴奋,还是在九十年代中后期互联网刚出现的时候。我当时就觉得,“哦,这真真切切地改变了一切。“它与所有事情都相关,所有东西都会被这个叫互联网的事物所影响。那时我才二十出头。
我也经历了移动时代,经历了社交时代,我见过很多让人们极其兴奋的东西,但在我看来说没有哪件事在震级上能比得上 AI。原因在于——如果你想想早期的个人电脑,PC 大规模地赋予了我们计算能力,这非常了不起。互联网则实际上大规模地赋予了我们分发能力——它让我们能够把信息、产品等等大规模地传播出去,因为一切都被连接了起来,而在此之前不是这样的。AI 现在赋予我们的,是大规模的认知能力——真正放大人类大脑的能力。这从未实现过。不是说我们没有在尝试,但现在我们有了切实的证据和真正可用的产品。我们用过 ChatGPT 和类似的产品。所以它是改变人生的。对于做产品的人——你们中很多人都是——AI……
创业者的历史性机遇
Dharmesh Shah: 顺便说,互联网释放了一个巨大的机会,尤其是对创业公司来说。就是那种,“哦,我们可以用这个叫做互联网的新东西重新想象世界,这在一年前还是不可能的,现在可能了。我们能解锁什么?“大量伟大的公司由此诞生。我认为类似的事情在这里也会发生。
从命令式到声明式的范式转变
Dharmesh Shah: 有几个方向性的东西,从实操角度来看,可以作为我那个创意筛选表格里的候选向量,这些方向适用于各个行业。作为软件公司和产品构建者,我们大多数人一直称自己的产品为”直观的产品”,我们确实会这么说。但事实证明,这并不是真的。因为作为某个产品的用户,不管它是什么,你脑子里有一个心智模型——你想要完成的事情——然后你必须把它翻译成一系列点击、拖拽、触摸和滑动操作,才能在你使用的软件产品中完成你想做的事情。不同的产品之间的区别只在于这种翻译的程度不同。
但这里存在一种阻抗失配——你必须把你脑子里的东西翻译出来,比如”我想在 Photoshop 里去掉这个背景”、“我想在 HubSpot 里创建这个报告”。现在的机会在于,我们正在从一个命令式模型——工程师的说法就是一步一步地,你给计算机下达指令,“做这个,做这个,做这个,做这个,“然后期望得到你想要的输出——转向工程师所说的声明式模型。声明式模型是你描述你想要的结果,而不是达到结果的步骤。大多数人最熟悉的类比就是 SQL,那个查询语言。
Dharmesh Shah: 在 SQL 出现之前,如果你想从系统里提取数据,你得写代码——“好,遍历所有客户记录,然后过滤掉收入低于一百万美元的,再把结果给我。“而 SQL 说,“不,直接加个 where 子句就行了。描述你想要什么数据、要哪些列、按什么排序、应用什么筛选条件,“然后数据库自己搞清楚怎么把你想要的东西给你。
现在我们有能力对软件界面做同样的事情了,对吧?不再是”点这里,拖到那里”,而是”给我生成一份报告——所有欧洲客户,上季度签约的,ARR 超过 10 万的。“你直接把脑子里想的东西表达出来,用什么媒介都行——文字、语音,无所谓——软件现在真的可以……所以现在技术已经能把这件事做成了。以前不行,现在行了。
我知道以前不行,因为六年半以前,我做了一个叫 GrowthBot 的产品。那是一个面向 HubSpot 软件和一般商业软件的聊天机器人,主要服务营销和销售人员。本质上就是现在的 ChatSpot。GrowthBot 这个产品非常棒,只有一个微不足道的小问题——它不能用。构思非常精妙,想法确实出色,就是不能用,因为技术还不到位。你根本无法以任何可信度来实现它,因为 AI 还没有……而现在可以了。这也正是促使我去验证自己理论的原因——我对那个想法一直耿耿于怀,对吧?但我认为——我并不是说 Web 和移动界面会消失,但在几乎所有软件中,确实存在一类场景,我们可以通过消除那层翻译来让事情对人更简单,真正实现”直观”。是的。
所以,感谢大家来听我的 TED 演讲,或者我的播客。
Lenny: 看你做事很有意思的一点是,你在 AI 这块基本是公开构建——你的学习方式就是动手做、动手试。你发布 ChatSpot 的时候发推说,花了你很多很多个深夜的工作。
对于那些想搞清楚正在发生什么、想走在趋势前面、想培养未来仍然有价值的技能的人,你有什么建议吗?就是不停地做东西吗,还是你有其他推荐?
Dharmesh Shah: 有,这是一个非常简单实用的建议,对软件工程师也一样——不要因为你觉得某个东西”值得学”就去学它。找一个你真正关心的问题,然后用那个新东西去尝试解决它。如果你有一个具体要完成的事情——比如回到公开演讲那个话题——如果我当初说”哦,我想学学幽默和脱口秀”,我觉得我什么也学不成。但我心里有一个目标——“我需要能站上舞台,掌控全场,而且我有这个可衡量的指标。”
所以不管你在什么行业,不管你做什么职业,我保证 AI 有办法帮到你,哪怕就用目前市面上已有的工具。要么使用这些工具,要么在现有的 API 之上构建工具——API 现在非常容易获取,启动成本也非常低。对,直接做,然后迭代。我也非常推崇公开学习。这就是为什么 Culture Code 放出去了。既然你接受”文化是一个产品”这个前提,那与其只从 HubSpot 员工那里获取反馈,不如把它展示给全世界?因为互联网在告诉你”你是个白痴”这件事上,能力极为出色——这可以说是它的核心竞争力之一。所以我建议,心里先有一个目标,然后把它写出来。我总体上非常推崇写作,但没错——公开做。
Lenny: 我很喜欢这个建议。
还有什么想分享的、想聊的,或者在结束前想留给听众的吗?
Dharmesh Shah: 我们覆盖了很多领域。
Lenny: 是的,确实。确实。
Dharmesh Shah: 我可能该见好就收了。我就留一句话吧。很多年前我给自己下了一个成功的定义,至今仍然认同,HubSpot 其实也认同——成功就是让那些相信你的人看起来像天才。就这一句。我喜欢它的地方在于,它不是向内看的。这指的是你的客户,那些在你还是创始人的时候就加入公司的员工——他们傻乎乎地愿意跟你一起干——那些相信早期那个完全不堪入目的产品的客户,那些相信你的投资人,所有这些人。尽你所能,让他们回想起来的时候觉得:“哦,我当初相信 Lenny、相信他在做的事情,真是太明智了。“所以……
Lenny: 这真的很美。Dharmesh,你太棒了。非常感谢你来。
Dharmesh Shah: 这次很有趣。很高兴我们最终把它做成了。
Lenny: 同感。我们聊了特别多。我想聊的全聊到了,还多出了不少。大家如果想在网上关注你的动态,或者有问题想联系你,去哪里找你?听众可以怎么帮到你?
Dharmesh Shah: 好。我在所有平台上都是 Dharmesh,D-H-A-R-M-E-S-H。我不仅找得到,而且基本上在互联网上无法回避。在社交媒体上关注我,告诉我哪里蠢、能学到什么。还有一件事——留一条 YouTube 评论,告诉我你最喜欢的 Lenny 的哪几期节目,我好回去看,因为我最近正在刷。很想知道。
Lenny: 说到这个,我知道你有自己的内容网站,有你的 YouTube 频道。大家去哪里找这些东西?
Dharmesh Shah: 直接去 dharmesh.com 就行,那里有其他地方的链接。实际上我最常活跃在 LinkedIn 上,这本身是一个可以聊一整天的话题,不过……
Lenny: 好。太棒了。那就不耽误你了。Dharmesh,非常感谢你来。
Dharmesh Shah: 谢谢。再见,大家。
Lenny: 非常感谢收听。如果你觉得这期有价值,可以在 Apple Podcasts、Spotify 或你喜欢的播客应用上订阅。也请考虑给我们评分或写评论,这真的能帮助更多听众找到这个播客。你可以在 LennysPodcast.com 找到所有往期节目或了解更多关于节目的信息。下期见。
术语表
| 原文 | 中文 |
|---|---|
| 4D chess | 四维象棋 |
| all-hands meeting | 全体会议 |
| ARR (Annual Recurring Revenue) | ARR(年度经常性收入) |
| bi-modal | 双峰 |
| Bowling Pin | Bowling Pin(保龄球策略,保留原文) |
| Brian | Brian(Dharmesh 的联合创始人,保留原文) |
| Cambridge, Massachusetts | 马萨诸塞州剑桥市 |
| carrying cost | 持有成本 |
| chat.com | chat.com(域名,保留原文) |
| ChatSpot | ChatSpot(产品名,保留原文) |
| Chris Miller | Chris Miller(人名,保留原文) |
| code of conduct | 行为准则 |
| core values | 核心价值观 |
| Culture Code | Culture Code(文化代码,HubSpot 的文化幻灯片,保留原文) |
| debate, decide, unite | debate, decide, unite(辩论、决定、团结) |
| declarative model | 声明式模型 |
| designated insiders | 指定内部人 |
| Dharmesh Shah | Dharmesh Shah(人名,保留原文) |
| dimensional complexity | 维度复杂性 |
| disagree and commit | disagree and commit(不认同但执行,Amazon 的决策原则) |
| DRI (Directly Responsible Individual) | DRI(直接负责人) |
| dying on the hill spectrum | 殉道光谱 |
| efficient market | 有效市场 |
| Elon | Elon(指 Elon Musk,保留原文写法) |
| entropy | 熵 |
| expected value | 期望值 |
| feedback loops | 反馈环路 |
| first principles | 第一性原理 |
| first-order thinking | 一阶思维 |
| flash tags | flash tags(Dharmesh 提出的分级标签系统,保留原文) |
| flat organization | 扁平化组织 |
| founding principle | 创始原则 |
| freemium | 免费增值 |
| functional decomposition | 功能分解 |
| Geoffrey Moore | Geoffrey Moore(人名,保留原文) |
| go-to market strategy | 市场进入策略 |
| GrowthBot | GrowthBot(产品名,保留原文) |
| guardrails | 护栏 |
| hashtag | hashtag(标签) |
| heuristic | 启发式方法 |
| high conviction, low consensus bets | 高信念、低共识押注 |
| high-stakes speaking | 高风险演讲 |
| HubSpot | HubSpot(公司名,保留原文) |
| impedance mismatch | 阻抗失配 |
| imperative model | 命令式模型 |
| inalienable rights | 不可剥夺的权利 |
| Kipp | Kipp(人名,保留原文) |
| learning in public | 公开学习 |
| Lenny | Lenny(播客主持人,保留原文) |
| lightning round | 闪电问答 |
| liner notes | liner notes(旁注,保留原文) |
| LPM (Laughs Per Minute) | LPM(每分钟笑声次数) |
| marked to market | 标记市值 |
| market cap | 市值 |
| mathematical induction | 数学归纳法 |
| megaphone issue | 扩音器效应 |
| mental model | 心智模型 |
| meritocracy | 精英制度 |
| MicroWare | MicroWare(为小规模用户群定制的软件,保留原文) |
| N+1 | N+1(逐次递增策略,保留原文) |
| NPS (Net Promoter Score) | NPS(净推荐值) |
| one-way door / two-way door | 单向门 / 双向门 |
| operating system | 操作系统 |
| opt in | opt-in(选择加入) |
| Peter Thiel | Peter Thiel(人名,保留原文) |
| prowess | 能力 |
| punchline | 包袱 |
| recursive loop | 递归循环 |
| revenue concentration | 收入集中度 |
| reverse gravity | 反向引力 |
| roadshow | 路演 |
| SDR (Sales Development Representative) | SDR(销售开发代表) |
| second law of thermodynamics | 热力学第二定律 |
| second-order thinking | 二阶思维 |
| SMB (Small and Medium Business) | SMB(中小企业) |
| SoloWare | SoloWare(为单人定制的软件,保留原文) |
| SQL | SQL(结构化查询语言,保留原文) |
| stack-rate | 排序 |
| third-order thinking | 三阶思维 |
| truth function | 真值函数 |
| Typeform | Typeform(产品名,保留原文) |
| unicorn | 独角兽 |
| VP of engineering | 工程 VP |
| where clause | where 子句 |
| Wordplay | Wordplay(产品名,保留原文) |
| Wufoo | Wufoo(产品名,保留原文) |
此文档由 AI 分片翻译(translate_long_document)