在 Atlassian 内部从0到1构建的来之不易的经验 | Tanguy Crusson(Jira 产品发现(Product Discovery)负责人)
Hard-won lessons building 0 to 1 inside Atlassian | Tanguy Crusson (Head of Jira Product Discovery)
Tanguy Crusson: Been in the product management team at Atlassian for roughly 10 years now. I worked on HipChat and Stride, and more recently I started Jira Product Discovery.
Introducing the Guest
Lenny Rachitsky: Why is it so hard to start new products, go zero to one within large companies?
The 0 to 1 Journey at Atlassian
Tanguy Crusson: The company has a tendency to over-invest. Startups have the benefit of starving, and so you need to create scarcity. What we try to do is remind everyone things are going to fail, let’s not drag the rest of the company into it.
Challenges of Internal 0 to 1
Lenny Rachitsky: Sounds like one of the biggest lessons is super silo sort of team.
Big Co Advantages for New Products
Tanguy Crusson: I needed the rest of the company to go away so we could get the autonomy to test the things that we needed, but it’s not going to scale. That is not going to respect all design guidelines.
Why Advantages Fail to Drive Success
Lenny Rachitsky: The biggest challenge I think a lot of companies have is just, it’s been six months, no one wants this, we’re going to kill it. How do you protect that?
Misaligned Success Metrics and Thresholds
Tanguy Crusson: Be very clear about what we’re testing, doing that with data, doing that with personal customer stories, give people a sense of velocity and speed. No one wants to fuck with a high-speed train.
Failure Case Study: HipChat
Lenny Rachitsky: Today, my guest is Tanguy Crusson. This is a really unique and important episode because we get into something you don’t hear much on podcasts like this, the real talk challenges of trying to innovate and build zero to one at a large company like Atlassian. Tanguy has been at Atlassian for over 10 years and has worked on a bunch of internal big bets, some that have worked and some that have not, including products like HipChat, which I was a huge fan of back in the day, also a product called Status Page, and most recently Jira Product Discovery, which is one of the fastest growing products in Atlassian history that Tanguy led from idea to launch. We go through each of these stories, and Tanguy shares what went wrong, what went right, and everything that he’s learned about creating space for innovation within a larger org, including how they structured their internal incubation program called Point A.
There’s a ton of gold in this episode and a bunch of really interesting stories, which is part of the reason that it went this long. It’s the longest episode I’ve done yet. If you’re looking to create change in your organization and foster more innovation, this episode will be worth your time. If you enjoy this podcast, don’t forget to subscribe and follow it in your favorite podcasting app or YouTube, the best way to avoid missing feature episodes and it helps the podcast tremendously. With that, I bring you Tanguy Crusson.
Tanguy, thank you so much for being here, and welcome to the podcast.
The Early Rise of HipChat
Tanguy Crusson: Thank you very much for welcoming me here, Lenny. I’m actually super proud to be on this podcast. I’ve been a huge fan. Whenever I get the chance, I listen to you when I drive somewhere so yeah.
Lenny Rachitsky: What we’re going to be talking about in this episode is we’re going to be talking about building new products and going zero to one within larger companies, and in particular, the pain and the challenges that come along with that, but also the lessons that you’ve learned from doing this many times and seeing it done many times. You’ve seen a lot of this happening at Atlassian, you’ve been there for over 10 years at this point, and Atlassian has, I don’t know, over a dozen different product lines at this point, something like that. I know a lot of people come to you asking for advice on how to build zero to one within a large company. So let me just start with a general question of just could you just share a bit about your history of building zero to one and just seeing zero to one happening within Atlassian?
HipChat’s Costly Rewrite and Scaling
Tanguy Crusson: Yeah, so like you said, I’ve been in the product management team at Atlassian for roughly 10 years now, and I’ve been working mainly on bootstrapping new things. Initially, I joined to start the cloud developer ecosystem so developers can build apps on top of the Atlassian platform and sell them on the Atlassian marketplace. I worked on HipChat and Stride. HipChat was well-known, Stride less so. We were trying to win the enterprise communications market before Slack and Microsoft Teams came about. I did lead a business case to invest more in IT operations, got nowhere with it than we acquired Statuspage, Opsgenie, and something I tried to do with didn’t quite get off the ground. More recently I started Jira Project Discovery, which was part of our internal incubator for two to three years, and came out of the incubator and generally available a year ago.
My track record at Atlassian has been 50/50 at best. Jira Project Discovery is actually my first, what I would call big success here. That one worked, but it was hard and all the ones that I worked on before were really hard to, and that’s the kind of stuff that really bothered me for a super long time. The good thing is working on a product for product managers, I got to talk to a lot of product managers and across all sorts of industries in the past three to four years, and I realize that 50/50 is actually not that bad.
Lenny Rachitsky: Awesome, and I know there’s going to be a lot of real talk in this conversation. I’m excited to share and hear all these stories.
WorkOS also recently acquired Warrant, the fine-grain authorization service. Warrant’s product is based on a groundbreaking authorization system called Zanzibar, which was originally designed for Google to power Google Docs and YouTube. This enables fast authorization checks at enormous scale while maintaining a flexible model that can be adapted to even the most complex use cases. If you’re currently looking to build role-based access control or other enterprise features like single sign-on, SCIM, or user management, you should consider WorkOS. It’s a drop-in replacement for Auth0 and supports up to 1 million monthly active users for free. Check it out at workos.com to learn more. That’s workos.com.
Just broadly, why is it so hard in your experience to start new products go zero to one within large companies? What have you seen are the biggest challenges and hurdles generally?
Impact of Microsoft Teams
Tanguy Crusson: Yeah, so on the opportunity side, so Atlassian, 300,000 customers. We play in a whole bunch of different markets, everything in the collaboration space. We have a lot of markets that we play in, which means that we’ve got a lot of competitors. But basically when we look at the areas we could go in, there’s an endless list of areas that we could go in, play and have a decent chance to win. Much harder to do if you’re a startup. The breadth that we’ve got makes it easier to try find areas where we could expand into. We’re not starving like a smaller company, so we can actually afford to try to play somewhere, to do some bets, to have some of them fail and some of them succeed, so that’s amazing. Now our customers, I said 300,000. The thing that I admire the most about the Atlassian business model is it’s very broad. It’s across small and medium-sized companies. We have startups using our products, and we’ve got also enterprises and very large enterprises using the same products, which means that there’s a lot of areas where we could find a niche and go after it and expand progressively into all these areas so there’s a massive distribution potential that comes with that. When I worked on Jira Product Discovery, I didn’t start with, okay, I’m going to need to start finding product managers and it’s going to be how to find them. No, they were already all using Jira. Atlassian is a company that has a relatively deep organization hierarchy but relatively flat decision making. So it’s more like a, imagine a network of key decision makers across the organization. It doesn’t really matter the job title or whether you are a manager of people or not. The decisions are made by people who drive change.
So there’s a lot of empowerment that comes from that, but also it’s a mix of top-down, bottom-up happiness I’d say. And so it can feel really chaotic at first, but once you know how to navigate it, it’s actually pretty easy to try to go after something that you care about. And of course, we’re a big company, so there’s lots of ways we can get help. Corporate development, research analysts that we can talk to whenever we want to explore something, thousands of customers that I just have to put something in the addressable community group and get hundreds of people applying to talk to me from one day to the next. So that’s amazing, any startup would want that.
Lenny Rachitsky: This sounds like how can anything not work when you launch a new product? You have 300,000 potential customers to launch it to. You have all the resources to build it. It sounds like decision making is efficient relatively, it’s flat. You have all these different customer segments that use all these different version product lines of Atlassian. It’s just like all of the opportunity possible to launch new products and still, many things do not work out, so I think this is a really important point and I think many big companies are in this. We have so much opportunity. Everything we’re going to build is going to, it’s going to grow like crazy because we have everything we need, but still it doesn’t work out. And that’s why I think we’re going to talk about, it’s going to be so important.
Market Exit and Growing Pains
Tanguy Crusson: It’s going to be a little bit of therapy for me. Hopefully some people out there could go, “Okay, it’s not just me having a hard time.” It can happen in companies like Atlassian too.
Don’t Drink Your Own Kool-Aid
Lenny Rachitsky: Amazing.
Why Playbooks Fail in New Markets
Tanguy Crusson: So yeah, let’s talk about the challenges.
Buyer vs. User Influence Gap
Lenny Rachitsky: Let’s do it.
Tanguy Crusson: Yeah, for that. So you want to start a new thing. This thing is going to take time, and you need to be able to have that time for the time it takes until you can prove whether there is a thing or not. The thing inside Atlassian is that the path for success is super high for a new bet. If you come in and you create a product and it’s got 100 customers, it’s going to look cute. Remember, we serve startups and enterprises, we have self-service and sales, we’ve got all these motions that are in place for our bigger products. A 100 million business is a home run. For us, it’s not like that. We’re trying to build businesses that grow really big and keep growing big over time.
Now, evaluating success can look very different between early stage and established products. For a long time at Atlassian, we were treating everything a little bit equally in that the metrics success for the same. For example, things like monthly active users is the way that you, for a long time we looked at you go, is that product being successful on it? And what if you are building an “internal startup,” your monthly active user number should look very low for quite some time up until you know that your product is ready to serve the vast majority of the customers that you want to put it in front of, so that they don’t just look at it and go, “It’s not ready for me.”
They’re going to try it and then they’re going to churn and it’s going to take forever to claim them back, so that makes it pretty challenging to try and start new things unless we’ve got the right metrics and processes and everything internally that can give room and breathing space for the bets to succeed internally. And for many years, we were not there. We started getting there more recently with the start of Point A, which was our internal incubator program, which is one of my latest that was successful. The ones before didn’t have that, and I really struggled from that, and I see many companies struggling with that exact aspect.
The Importance of Validating Hypotheses
Lenny Rachitsky: Amazing. Let’s dive into an actual story. There’s three that I want to talk about. There’s HipChat, which you mentioned. There’s Status Page and then there’s the product you’re working on now, Jira Product Discovery. So with HipChat, funny story, I loved HipChat. I was a huge user of HipChat at my startup back in the day. I can never forget the billboard that you all put out promoting HipChat where there’s this little stick figure meme guy and it just said, “Why use HipChat?” And I thought that was the funniest thing, and the product was so delightful. There’s just all these little emojis in there, and the idea with HipChat for Atlassian was basically to become the Slack killer, that was the vision.
Early Validation in Product Discovery
Tanguy Crusson: You just killed me. We were way before Slack.
Lenny Rachitsky: Okay, so first mover advantage, amazing product.
The Danger of Competitive Myopia
Tanguy Crusson: It works, yeah.
The Statuspage Experience
Lenny Rachitsky: And it was an acquisition for Atlassian.
Tanguy Crusson: It was an acquisition.
The Power of Transparency
Lenny Rachitsky: Awesome. So let’s talk about what went wrong with HipChat. What did you learn?
Discovering and Acquiring Statuspage
Tanguy Crusson: All know Atlassian as company, as the tool that could almost have a, okay. [inaudible 00:13:57].
M&A Pain Points and Lessons
Lenny Rachitsky: Yeah, this is the therapy. The therapy session begins.
Tech Stack and Product Integration
Tanguy Crusson: Yeah, it’s going to start right there. Me and everyone else from the HipChat team I can tell you. Okay, so yes, HipChat was an acquisition, team of 20 people or so. It was Slack before Slack was there. Great traction and lot of, it was a darling with startups. It was a new way to collaborate back then. There were a few of these smaller apps that were trying to do this thing. I remember actually joining Atlassian. Before that, I was working with financial services, banks and stuff like that, and we were big meeting to talk about stuff, or going to someone’s desk to talk about stuff. I joined this company where my colleagues who sit on the same floor as me and on the same table, we talk over the computer via chat. I often felt weird at first looking over my shoulder to the person I’m currently talking to and we’re having an argument, but we’re doing it over text. Anyway, it might seem a bit cute to the people who had been born in the Slack world, but it was a major change back then.
Lenny Rachitsky: Yeah, I remember that. I remember that, I was in the same office with my team and we’re using HipChat to chat and it felt strange. Now it’s completely normal.
The “Why Now” Urgency
Tanguy Crusson: It’s just normal, and Hipchat was one of the first to move there. Slack came out of nowhere. Company actually initially was focused more on gaming and they really took the market by storm. The growth numbers were dizzying when we’re looking at them. And so at some point, Hipchat was left relatively alone for a while inside Atlassian. You’re doing something good, so keep going after it. But with Slack, we now had to try to go bigger, so we started this thing called HipChat Go Big, the team, they recruited-
Lessons Learned from 0 to 1
Lenny Rachitsky: That was the name of the project? HipChat Go Big?
Point A Incubator and Jira Discovery
Tanguy Crusson: Yeah, HipChat Go Big. And then it was HipChat Next Gen, there was a few different. Anyway, yeah.
Lenny Rachitsky: Very clear. HipChat Go Big, I love it.
The Four Stages of Point A
Tanguy Crusson: Go big. But it was really a go big, lots of new developers, lots of new product managers, designers, the full company behind this product kind of thing. We tried to grow it very aggressively for a product that did not change that fast before it had reached good product market fit already, hundreds of thousands of users on a daily basis, and all of a sudden that you get a lot of people who want to make changes to it, to compete against this new threat. The platform is not so ready for so many people to work on it. And so we got to the inevitable, okay, it’s too much tech, we can’t do much about it so we made the decision to rewrite it. There’s lots of literature around there around should you do rewrite? Should you not do rewrite? You ask me now, I tell you never. Trust me.
Lenny Rachitsky: That’s what most of the advice is, never do a rewrite, and people still do it.
Review Gates and Decision Mechanisms
Tanguy Crusson: Never do a rewrite, and there’s good reasons for that. But basically we did that and out of it came up actually a new product called Stride, which was initially “HipChat next gen.” The problem is that the product was great, but by the time we were done, Slack was just miles ahead of us. At that point, Microsoft launched Teams. I don’t know if you remember this moment where Slack put an ad in the, I think it was The New York Times, copying the Apple versus Microsoft thing from ages before going, “Welcome to the game, welcome to the party, we’ll welcome you competitors,” and stuff like that and Slack got pretty much destroyed by Teams. We started coming because it was like Microsoft distribution advantage. Everyone in Office is going to get it. They’re giving it for free as part of Office, it was unbundled I think a few months ago.
Project Funnels and Executive Involvement
Lenny Rachitsky: Which is ironic because in theory, Atlassian also has that same advantage, right? You have all these products, you could bundle it.
Success Patterns and Breakthrough Strategies
Tanguy Crusson: We are going to talk about that actually in a minute.
The Value of Highly Isolated Teams
Lenny Rachitsky: Okay, great.
Innovating Safely Within Mature Products
Tanguy Crusson: Because that was what helped us, how we thought we would win, and it was how I think we lost.
The Safety Funnel and Business Risk
Lenny Rachitsky: Amazing. Let’s get in.
Lighthouse User Program and Metrics
Tanguy Crusson: So anyway, this all happened, and in the end we executed the market. We sold HipChat and Stride to Slack and basically exited the enterprise communications market.
Lenny Rachitsky: Just to double down that, but you sold that to Slack and now Salesforce basically, that’s the word ended. I don’t know if people know that.
Internal Value of the Lighthouse Program
Tanguy Crusson: Yeah, actually, it was noticed by the market in that as soon as we did that, our stock price went up. I think it was $60 back then and it went up to 70. I mean, that really sucked for us, the team working on it. First, no one tells you, but failure, everyone tells you failure is great because you learn so many things, but failure to start with really sucks. None of us here were really happy about this on the team because we had spent years, on my part, it was three years, but the people who were before that on HipChat was longer than that. Obsessing over it, obsessing over every detail, every customer conversation, every solution, should we rewrite, should we this, should we do that?
So many intense conversations, and from one day to the next, it didn’t matter anymore. We had worked on something and that’s it, that’s the end. I’m sure many startups have been through that before. For me it was the first time it felt that personal, and the market the next day went, “Oh, you’re stopping doing what you’re doing? Awesome. Yes, 10 more dollars to your stock price.” So yeah, there’s a personal side to all the stories. For us, there was a bit like the seven stages of grief after we shut down HipChat.
Surviving the Ugly Baby Phase
Lenny Rachitsky: How long was that period of mourning and stages for you and the team?
Balancing Change and Self-Preservation
Tanguy Crusson: It lasted a few months. I was not part of the decision making team for shutting down HipChat. I was one of the product managers on the team leading one of the three pillars. I got brought in, I think it was a month or two, a month before it was announced for the team to go, “Hey, Tanguy, by the way, just so you know, HipChat is no more, and now your mission is to find a new mission for the team after that.”
We basically spend the next two, three months trying to make sure that the squads were created to fully own what they do there, to make sure that they’re the ones talking to the customers, they’re the other ones trying to come up with strategies, trying to come up with solutions and everything in that area. As long as it was people interacting with Atlassian products on other surfaces, everything was fair game, and so there was a lot of ups and downs and everything. I think it took about two, three months before we got back to a new rhythm. And some people, when we talk about it, were still scarred by it basically.
Lightning Q&A Round
Lenny Rachitsky: What are some lessons from that experience?
Top Book Recommendations
Tanguy Crusson: Yeah, so the main one that I personally got from this, and it’s back to the hypothesis that you talked about, which we have all these successful products, we can expand into this one which is, and I quote it, but it’s just myself, don’t eat your own bullshit, which is a mix of two things. We’ve got a company value that says open company, no bullshit. So we need to be able to talk about the things like they are, and we don’t try to make things sound smarter than they are. We don’t try to hide the truth, we go after that truth, and we don’t hide stuff from each other. We share with everyone, we are open by default. So that’s one of the values and we do a lot of dogfooding, so eat your own dog food. We do test our own software a lot.
I’ve noticed that sometimes there are things that we do while we tend to believe stuff because it’s worked for us before, and we have this assumption that it’s going to keep working for us forever. The founders keep telling us what took us here won’t take us there. That’s a thing we keep hearing over and over again. But it’s very easy for teams when they see success of something to think that it’s successful because of X, but X is not validated. That’s where we go back to the topic we’ve got today, which is why is doing this in a successful company harder sometimes. Well, Atlassian was successful with the playbook, and the playbook was we’ve got people in developers or tech or IT. They choose Atlassian apps, they love them. They start to recommend them to people in the business, and we start to see adoption, bottom-up adoption across the company before people decide to standardize on Atlassian.
We made the bet that we can apply this playbook to this market, which is basically we can, from people to use Jira, introduce HipChat, and then people will go into HipChat first in tech teams, and then it will expand into business teams, and it’ll go world to world basically from that. The thing is we didn’t, in my opinion, do enough to validate that assumption early enough. We did a lot of work, a lot of work even on other things, even when faced with signals that this might not work. I do remember talking with a lot of customers who were like, “Well, we’ve got, the IT is on HipChat, but the business prefer Slack.” And then we started to see those businesses choosing Slack, which is the, initially it was like the developers try things and they like it and then everyone starts to adopt. In that case, the Slack managed to create a very strong fund base in roles that were not tech and IT.
It was the moment where the consumerization of apps, that trend was starting to get really high. Slack really rode that and they focused everything in their experience to catch that. They gamified onboarding, they focused a lot more on the look and feel. They try to make it pleasant, modern, functional to use. There’s a lot of stuff that we learned since then from what they did, we had missed that part. The part that I took personally from that is that there’s a lot of assumptions in what made us successful, it doesn’t mean that it’s going to work.
Product Interview Tips
Lenny Rachitsky: Just I understand what you’re saying, which is really interesting that Atlassian was really successful selling basically to the buyer within the org, the IT team because they had everything they needed. They checked all the checkboxes, but it turned out in the Slack case, it was the users that ended up having the most influence over what tool they’d opted.
Current Favorite Products
Tanguy Crusson: I’d actually phrase it more as both were going after the users. Atlassian was going after the users in tech teams, Slack was going after the users in business teams. And in both cases, what happened was a bottom-up adoption. The people on the other side, the business would prefer Slack, the developers, they prefer the HipChat. We did a lot of work in the streams I was working on. We’re integrating with every developer tooling out there to make sure that every tool that they use goes into HipChat and from HipChat back to those tools. And basically, they can do a lot of work by seeing an activity stream in HipChat. Business? Eh, not so excited by this. Emojis, a lot of other things that may at some point I remember we were thinking those things were trivial. No, they were not trivial. It was just a different approach for using the tool by a different set of users that we did not talk enough to.
Lenny Rachitsky: So is the lesson here, don’t underestimate the challenge you’ll have convincing a new segment to buy your thing. You may think they’re close or similar, but they’re probably not.
Personal Life Motto
Tanguy Crusson: Yeah, that’s one of them. The other one is what took you here is not going to take you there. And so go back and try to explain why you are successful today, and then if you think you can use the same thing on the next thing, find ways to validate it, find ways to test it. Don’t just go and build on those assumptions. That’s the main thing I got out of this ordeal basically for the stuff I did after.
My Freediving Experience
Lenny Rachitsky: How would you do that? How would you go about and test it? Is it use research? Is it the PMs talking to potential users? What would you have done there?
Tanguy Crusson: For example, when we started Jira for the Discovery, which is, so maybe I should introduce this for a second. It’s a product for product managers, which is mainly used for prioritization and roadmap. People use Jira to plan and track work when it’s committed. We wanted to create a space before that so people debate priorities with everyone that should be involved in that prioritization process, whether it be the developers, designers, so people in the product team or people outside of it, customer success, salespeople, support, leadership and so on and so forth. When we started, that we thought, okay, the product managers are already in Jira. You know what? We can reach them. So we create the tool, and then we’ll distribute it from Jira. We could have gone down the path of building that and then started to start to distribute it.
Instead, we did things like before we brought a single line of code, put an ad inside a Jira newsletter going, “Hey, we’ve got this thing for product managers coming up.” And then we had a website that before we had any line of code written that said, “Hey, product managers, your job is hard. We want to help, put your name here if you want to join us on the journey,” that kind of thing. And that’s when we saw, I think it was in two weeks, we got more than 3000 signups to that waitlist. We’re like, “Okay, cool. Validation of demand.” We are talking to people who are interested and we can reach them. That’s one examples of the things that I tried later, just to make sure that we validate the hypothesis that we’ve got much earlier on in the game basically, and not once it’s too late.
Closing Remarks and Contact Info
Lenny Rachitsky: Awesome. That’s an awesome, very tactical example.
Tanguy Crusson: Yeah, most of those are, none of what I’m going to talk about today is revolutionary. A lot of it is just trying to apply, asking the right question at the right time, and trying to go by whatever means to answer it really.
Final Podcast Outro
Lenny Rachitsky: Any other lessons from the HipChat experience before we shift to a different product?
Tanguy Crusson: I’ve got two, I’m going to do them quickly. The first one is competitive myopia, don’t fall for it. At some point, the Slack was really gaining round and gaining round and capturing more of the mindshare. And everyone on Twitter was always loving them. Even when they had outrages, they were getting congratulated. I was like, “This is working mad.” But the love was so strong, and the way we tended to revert back is to our functional side of the brain going, okay, we just need this one more feature. We just need this one more feature, we just need this one more feature. And we ended up reacting to whatever.
Tanguy Crusson: … more feature. We just need this one more feature. And we ended up reacting to whatever the competitor was doing, which I think is really, really bad because that’s when we lost basically what made HipChat successful so far, which is to serve some users really well. And instead we ended up fast following based on what the competitor was doing, which is super bad because your competitor, if you think of what they do as an iceberg, the top side, what comes out of the water is what they’ve shipped in terms of features, but it’s based on all this stuff that they’ve built in terms of research and understanding of their customer base and everything else. And so you’re just seeing the manifestation of what they were thinking maybe a year ago, based on what they’re shipping now. But we got there.
And now whenever I work, I tend to try and ignore competition other than watching every three months or so, seeing what came out, if there’s anything we should be worried about, afraid of, stuff like that. But really just try to disconnect all the creative process and the research process from what competitors do, because you can’t compare. The market is huge. There are hundreds of thousands of companies out there. Not everyone has the same needs. We serve a particular set of segments. We would do better learning from them to then expand to the others than watching what competition is doing.
Lenny Rachitsky: And this is advice you give to your teams, just ignore the competition, maybe pay attention at big announcements and things like that?
Tanguy Crusson: We always see in the Slack Channels team sharing, okay, they just did this with AI, they just did that. So it keeps coming up, and so we often have discussions to go back, to okay, let’s watch again the user interviews that we did over the past three weeks. Let’s watch them all together now and remember who we are building this for. So there’s a lot of trying to anchor back on what we know and how, basically building our own journey on it. And I think it’s much richer for everyone to be involved.
Lenny Rachitsky: I love that. So you’re finding that when there’s a big announcement and everyone’s like, “Oh my God, look what Slack’s doing,” or “Look what this company’s doing.” It’s like, “Okay, now let’s spend a little time reminding ourselves what our customers have been asking us to do and let’s watch a couple of user interviews.”
Tanguy Crusson: And even when there is something that competitors do that is right, remember, we’re playing the long run and we don’t always need to be first and shinier. We need to make sure that people have a problem, we solve this problem. They tell us, we solve this problem for them. They are delighted when we solve these problems for them. And so that’s the stuff we should be obsessing about.
Lenny Rachitsky: I love that. And you said you had one more lesson from this experience.
Tanguy Crusson: Yeah, the last one, startups have the benefit of starving. Right? We’re a big company, we can throw a lot of resources at something that we’re excited about. So this notion of we rebuild the HipChat, was coupled with we rebuild the HipChat and we’ll do this on a new platform, which is microservices and everything that we built can be reused across all the other products. So the chat text box that you’ve got, it’s an editor that can be open full screen and it’s a Confluence editor with everything that you can do there. It’s built as a platform component, which is amazing when the platform is there when you start building the product. Where it’s really difficult is when you try to do the two at the same time. So I think that part, if I were to work on HipChat again and say I was leading that thing, that’s the part I would go, okay, we need to win the problem space first, and if the platform is there, let’s use it. If it’s not there, we’ll hack it, test it, iterate on it with customers, and then whatever is good there, let’s platformize it later.
But that’s where you see what makes us super powerful as a bigger company, can also slow us down and make us focus on the wrong assumptions. So in that case, I think we thought we will win. Interestingly, I think we were convinced that we had a great shot at this market. And at the same time we thought that we could tackle the rewrite and we could tackle the platformization. All these things were necessary, but all of them at the same time was probably a bit too much to bite. Now I’m saying that, but today the editor you see in Confluence started with what we did in HipChat back then. So I would say that in terms of code, purely in terms of code, 70% of what we wrote for HipChat is probably still in the Atlassian platform today. But for a new bet, local Optima, that was really bad.
Lenny Rachitsky: Yeah. Just thinking about the fact that Atlassian could have had this $30 billion business if this worked out, and I could see why people would be frustrated that it didn’t work out. So thank you for sharing that story. Let’s talk about Statuspage, another journey that you were a part of, that also didn’t quite work out the way folks had hoped. Our therapy session continues. Talk about what that product was and what happened there.
Tanguy Crusson: Yeah, so this one is actually a success story, but became a success story after I was gone [inaudible 00:34:08]
Lenny Rachitsky: Okay.
Tanguy Crusson: That’s more like a story of big companies can play the long run and for you individually inside that process, it might look like a loss and you might feel like you’re going nowhere, yet the company stays on the opportunity for long enough to make it happen. So what we’re going to talk about here are my own challenges working through it for something that ended up being very successful in the end, but I felt as a failure personally back then. So Statuspage at some point… So Jira is used by developers, right? And back then, I think it was back in 2016 or even before that, 2015, everyone was moving to the cloud, everyone was adopting DevOps, you build it, you run it. So basically developers would implement the software and then they would put it to production. And after they put it to production, they don’t throw it over the wall to operations people. It’s the same developers who operate the software in production. They go on call, yada yada.
So back then I did market research to see whether Atlassian should play there, whether we had a crack at going after all the jobs around IT operations. So Jira was basically Jira software to build software. Could we have Jira for operations? So for operating this software. And so I did market research, found a few companies that were doing super interesting things there, like [inaudible 00:35:37], Opsgenie, New Relic. BigPanda was a small startup doing lots of cool stuff there, and Statuspage. Now I discovered Statuspage and found their offering super interesting, in that Atlassian is not an operations… We don’t really build super deep operations tooling, what we focus on a lot is the collaborative aspects around everything. And when I was watching teams in incidents, I realized that there’s a lot of chicken without a head syndrome, headless chicken, people running around like bunch of headless chicken.
Shit hits the fan, and then what you see is teams just scrambling and everything gets mixed up all together in trying to fix the problem straight away, but at the same time questioning what happened, arguing over why we got there. Your boss is pinging you to go, “Hey, what’s going on? I’ve been hearing that the app is down. We’re losing money.” Customers are emailing you and asking you what’s going… Your support channels get blown up. Sales people are worried because their customers are calling them. So basically, it ends up being a super stressful experience for everyone.
And what Statuspage was offering is something seemingly super simple, which is, well, what you should have is a status page for your services and you tell your customers about it and you can subscribe to your status page. Over there, you’ve got your services, and for those services you can publish an incident whenever there is one, and your customers will be notified, which means that they’re not going to get in touch with support because they know you’re on it. It’s going to build trust with them because basically you are open in your communication with them. They’re more likely to empathize with your position and be supportive as opposed to, “Oh, that stuff is done again.” So there was just so many benefits of that.
Lenny Rachitsky: I’m going to share a quick story while you’re on this topic.
Tanguy Crusson: Yeah, of course.
Lenny Rachitsky: Funny enough, a decade ago, I used to work at a website performance monitoring company and I started a blog called Transparent Uptime. And my whole blog was about the power of being transparent about being down, telling people the status of things are broken right now, here’s when things are going to return. It was a whole thing I was really obsessed with, and I was deep into this space. So when I saw Statuspage back in the day and Atlassian working, I was like, “I love this.” And I actually chatted with the founders a bit, because they were fans of that work I was doing back in the day. Completely unrelated to anything else I’ve done in my life. But I was really passionate about this very strange topic and I love that companies are embracing it.
Tanguy Crusson: I really loved diving deep into it for a few years, it’s an amazing topic. And so anyway, there’s so many fascinating things about this particular domain, but back then… So find Statuspage and we invited a few of those companies to work with us for a week to go, hey, so we’re at Atlassian, we’re basically the collaboration hub for everything that happens for developing teams. What would an experience look like that puts everything together and where [inaudible 00:38:48] could actually help you get the right information to the right team so they can act on it?
And so we did a week of hackathon in San Francisco all together. So people from New Relic, Statuspage and a few other companies. And out of that came really interesting concepts and there was really something there and I was like, “Okay, I should start to work on a business case for Atlassian for basically IT tool/operations.” Now, big companies like Atlassian, we’ve got money. So in every strategy that we’ve got, we’ve got cash in the bank, we always look at should we build, should we buy, should we partner? Acquisitions can be really powerful to accelerate you. We’ve got quite a few success stories there. Can’t remember how many we did, but Trello is one of the good examples of that, for example. So we decided to buy Statuspage and I was running the integration of the Statuspage business inside the Atlassian business.
Lenny Rachitsky: Thanks for sharing all that context. What are some of the things that you learned from going through this experience? It sounds like basically it was really painful when you were a part of it and then it ended up being really successful. What are some lessons from the pain?
Tanguy Crusson: My learnings from it were on the acquisition side. So big company, we’ve got cash, we can buy a company, it will make us go faster, is not always the case. In my case, there were quite a few things that were not as easy as I would’ve thought. The first one is the culture shock of a startup that joins your company. So imagine you’re a startup, you’ve got, I don’t know, 20, 30 staff or something and things are going well, and you are in full control of your destiny. And then a company buys you to accelerate you, but then you stop owning all the decisions. So the CEO, maybe you become a product person. The person who was running GTM but was also doing a bunch of product stuff, and was also doing a bunch of maybe engineering stuff, all of a sudden is just working in marketing.
There are decisions that are made above your head. For example, portfolio fit, which should be part of your product, which is things that should be used from the platform that we’ve got. So there’s a lot of decisions that we’re able to make on the day-to-day basis, that start to escape you. Big companies look much further out in the future. So Atlassian would look at the long game. And so I remember with Statuspage when they first joined, we asked about the roadmaps and they were like, well, for the next three months we’re working on that and for the following three months we’re thinking about potentially X, Y or Z. And so when we’re like, “Okay, so what’s the three-year plan?” They’re like, “What do you mean the three-year plan? I don’t know, we’ll survive, I guess.” Which is what startups do. And they were very penciled ideas about the future, but not to the extent of what a big company would expect. And so that’s a really big culture shock.
And what you end up with as well, and that’s for companies who are looking to get acquired out there or buying other companies, before, you had a startup and everyone was one team, depending on how the buying company is organized, you might land with silos from within your team. So the way Atlassian is organized is like many Atlassian companies, we’ve got a product organization, we’ve got an engineering organization, we’ve got a marketing organization, design organization, and they each roll up to a different leader which may then roll up to the same person, but still by and large different organizations that are then assembled into squads and those squads operate together. But there are rituals that are part of the squad and there are rituals that ladder up to where you are in terms of craft. So I do remember how daunting it was for the Statuspage team when they joined, to understand how to navigate that.
If I want to hire a new designer, I’m not talking to the Statuspage CEO anymore, I need to talk to the head of design for this area, which has pretty much nothing to do with the Statuspage business up until the acquisition. So that is the things that people told me when I started working on the integration, which is, hey, remember, you don’t know yet, but integrations are mostly about people. They’re not about technology as much or product vision. All of that stuff is the easy stuff. The hard part is the people. And I didn’t quite understand at first and then I really got it by the end of it.
Because what we tell those companies is we can accelerate by buying, but in reality they are going to be faced with more internal processes around how they manage staff, performance reviews, this concept of engineering allocation, revenue forecasts, OKRs, long-term roadmaps, all that stuff comes in very new to them. So one part of the company top-down says, we bought you because you are successful, keep doing what you’re doing. While many other teams without meaning to, basically end up constantly interacting so that the right processes are followed, so that the right things are done.
Lenny Rachitsky: That’s a really interesting point, that one group is keep doing what you’re doing. We’re going to leave you alone, you’re the experts. And then other people that are on the ground actually building it are like, “Hey, build with this component. Hey, we need this process, we need this document.”
Tanguy Crusson: And it’s the things that you used to be able to focus 90% of your time working on your product, and all of a sudden all of that stuff may seem parasitical but comes in and interrupts you all the time. And often the new joiners, they don’t know that they basically are not only being acquired, they’ve been hired by the company. So they basically are joining a different company with different sets of rituals, with a different culture. All of that is very different, basically. It’s not going to be the same for every acquisition. Like I said, we had some acquisitions that were great. Statuspage ended up being a huge success inside Atlassian, but when I was there, I basically worked through the difficult parts of it, where I was like, “It’s not as simple as people may think.” So one warning, if you’re planning to do acquisitions, make sure you factor all of that in and think about the integration plans to compensate for these aspects. So you don’t expect the business to join and keep running at exactly the same pace, because going to be a big slowdown before it accelerates again.
Lenny Rachitsky: So maybe as a last question here, is just say you were to do an acquisition again and I don’t know if you’ve gone through more, what’s one thing you would change? What’s one thing you’d recommend of let’s make sure to do this thing very differently?
Tanguy Crusson: One aspect that I’m going to talk about to then go there, but one aspect we didn’t talk about which is we bought a product, how does that product fit with the rest? And so we had different types of acquisitions that we did. One is we buy the company and we keep the product running and then we try to integrate it with our tech stack. And then the other acquisition is we buy the company and then we kind of rebuild on our platform, or we buy with this huge synergy with our platform. And so we call that the frankenstack, otherwise. Which is, you get one tech stack here, one tech stack there and they can’t quite talk to each other, identity is different, integrating is different. And so it looks like a patchwork of products and that’s not what our customers want from us.
So the next time I try personally, I would treat it like hiring over treating it like buying business only. So there would be a huge component which is to both educate and tease out what it means to actually hire the team inside the company. At the same time I’m saying that because I don’t want to do a big one. The next one I want to do is relatively small. Find a company that has amazing product that we can bring in as a tech-in, shut down the product, rebuild it on our platform and so it’s basically the equivalent of an acqui-hire, because basically what we’ll be buying is the acceleration of our roadmap.
We could try and form a team to do what they did, but they’ve been successful so they know what they’re doing. We could probably get there one year faster. What does that mean for our revenue at the scale of Atlassian, if we can enter a market one year earlier? It’s probably going to pay the acquisition back on its own. So my learnings from it were basically that it needs to be treated like hiring and what I would like to do next time is more doing it like that, basically.
Lenny Rachitsky: What about in the case of HipChat, where it feels like it was a mistake to rebuild a thing. Is there for this kind of startup, don’t rebuild for this kind of startup? Start again. Do you have a thought of how you separate those two?
Tanguy Crusson: There’s actually a big difference. There’s a big difference. Which is in one case what you try to do is… So when you rebuild, you’ve got a successful business, you’ve got hundreds of thousands of customers, for example, and you’re trying to rebuild the same thing or a different thing, but for the same customers that have got expectations about your current product. In the other one is you buy a company that’s got some traction, not too much, so you can shut down the business and then rebuild on your platform to reach your customer base. So it’s not the same as trying to rebuild the plane mid flight, it’s delaying takeoff. It’s like, “Yeah, we did a trial run, the plane landed. Okay, cool. Switch to another plane, takeoff again.”
Lenny Rachitsky: Got it. I love that. Any other key insights and lessons from this experience before we get to your product discovery?
Tanguy Crusson: One last one and then I’m going to stop with the therapy session, thank you very much for offering.
Lenny Rachitsky: No, no, we got to keep it going. I got to rack up the bill on this therapy session.
Tanguy Crusson: Which is, remember I mentioned I was working on a business case for… Basically going bigger in IT operations. And so there was as Statuspage as part of it, but there was a lot of stuff that I wanted to be able to do inside Jira and a whole bunch of other products to basically go big in that area, IT operations, before Jira service management, which is a very successful product we’ve got around that. Before it was really entering that part of the market. Everyone was excited. It was before we had an incubator internally. And so I was trying to pitch it like let’s build a new product that’s centered around that thing. It’s on Jira and it integrates with these tools that we discussed and we can put in Statuspage there and that’s actually how we could accelerate them as well, and so on and so forth. Everyone was excited, thought it made sense, lots of encouragement. I pitched it to every level of the organization from business leaders, to the CTO that we had at the time, to the CEOs. No one said no, no one said yes. For months.
For months, I was in this limbo of, I think everyone’s excited about this, everyone wants this to happen, but it’s not happening. And so I remember talking with my boss at the time going, what’s going on? When do I stop pushing? Because at some point I’m sure I’m going to start pissing people off. And he was like, “Well, basically when you lose the passion for it. Keep going up until you feel like it’s not worth pushing anymore.” And I remember looking at this advice and going, wow, that was not helpful at all, based on the situation that I was in. But basically at some point I basically gave up. What happens though, and I understood that after because the company ended up going there just a year later, was that I misread the appetite and sense of urgency around that topic and the fact that Atlassian being Atlassian, we invest in so many markets, we have many opportunities like this that sit on a shelf.
Someone did the analysis, someone created a business case, that thing makes sense. There may not be a trigger for the why now. So we need a very strong trigger for why now to go after it. And I did not do a good enough job at articulating this. Why now? Why do we have to do this now versus in a year’s time, in two years time? And the next team that came in afterwards did a much better job at that. But for me, that was a great learning, which is great work can get parked and [inaudible 00:51:20]. And yeah, that’s just what happens, which is because we have so many opportunities and there’s many companies out there that are probably faced with that, doesn’t mean that we have to go after all of them. But it does make sense to explore them, and then decide when to pull the trigger, and that rationale for the sense of urgency needs to be there. It needs to be beautiful to the business days.
Lenny Rachitsky: This reminds me of my chat with Mihika, who does similar work at Figma where she works a lot of zero to one stuff. And she described it as your job is to keep the flame alive and help it spread throughout the entire business if you’re trying to get everyone on board with a new idea. And I like this very tactical piece of advice you’re sharing here of how to do that, is make it clear why this is… I think of it as why is it perishable? Why is this opportunity perishable? Why is it going to disappear if we don’t act on it now? And you could think of it as why do we have to do this now? It’s not just like “This is a huge opportunity.” But we also need to do it right now to give people motivation.
Tanguy Crusson: Yeah, that was a huge learning for me. I wasted months on it, but like with every failure, learnings came out of it in the painful way. So product managers are often biased towards action, or at least that’s my case. I want to go and do stuff and build stuff, and try stuff with customers. And so being idle, waiting for a confirmation is not a great spot to be, but sometimes you need to recognize when you’re there and it’s good to step back.
Lenny Rachitsky:
If you want a platform that empowers your team to strategize, plan and track goals together, you can get started with Coda today for free. And if you want to see for yourself why product teams at high-growth companies like Pinterest, Figma and Qualtrics run on Coda, take advantage of the special limited-time offer just for startups. Head over to coda.io/lenny to sign up and get 1,000 in credit. Coda.io/lenny.
Let’s try to summarize some of the biggest lessons so far that you’ve shared before we get to product discovery. So a few things I noted here, just how to be successful, building zero to one at a large company. One is be very clear, are the users you’re going to be building this new product to actually the same users you’re already selling to? And it may feel like they are close enough, but in the case of HipChat, you learned maybe not, and it’s a lot harder than you expected.
Two is be careful when you rewrite. In some cases, shut it down, rewrite it immediately, accelerate this new idea internally. In other cases and you shared, and so you could rewind if you want to get the actual details of when it makes sense to go one or the other. Sometimes doesn’t make sense to rewrite, just keep what you’re doing and focus on the user problems and don’t slow down. Another tip I wrote down is ignore the competition. Don’t be obsessed with what they’re doing. Focus on what your users are asking you for. And then this idea of paying attention to the why now, that’s a really good one. Just like when you’re trying to make a case, make it clear why it has to happen now. Is there anything else that comes to mind as I try to summarize some of the advice so far?
Tanguy Crusson: No, that seems like a good one. The main one I’m getting also out of this, is if you try to start new things, it’s going to be coming from your drive and your passion, and that’s what pushes stuff forward. So don’t give up. Because I try really hard before getting to one that worked. So I guess that’s probably a testament for it’s possible.
Lenny Rachitsky: And a testament to your grit and desire to make something work. So on that note, let’s talk about your product discovery. I know this is a success at this point. I know there was also a lot of pain that went into this and things that didn’t work. So I’d love to hear both sides of it, just like what was hard about getting this off the ground and then also just what worked, what allowed you to make this work? So start wherever you want to start.
Tanguy Crusson: Yeah, cool. So let’s start with the good stuff before we go back into therapy. Some of the good stuff here is Atlassian at that point had recognized we are innovating in our big successful products all by doing acquisitions, we have to correct that and start building new products ourselves as well. And so there was a huge push from the founders to go, “Hey, we need to restart that.” Out of it came Point A, which was an internal incubator program that was meant to fix that. And the way they framed it was that innovation is like a muscle. Unless you exercise it, it becomes weak, and what we have to do now is to work on it again. And so out of that thing, Point A and Jira Product Discovery was one of the… I think it was one of the 100 pitches that went through that innovation program. There were 100 pitches and out of it came out three products that basically went through all the different stages of it. And so it started with Jira Product Discovery was actually made possible because of that and because we’re inside Atlassian. So we started with a lot of, I could take time to focus on this stuff with nothing else to do it. It was a full-time job, because of this incubator. I was able to form a team easier because there was budget allocated. I was able to form a team that was not worried about losing their job because the program was made so that technically speaking, you would borrow people from other departments. If that thing doesn’t work out, they go back to your department. So there’s no fear of losing your job, basically.
We were able to tap into all the research that was done by our research and insights team internally. We were able to have the corporate development team working with us. I met with probably 20 companies that played in things around product management before forming a view that, hey, maybe we should play there. And all those teams are willing to talk to us because Atlassian is a big player in this market. And so there’s always the opportunity of integrating, being bought, partnering that can…
The opportunity of integrating, being bought, partnering, that kind of stuff. I was able to meet with analysts to say, “Hey, so what do you think about the product management market?”
Lenny Rachitsky: This was all part of the Point A structure?
Tanguy Crusson: So the Point A structure, not all of it was formalized in Point A, remember Point A back then was created alongside the first bets that went through it, and ours was one bet as part of that. So we forged the path for all the ones that came afterwards. But basically, it gave us the crags to go in and ask for help from everyone, and everyone knew it was important. Because everyone knew the company priorities and the new products were top company priority. And so Atlassian, playing the long game, had decided that it was okay to invest in these bets and to reassess them into every three to six months to understand whether we should put more chips in it.
And so the psychological safety of everyone’s job was safe, coupled with access to all the resources from the company. That part was just invaluable to the success of what JPD is today.
So to give you a bit of context, Jira Product Discovery started four years ago with some research. I was alone back then, and then we were three. The first line of code was written three years before we launched officially, as generally available. Which means that for three years we were dogfooding alpha or beta and able to do that with the full support of the company. So when I mention the long game, I mean it. It’s something that’s very hard to get in most companies out there.
And when people ask me about how to stop incubators, it’s like think about it with when you’re going to get your dollars back and forget about it for a while. What you need to see is how the teams are answering the right questions that they ask along the way, and seeing whether you are still excited about the bets as they do that.
Anyway, so we launched a year ago. Fast-forward to today, we’ve got 8,000 customers, amazing CSAT, great traction. It’s one of the fastest growing products in Atlassian history, which is great. So what was hard though, the first one was reminding everyone that failure is the most likely outcome. And I will die on that hill to explain to people when they want to stop things internally, frame it like that, remind everyone. There’s a [inaudible 01:00:28] 70% chance, [inaudible 01:00:30] completely pulled out of thin air that whatever you’re working on is not going to exist in six months.
We’re trying to launch a new product, enter a new market. Our goal is to get to $100 million businesses. So there’s not that many of them out there. We have tried and we have failed. And a few of them are Atlassian. And so remember that, and it’s super important to remember that because otherwise the company has a tendency to over invest, not the company top-down, parts of the company have the tendency to come and try to help.
So for example, “Here, we want to build this.” “Oh, if you want we could change this service to be able to XYZ for you.” “No, we are a bet, which is seven people. Let’s not drag the rest of the company into it. The appetite that the company has right now is these seven people. We’ll see what we can do with these seven people,” was what I was telling everyone.
The reason I was saying that is that, otherwise, yes, you get the help, but the help always comes with condition and the condition is usually things slow down. So what we try to do is remind everyone, things are going to fail, so that we could basically buy the opportunity to hack shit together that is not going to scale, that is not going to respect our design guidelines, that is not going to fit with the JIRA target architecture. But we’re going to test this with customers and see if the concepts make sense, if the prototypes make sense, whether they get value out of those before we tell you, “Hey, you know what? That thing is a thing and by the way, now we’re a proper business. We should build that thing into the platform.”
Lenny Rachitsky: This is really interesting because it’s counterintuitive to think that you should position your new bets as, this is most likely going to fail. This is just the thing we’re trying, don’t worry, don’t commit too much this yet, don’t worry about giving us all these resources. Why do you think that’s so important? Because I don’t think most companies position new incubations that way. Why do you think that’s so effective and more effective?
Tanguy Crusson: Yeah. So the thing that we’re trying to do when starting new products is to basically emulate a startup in an environment that is not hungry, is not starving. And so you need to create scarcity. What I wanted with my team is to make sure that they feel the urgency. That thing needs to move. I also needed the rest of the company to go away, so we could get the autonomy to test the things that we needed to know whether this thing is even going to work or not.
We could not go into a planning session for the next six months to negotiate something with a platform service so we can build a feature to then test with users. I said, “Now, we’re rebuilding this component and we’re testing this with customers next week. It’s not perfect. It’s not perfect.” And so that helped us a lot, but otherwise there’s always this tendency of the process that works for everything else is going to work for this and we do need to keep reminding them, “Hey, we might not exist in six months. Do you really care that much about this process right now? The product might not exist anymore.” The process goes away usually.
Lenny Rachitsky: So it’s basically a trick to keep everyone else within the org away and not worry about what you’re building? Because you just tell them, “Don’t worry, this is not going to work out. We’re just going to try this thing just to see.” So it’s not for your team to feel like, “This is probably not going to work out.” I imagine the team is like, “Oh, we got to make this work. It’s such a good idea.” It’s more as a trick to keep the org from swallowing you up and pushing you around?
Tanguy Crusson: So there’s a part of that, but there’s also really a need from… We need to respect Atlassian’s dollars here, and if we don’t know whether this thing is going to work… I do not want to drag a team of 50 people into this. I want to know that this thing is worth the investment of a team of 50 people. So it’s a bit of both, actually. Now, it’s of course, easier said than done. And so that’s where, again, Point A helped. We had four stages called Wonder, Explore, Make and Impact, where in the first stage it was all about proving that there was a problem area we could go into, there was a market.
We could answer very clearly, articulate why Atlassian should move there. We could articulate why now that stuff that [inaudible 01:04:35] struggled with before, and have enough data to validate all those claims. Explore was about exploring solutions, which doesn’t mean build it and throw it out there and see what sticks. It’s about if you get a bunch of customers raising a problem, can you get them to play back to the solution, [inaudible 01:04:53] address that problem?
And so in the case of Jira Product Discovery, because we were not building, we didn’t need any new technology, it was mainly new UX and new workflows, we basically validated a lot of that with Figma in dozens of Zooms. But it’s basically coming back saying, “Here’s how those companies are framing it. Here’s their problems, here’s how this thing would be solving it for them.”
So that’s part of Explore, which is validating that it’s worth investing in in terms of solution. We don’t only have the right problem, we’ve got the right solutions. Then Make is about making it happen in stages, starting with an alpha, then a beta, and then going out there. And Impact is that stuff is actually ready to go [inaudible 01:05:37]. And now let’s see the impact it has on Atlassian’s business and keep monitoring it from there.
And it turns into a real business from that point onwards. But everyone at Atlassian knew these four stages, Wonder, Explore, Make, Impact. Whenever we were talking with teams, we were telling them we’re currently in Explore. And we started doing that with the full bet itself. Then we started doing that to talk about the different features that we were working on or problem areas we were going after. Which means that now every time we go into conversation with other teams and we’ll mention, “We’re in Wonder,” or, “We’re in Explore,” they know what to expect. When we go to someone in the leadership team and we say, “We want to go from Explore to Make,” they know they’re going to [inaudible 01:06:22] because they need developers now.
So all that vocabulary and clear expectations set for every stage of the process really helped us to facilitate all the conversations that we had with everyone in the organization. And really protected us again from all those, basically, teams that felt like they had to chime in. Now, we’re in Explore, we don’t have anything ready that’s been validated. Let’s not have opinions about how the architecture is done before we have validated that customers want something.
Lenny Rachitsky: This is super cool. I feel like we could do a whole podcast just on the structure of Point A and how you all do this. But just one question, what does the gate look like when you move from Explore to Make, Make to Impact? Is there a group of people that sit in a room and decide thumbs up, thumbs down? How does that decision work?
Tanguy Crusson: So we basically write a six pager that looks at all the different aspects of all the questions that we’re going to answer. And then we are in a meeting with the Point A stakeholders and the founders of Atlassian. And everyone reads that page for about 15 minutes and then question, answers, comments, and [inaudible 01:07:31] of that, by the end of this meeting, we know whether we are clear to go to the next stage.
We got booted back one time when we were like, “Hey, we’re ready to go anywhere from alpha to beta as part of Make.” And they’re like, “No, you’re not.” And I was like, “No, we’re not.” And so we stayed and we basically got more time than what was initially allocated to us. But basically, the founders and the leadership of Point A as well as heads from the different lines of businesses at Atlassian were participating in those sessions, which-
Lenny Rachitsky: That is super cool.
Tanguy Crusson: Basically, visibility all the way up.
Lenny Rachitsky: And they might also just decide, “Let’s kill this thing, it’s not working,” at one of these meetings.
Tanguy Crusson: So it might be, “Let’s kill this thing,” which happened to a number of bets that we did, or it might be, “Let’s roll this thing into something else.” So for example, our new whiteboards product. I say product, whiteboards feature part of Confluence initially came out of Point A. And was eventually rolled into Confluence instead because [inaudible 01:08:32] made more sense there.
Lenny Rachitsky: That’s so interesting. So you said there’s 100 projects that went through Point A. So there’s this funnel. Is there a meeting for all 100 of these that the founders go to for all these incubations or did they come join later down the funnel?
Tanguy Crusson: Not the full 100 of them.
Lenny Rachitsky: Okay, good.
Tanguy Crusson: And I’m saying 100, it’s over a few different quarters of teams coming in and [inaudible 01:08:53], but not for the initial stage of entering Point A.
Lenny Rachitsky: I see.
Tanguy Crusson: It’s usually when they’ve been accepted.
Lenny Rachitsky: I interrupted you and took us on a tangent, you were sharing essentially the things that went well and how this all came together.
Tanguy Crusson: So failure as the most likely outcome. It’s the one thing I would stand behind in everything because I’ve seen before what happens when we get too complacent with it’s going to work. Lessons from the previous things I was talking about. So this one really key. Second one, this one is much harder, which is if you’re starting something like this, your teams will need to break a lot of rules that are established. But they need to be able to do that without breaking the trust of everyone in the organization.
The rules were created to support the business at the stage where it was successful, and it just so happens that they might not work for new bets. So the trick here for me has been to… The way I pictured it is, I’ve got a bunch of chips. Since I joined my class, I’ve been accumulating chips and those chips are-
Lenny Rachitsky: Like social capital.
Tanguy Crusson: Yeah. It’s like trust I’ve built with the founders, with the different business leaders, with succeeding on stuff or failing and explaining why and trying to do better next time. And all that stuff gave me, like you said capital. I’ve got these chips and I decided on this bet, “You know what, I’m going to go all in.” So I always said, “If it works, it works. If it doesn’t work, I’m probably out of here, but I’m going to go all in.”
So if I see something that’s not going to work, I’m going to say it. We’re not going to do it. And so I knew I was going to put myself in tough conversations because people are here to protect things that need to be protected. They just don’t make sense for the stuff I was working on. So one example, breaking rules without breaking trust where it was tricky.
We have a lot of rules in a company like Atlassian, and this is how it works in engineering. I mean, engineers are the biggest part of our workforce. Basically, shit gets done because engineers work on it. And what I needed was to be able to hire the right team, only principal levels, people who have a lot of internal credits so that they can commit to any team’s repo, no questions asked. People who are not looking for the next promotion, but they want to make a splash. And when that’s not possible, I wanted to be able to hire contractors to fill in the gaps and stuff like that. And I was like, “A lot of the rules that I need to be able to break is basically in engineering.” So I decided not to have an engineering leader in the team, and to do it myself.
So I was the product leader and the engineering leader. I was technical enough to be able to have these conversations, but mainly I was just working with amazing engineers who just could self-drive themselves. So they were able to make changes in areas that were not owned by them. They were able to do changes that do not respect any of our standards. They were able to hack their way around rebuilding services and whatnot. We are now not in the position where we need to do stuff like that, but at that time I needed to take a position like that so we could actually go and move fast with basically the equivalent of being a startup inside Atlassian.
That’s not comfortable to do stuff like that. And I would not recommend that in many environments. Atlassian was very forgiving throughout the whole thing. It doesn’t mean that I didn’t [inaudible 01:12:27] on that.
That was not simple. For example, one of the rules is, at that time we didn’t want to have a footprint in Europe back then for more engineering teams. It’s different today, but back then it was like that. I was like, “Shit. I’m based in France, I’m trying to start this stuff. I don’t want to move to the US or move back to Australia just for this right now.” So I hired contractors, lots of contractors when we started, or not lots because we’re not a big team, but contractors to build this stuff.
And again, contractors, they don’t fall into the same rules as the rest of engineering staff. So basically, all the rules around you need to contribute to… For example, the engineering team could say, “[inaudible 01:13:10] to the next, leadership could say, “You need to invest 15% of your time in reliability.”
And for us it’s like, “We’re not there yet. We don’t even have a prototype out.” Yeah. But that’s the company rule, all right. Again, no engineering manager and contractors, boom, none of these rules applied. So it was people who were looking at it from the outside who were like, “Wow, dude, are you sure you’re okay with all this?” And it felt super uncomfortable, but we have the support of leadership from it. It’s just a tough transition for Atlassian from the, “We only invest in the big ones for acquisitions,” to “Let’s invest in [inaudible 01:13:46].”
Lenny Rachitsky: This is a crazy story you’re telling me. So you’re leading this team, you hired a team of contractors to build this product. You’re in a whole different country from the rest of Atlassian, basically. And the whole idea here was just to do stuff that wouldn’t be necessarily allowed at Atlassian. They wouldn’t let you work this way. And you found that to make the thing you needed to make, the thing that you were betting your career on, it sounded like, you’re just going to be this pirate working on this thing in France and it worked out?
Tanguy Crusson: So the Point A emoji and Atlassian is a [inaudible 01:14:24]. I was not the only one. There were quite a few of us working on new bets, basically operating like that. Each questioning different rules. The end goal was not to question the rules, the end goal was to get to the stuff that we needed to do, which is we just need to clean this space, to work with users to test prototypes up until it works and progressively get to, as I said, a product that’s going to be enough to launch. So we never intended to break the rules, which is the things that we were going to choose the ones that were going to work to support us on this mission and say no to the others.
You mentioned Europe. At the time, a lot of other Point A founders were struggling with the fact that they were operating from their mothership in Sydney because they’re still with everyone else. You were talking about doing all this stuff, but remember that we’ve got this OKR thing, we’ve got this vision for GR that we’re building, you need to participate in all this. And so I was in Europe going, “Fine, schedule it at your time.”
And I think a lot of teams were like, “We don’t care enough about this test. They’re telling us it’s not going to exist in six months. I don’t care enough about this to stay up late or wake up super early every day to disagree with them.” And so there was a lot of the early success that we had in being able to move super fast that came from being so far, that people just did not engage to stop us.
And that’s why we were initially the group with the fastest speed. Then it was possible to institutionalize those things into Point A to then make it possible to do it from within Atlassian. But at first, we just needed to blaze our way through. So that’s what we did.
Lenny Rachitsky: So it sounds like one of the biggest lessons and things that work well for something like this is a super silo sort of team, as much as you can disconnect from the core business and let them just do what you think you need to do?
Tanguy Crusson: You’ve interviewed Megan on the show before.
Lenny Rachitsky: Yeah.
Tanguy Crusson: And I listened again to the way she pitched Point A, and yes, that’s exactly what came out of what she was saying. Which is, we gave the space for those teams to be able to do this with a lot of autonomy. And that was the outcome of all that work, basically, which is the program was then set up so new teams could do it and fitting within the mothership, basically.
Lenny Rachitsky: And I think people hear that, like yeah, okay, of course. [inaudible 01:16:59] incubations, silo, separate, they do their own thing. People hear and they try to do that, but I imagine it’s not actually what they need to do. And what I’m hearing is you went to a pretty far extreme of making that happen, not basically breaking the rules to allow for a silo to actually exist. And I love that. What else was key to the success of making this work out?
Tanguy Crusson: So the one that I’m most passionate about is how we work with customers is very different in a super early stage bet versus to what you do in a very established product like Jira. So the first part is, how can we innovate in a way that doesn’t fuck up existing customers? Jira, 120,000 customers, Atlassian, 300,000 customers. We can’t just go in there, start experimenting, breaking a whole bunch of shit that millions of people experience and go, “What are you doing, Atlassian?” So what we needed is to create this area where we could experiment that’s away from Jira while being inside Jira.
That part was a bit tough. It’s possible to do. In fact, I came across an article from Ben Weiss. No, sorry, Noah Weiss, it’s from many years ago now. Where he was basically talking about innovating in a successful material product and talked about three stages of incubate, iterate, integrate. And as part of that he gave the examples of Instagram, Twitter, and Foursquare. Where basically he was explaining, at the beginning, for example, Instagram was a feed that was chronological. And then the team started to experiment with a popular tab that was on the side that was not integrated into the core of the product that everyone uses.
They iterated on that for a while. That became the explore tab and then the explore tab became the main feed. The feed is not chronological anymore, it’s based on your interests. And that part really resonated with me where I was like, “You know what? We’re going to try to apply that to Jira, which is we’re going to build it in Jira, but we’re going to extract ourselves from a lot of the core components of the core base to rebuild the UX that would work for the audience we’re going after.”
Our audience is product managers. They have no patience for spinners, things need to be visual. They need to be able to quickly move things around to visualize the potential options they’ve got on their roadmaps and stuff like that. It needs to be snappy, it needs to get out of the way, but also something that they can feel proud to present to their stakeholders to have discussions around that.
It can be bogged down into the details. It can go to very, very deep [inaudible 01:19:43]. It needs to be an area, a space where you feel like you can breathe and have creative conversations. And Jira’s UI was not exactly known for that. And in fact, most of the product managers I talked to were like, “I don’t want to do this in Jira. It’s too strict, there’s too many workflows, it’s owned by IT,” yada, yada. And so we said, “That’s fine. We’re going to experiment with an experience that’s going to be detached from Jira still on the same platform.”
And that’s what we did, basically. And so this concept of incubate by working on something on the site, iterate until you’ve got it right and then integrate it back into the main product is something I would definitely do again. We’re currently in the middle of the integration phase for Jira product discovery.
Lenny Rachitsky: So the lesson there is don’t force yourself to be a part of the broader product, initially. First, figure out what it could be if it’s its own thing, and then eventually you can integrate?
Tanguy Crusson: It depends who you’re doing it for and the programs that they’ve got. Don’t feel limited by your platform to answer the core things that are necessary for your product to work, basically. So if the platform is good enough for us, if the UX was good enough for the product managers, we would not have done that. And we would’ve had a hard time justifying it, but because it was not, we basically gave ourselves the stamp to be able to do it. So still in don’t fuck existing customers, something I read from a site, from Reforge that talk about the Safety Funnel.
So they have this thing where the typical growth funnel goes from acquisition all the way to revenue and there’s a whole bunch of different stages in between. Your goal is to maximize the number of people who go through this funnel successfully.
And I was trying to explain to people what we were trying to do with Jira Product Discovery, and I was struggling to put a name on it. And it was we’re going to work with a few, small number of customers up until it’s right for them and then we’re going to expose it to progressively more people. And I know it’s very common sense, but that’s basically what we decided to do instead of the Atlassian way, which was more, “Hey, we measure our success based on…” For example, at that time, I think it was still before all those projects, ”… monthly active users.” If that was a measure of success, we would’ve been pushed very quickly to go and expose it to as many people as we could. We didn’t want to do that because if they would have a bad experience, it would’ve been very hard to claim them back afterwards.
So someone in Reforge put the name on Safety Funnel. So the Safety Funnel is amazing. People don’t understand that. You basically put a hard stop and you limit the number of people who had bad experiences. And you do that for a while, up until you can prove it’s amazing and then you invite more people. So we did a lot of that. So we basically created that pocket away from Jira to reimagine what Jira could be for product managers. And we only exposed that to a very small number of customers at first. So that’s don’t fuck existing customers’ first principle.
Lenny Rachitsky: And the reason that’s important, just to put a point there, is the business will start to get upset if you’re making many customers upset.
Tanguy Crusson: That would’ve been it if we had triggered an incident that brought down Jira for millions of users. And first time might have been okay. I mean, by the third time we would’ve been asked to pack up and go home, I think.
Lenny Rachitsky: Got it. So the advice there is just limit how many people get exposed to your very early product, even if it’s going to hurt your numbers. You don’t want to cause people to be like, “What the hell are you doing over there, man? [inaudible 01:23:17].”
Tanguy Crusson: So [inaudible 01:23:17] your numbers is now the interesting one. So how do you frame success and how do you define metrics when your goal is basically to work with a small number of customers for a super long time? So that’s what I ended up defining for this, is based on a term that was used at Atlassian that I tried to formalize it into something that everyone could refer to, which is the Lighthouse Users Program. And so the principle for it is it’s a program, so it’s an official thing. We have hundreds of thousands of customers at Atlassian, but we would only build the experience for a small number of them. And so there are stages to it.
The first stage is we work with 10 and we prove that the problems that they had are the things that we solved and where we spend most of the time for the first 10 Lighthouse Users is to explain why they are the Lighthouse Users.
Everything that explains why we believe they are a proxy for every customer that we serve afterwards. We then have this stage from 10 to 100 where we recruit more customers, but we’re still not in the fully self-service, don’t need a lot of onboarding stuff. We’re still testing out the value, but we’re testing the different variations in the scenarios that people face to make sure that the core solution that we’ve got can cater for those. So there’s a lot of different options and different security needs. There’s a lot of subtleties that you catch between 10 and 100.
Then we get to a stage where we’re like, “You know what? It’s good. It solves people’s problems, but it’s not self-service. We just need to explain stuff. We need to work with people on Slack. We need to work with them and answer a lot of support tickets or do a lot of Zoom calls to explain stuff.”
Now we need to get from 100 to 1,000, and by getting there we basically need to solve all of that and after that we graduate. So basically by detailing that and explaining all the success criteria in between, it helps to define success in that in the make phase, we’ve got 10, 100, 1,000, and then people can understand exactly where we’re at there, and the types of questions we’re trying to answer. In the 10, we’re answering every question with a user, snippet of a video call with them, either talking about their problem and solution but showing the product and how they solve with it.
That’s how we wanted our stakeholders to view it, which is it’s very qualitative, but have it felt from the user’s perspective. It’s different when we go from 10 to 100, from 100 to 1,000, but for each of those there’s a playbook for how you go from one to the other. That helped us because we could then clean the space to do the right thing and we will not ask to hit, for example, a multi-active user’s number or number of customers’ number or percentage of users who use this number. Because it’s very qualitative still when we’re at that stage.
Lenny Rachitsky: That is super cool and super interesting. And it’s along the lines of the gates. It aligns with, at each gate, here’s how many users we’re looking for. Basically, it sets expectations for what success is, keeps you from having to scale things too early. It’s such a cool idea. And you call it Lighthouse Users, like the Lighthouse Users Program?
Tanguy Crusson: Lighthouse Users, and there’s two sides to it. We just talked about the stakeholder facing side. The one I love is the team facing side of this, which is I’ve seen a lot of teams, Atlassian and other companies as the products grow, you tend to leverage a lot more user research, formal user research or CSAT service. And the way I call it, it’s hiding behind research and not being close enough to the ground with customers. And what I want to do in the teams I work with is to create a direct feedback loop with customers, but not like someone gives you feedback and you do it. We talk…
… with customers, but not like someone gives you feedback and you do it. We talk to those customers, we build stuff for them. So there’s something I’ve seen which is climate change. It’s a thing. Everyone knows it’s a thing. We all read reports about it. We’re all like … Every time we read an IPCC report we’re like, “Shit.” Do we change anything? Not enough.
What makes people want to change and actually gives them a trigger to change? I’ve seen is a lot more empathizing with individual people’s experience. If I know someone that’s impacted by climate change, it’s going to make me relate a lot more strongly to the concept of climate change and want to change myself. Sorry for the power note here. I’m-
Lenny Rachitsky: No, that’s such a good example. Such a good way of making that point very clear of just the power of talking to a small number of users versus thinking that the more data you have, looking at data, CSAT scores, NPS retention will tell you what you need.
Tanguy Crusson: So what we do with that is we recruit 10 people and we put these people in front of the whole team, not just the PMs, PM, designer, engineering. We meet on Zoom, we chat, we work with the same ones over months to build a product. They are with us on Slack. We have regular things.
What we’ve seen, what I’ve seen going fast is that the engineers would go into a planning meeting and the PM would say, “So we should work on X,” and the engineer will go, “Wait a minute. We’ve had a talk with this customer and they struggle with this so I think we should work on that instead and fix that part of the experience,” and so on and so forth.
So all of a sudden you’re not talking about a product manager, you’re talking about a product team with product engineers. So there was this thing at Atlassian where we called some engineers were more like system engineers, some were more product engineers.
In my opinion everyone can be a product engineer. They just need to be exposed to the right user context. The right user context, what is it? It’s 10 customers you know by name, you know their context, you know their problems. You can empathize with it. This empathy makes you want to act to change your product to solve their problem and you get really huge pride coming back out of it.
It can seem counterintuitive in a company like Atlassian. 300,000 customers, right? We should build for them. Well, you can’t, right? We’re limited in our ability to make changes that will work for the vast majority so might as well embrace that. Embrace that we are indeed biased, that we are indeed reacting based on feelings, like wanting to help or not wanting to help or reacting strongly to what you’ve seen, but we are embracing that to try and build the best product possible. I’ve actually seen that the outcome is, so far on this product it’s worked.
Lenny Rachitsky: If you actually think about this from the perspective of how would a startup approach this? This is exactly how a company that’s just starting out would approach it. So it makes tons of sense to think of it this way. It’s just people don’t actually do this. It’s hard to do.
This is a segue to a question I wanted to ask. So it took probably a long time for you to show real success, real progress, real like, “This is going to work.” First of all how long was that period of just like, “This is probably not going to work too. Oh wow, maybe this is going to work”? And along the same lines, how does Atlassian slash … What have you learned about how to protect this, Pixar people call it an ugly baby? When an idea is new they call this an ugly baby. People don’t want it. It’s just like, “Get rid of this thing.”
Man, that sounds mean to say, but that’s the way they talk about it. That’s the term in creativity.
Tanguy Crusson: That’s a good one.
Lenny Rachitsky: How do you protect that because that’s the biggest challenge I think a lot of companies have is just like, “It’s been six months. No one wants this. We’re going to kill it.” How do you protect it? So how long did it take for it to show success and what have you learned about how to protect?
Tanguy Crusson: Yeah. So basically, internal comms is everything there. The way I saw my job as, a large part as being very clear on where we were, on what we were learning on, how fast we were learning, how fast we were moving. Being super honest every step of the way, not try to make shit up and try to make it bigger than it is, more successful than it is.
On the contrary, be very clear about what we’re testing, where we’re testing it. Doing that with data, doing that with personal customer stories because we all can constantly relate to the heart and to the mind. So that’s what I always try to weave into comms that I then send weekly.
So we’ve got this product internally called Atlas, which have a project on Atlas. People can subscribe to this project and then weekly you send a tweet-sized update about your project. I was using that as a platform to communicate internally about this product and there were actually hundreds of people who ended up subscribing to this thing.
Every week when we were at the stage where we were trying to get the first version out to customers it was a weekly demo of the product and everything that we built. Show the momentum as much as possible. I was even saving features for one week to the next just to … When I knew we were going to be on vacation, for example, just to have something to show this train that keeps moving, it keeps moving. You don’t want to get in front of it.
But then what I needed to do is to balance that out with all the customer side of things. So when we started to put it in front of customers then it was snippets from customer conversations. No one is going to read a research report that takes 30 minutes to read. Everyone is happy to watch a three-minute snippet with four customers talking about something. So I was using that a lot every week by posting those out there and then sharing them widely on Slack to go, “This is what we learned. This is what this customer is facing in terms of problem with Jira. This is how we’re solving it. This is them talking about it,” that kind of stuff.
So I was publishing that there and then sending that everywhere on Slack. Those kinds of things are what give people a sense of velocity and speed, and no one wants to fuck with a high-speed train. You don’t get in front of it. That doesn’t help and you’re going to look bad by doing that. So basically that ended up buying us the time that we needed to get out of this, I don’t know how you call it, the ugly baby phase.
Lenny Rachitsky: An ugly baby phase.
Tanguy Crusson: So basically it helped us get out of this phase where we don’t have anything to show for yet because what we’re showing is based on the learnings we’re trying to get and we’re trying to do that with, “Here’s some numbers, here’s some demos, here’s some user snippets.” Doing that every week because people can consume bite-sized content every week. They just struggle with the stuff where you come in three, four months later and go, “This is what we have to show for.”
Lenny Rachitsky: Right. How long was that period before it got out of this ugly baby phase for you? When did people start to get like, “Oh wow, this might work”?
Tanguy Crusson: Honestly, we had something dogfooding within two months, in the hands of the first lighthouse user within five months. We iterated on the alpha for maybe six months. We then entered a beta that lasted close to a year.
So I think at every step of the way people could see us going somewhere, and I think initially they judged, not the outcome. They judged the team a lot more than the outcome.
So there’s all this talk of founder market fit and I think really that’s what you need to be teasing out, which is, is it the right thing, going after the right problems? Then if they can answer yes in ways that you haven’t thought about asking consistently week over week, it’s enough. I think that’s where we were up until the stage where we were able to give something to customers, but the first lighthouse user was five months in. So it didn’t even matter whether it looked good or not because people could see, “Okay, that’s what … Oh, there’s something there,” right?
So I’m not sure if there’s one precise moment where it happened. However, there was this one moment where we were like, “Okay, that’s it. We’re ready to go GA.” One of the founders, Mike went, “No, you’re not. That thing is ugly. I do not want to look at it. It needs to level up with the rest of the Atlassian design standards. Our customers have expectations from us on that front. Your stuff is functional, but it’s way too early.”
So we went and fixed it. Took us two, three months and then we were okay to go. But yeah, no, because we had point A, because we had this expectation set that it would take a while to get there and because we were able to show progress every week, we basically didn’t really feel that moment where we don’t know what to do with this because it’s too early.
Lenny Rachitsky: This design improvement phase, what I’m visualizing is you’re this pirate that is invited to a nice dinner and you have to start cleaning up. You have to start looking good and become inter-society. It makes a lot of sense.
Also, what you’re describing makes me think again of this concept of, it’s your job to keep the flame alive and help it spread within the organization. So you have this idea and you’re just like, keep momentum going, keep the flame growing by sharing constant updates, sharing progress, make it very easy to consume with little snippets and videos. So there’s a lot of great advice here.
So timeline-wise, interestingly what I’m hearing is it basically took a year from idea to alpha, something like that. Is that roughly right?
Tanguy Crusson: Just like seven months … No, five months to first alpha1 customer.
Lenny Rachitsky: Okay.
Tanguy Crusson: We stayed in alpha for six months.
Lenny Rachitsky: Okay, got it. So through alpha about a year total. Many people are hearing this. On the one hand it would be like, “Of course, Atlassian has all these resources. They’re going to spend a year on this idea that who knows is going to work out. It’s so easy for them.”
I imagine there was much pain and suffering and challenge along the way that happened that makes it not so easy. Is there something you could share by just the struggle to actually make this real just as a way to wrap up conversation, go back to therapy if there’s anything you want to share?
Tanguy Crusson: Yeah. Basically, it’s pretty much what we talked about earlier with lots of processes that were not changed yet because we’re at the very beginning of point A, and people still wanting us to chime in on the things that were important for their part of the organization where we would play a part later if we become successful.
So none of that stuff was just solved once and for all. It was just a constant process to play with that, but I think we managed to get out of these things pretty gracefully in the end and I think we gained some level of admiration for some teams from that and some found it a bit … It’s always interesting to see someone breaking the rules and getting away with it. I think it inspired some of the other founders to try and do that, which was pretty cool to see.
It’s actually something I’d love to see more, hearing back from the customers I’ve been talking to about this in their company, rather than feel suffocated by the current processes and they don’t feel like they can break the chains. So I’m hoping it inspires a few more teams to give it a go, really.
Lenny Rachitsky: There’s clearly more I want to learn about point A and how other companies do this incubation stuff. So this is a really good inspiration to dig deeper into those sorts of programs.
We’ve covered a lot of ground. We’ve talked about things that have worked and haven’t worked, all kinds of therapy and pain, but also things that help you succeed and build amazing products. Is there anything else you want to share or leave listeners with before we get to a very exciting lightning round?
Tanguy Crusson: Yeah. I would try to balance the point I just gave, which is it’s hard so someone has to try. Always feel like you can push. I’d like to balance that a bit with be careful for yourself as well and make sure that you’re doing this in an environment that’s ready to welcome it.
So as I mentioned, I worked at Atlassian now for the past 10 years. Before that I used to work in banks, in heavily regulated industries, in a whole bunch of different areas where that kind of stuff was not okay. I could have tried to push, it would have not gone well at all.
So I think I’m deeply convinced by the fact that we don’t need so much top-down leadership. What we need is a lot of autonomous leaders, regardless of their position within the org, able to push for change, fighting for the right things, but you need to do that in an environment that’s safe for you to do so. If it’s not, I think it’s okay to consider alternatives.
So in other words, don’t feel like you’re trapped. Right? You only have one work life. Do it in places where you really believe you can do amazing work and surrounded by people that you’re excited to work with. Otherwise, I’ve been there before where it’s possible to become cynical by the weight of the things that are not possible and ending up doubting your own ability to do stuff really. So it’s a bit of a balance of push for the right things, but also watch after yourself and if the environment is not right, it’s okay to change.
Lenny Rachitsky: This is a really hot topic and common question in product management and I imagine other functions in product of just how much can actually change as an IC at a company. I hear all this advice about making change, changing culture, incubating stuff, innovating. How much can you actually make an impact versus, “Nothing’s going to change. I should go work somewhere else”? Do you have any advice there? It sounds like basically you’re saying there’s oftentimes you have no impact on how the business and culture’s going to work and you probably should go find a company like Atlassian that does learn how to incubate and innovate and think differently.
Tanguy Crusson: Yeah. This one is a … I’m not sure how to answer that really because I’ve worked with, consulted, worked for close to 50 companies. Atlassian is the first company I joined where I was at home from day one, challenged in the right way, but fuck, I can really choose my battles and go after them. Things are hard, but it’s okay.
So I joined Atlassian. [inaudible 01:41:13] found the trampoline like that, but yeah, I would just go, not settle for status quo because you cannot be the only sane person in a room. At some point you will go insane. The environment will permeate on you. You are not this entity that is absolutely permeable to everything that happens around you. Everything around you will affect you and will change you as a person. I really believe that.
So you need to surround yourself with people and environments that can help bring out the best in you, otherwise you can turn bad. I myself have been cynical in the past working in environments that were cynical and I then decided it’s not me. I do not want it to be me. It was me back then. I do not want it to be me.
So yeah, that’s my only advice, which is have the courage to ask yourselves those questions because otherwise it might change you in ways you don’t want to.
Lenny Rachitsky: Amazing advice. Really important advice. With that we’ve reached our very exciting lightning round. Are you ready?
Tanguy Crusson: I am, looking forward to it.
Lenny Rachitsky: All right, here we go. First question, what are two or three books that you’ve recommended most to other people?
Tanguy Crusson: Yeah, the first one was a book that was recommended to me by Scott Farquhar, founder of Atlassian co-founder. Who: A Method for Hiring is basically a book about how to do the thing that I for a while was really bad at, which is try to interview to understand who’s going to be a good person, not a good person to join your team.
Second one, so that’s one book about work. Two other books that are not about work. Highly recommend reading Hakim’s Odyssey. It’s the story of a Syrian refugee moving out of Syria, trying to move into Europe. We hear a lot of these stats and numbers about people trying to cross the Med Sea or basically become refugees in other countries and it’s very easy to look at it complacently up until you meet a personal story. This personal story of Hakim that was told very beautifully by this illustrator is well worth a read on that front.
The last one is a book that I’m not sure has been translated from French. It’s called Vivre avec la Terre, To Live with the Earth and it’s about how we could build a different agricultural system to the one that we have today that might have better ways to feed us into the future that doesn’t require large monoculture of vegetables that may not be as sustainable.
They managed to create a really efficient structure on very small parcels of land with a lot of species working together to control and not need the use of any pesticides and stuff like that. They’ve been always surrounded by teams of researchers from the, I think it’s called the Inria. It’s one of the French research centers, have got really amazing results, but I don’t think many know that this stuff exists and that it could actually be used out there and I think that’s a shame.
Lenny Rachitsky: I think it’s the first book that sounds recommended that’s only in a different language, but may not be translated.
Tanguy Crusson: It might be translated. I hope it is. It’s pretty thick though and it’s very technical. So I’ve read the principles bit and I would get lost in all the technical aspects of growing food, to be honest.
Lenny Rachitsky: Okay, love it. So let’s ask you your favorite interview question. What’s your favorite interview question?
Tanguy Crusson: Yeah. So I struggled a lot with interviewing and I’ve read all the standard interview questions there is out there and I’ve heard them in your podcast as well. There is one that came from this book that Scott Farquhar also told me was working really well for him is when people describe an experience, you ask them the name of the person that they worked with back then. And you ask them, “So when I call this person after our call, what do you think they’re going to say about that?”
Apparently this does something and I’ve seen it happen where people are unable to project and invent on the spot something from the lens of another person talking about them. So they might be able to talk about, “I did this, I did that, I did that. So when I talk … “Who was your boss? My boss was … So when I ask this person, what do you think they’re going to say?” It’s like, “Ah, well actually, I was only leading a part of it and maybe you shouldn’t call them because XYZ.”
That part makes people all of a sudden go out of this scripted version they give of themselves and become more real for a second. You get a bit of the authenticity coming out, which often is very hard to get to for me in interview questions. I’ve seen it happen. I find that very funny to do now because people get really uneasy for a second and then you get to something real.
Lenny Rachitsky: That is genius. Makes so much sense. Makes me wonder how I can integrate this tip into my podcast interviews. Really good tip. Next question, do you have a favorite product you’ve recently discovered that you really love?
Tanguy Crusson: Yeah, it’s got nothing with tech, nothing to do with tech. I’m kite surfing a lot at the moment and I came across hydrofoils, which is basically these things where there’s an airplane underneath you on a mast which is connected to a board. You’re on the board with your kite and you basically just fly over the water.
That invention is just brilliant. It’s relatively easy to learn and I can spend basically, parts of my weekend flying over water in as much as a breeze of wind because there’s absolutely zero friction between the board and the water. So that’s, technically speaking I think is genius the way they created that and inspiring themselves from airplanes.
Lenny Rachitsky: Is that the thing that Zuck was riding with his sunscreen with the flag or is that something else? I don’t know if you saw that photo. Don’t worry about it.
Tanguy Crusson: No, I didn’t see that.
Lenny Rachitsky: I like that you say it’s relatively easy to learn even though there’s a kite on you pulling you in the water on this thing that’s above the water. I don’t know if I believe you.
Tanguy Crusson: Yeah. I mean, there’s sports and crafts that are difficult to learn. This one you can get comfortable in 18 months to two years, which is really, really short. It could be a sport that you would have done before.
Lenny Rachitsky: Just a couple of years. Okay. I like your bar for … I’m going to ask a question along these lines, but before we get there, do you have a favorite life motto that you often come back to, find useful in work or in life, maybe share with friends and family?
Tanguy Crusson: I’ve got two, one for work and one for everything in life. The one for work is initially a quote from Obama. I can’t remember the exact quote. I’m going to paraphrase, but let’s keep it about the work. So there’s going to be moments in your career where you don’t feel valued and you wonder, “Am I doing the right thing? Am I being recognized? Am I valued? Am I at the right place?” Whenever that happens, a lot of parasitic thoughts may come in. You may face imposter syndrome and stuff like that. Happens to me super regularly.
Whenever that happens, what I do is I remind myself, “Keep it about the work,” because as long as you make it about the work there’s always work to be done and there’s always a path that emerges from that work. It’s a bit of the same of this, you’ve got this thing in yoga where once you’re committed good things will happen. Well, that’s exactly what I’ve seen and that’s helped me every time. I reread this quote whenever I’m in these moments of turmoil around me.
Now when I face those moments I also remind myself of the other one, which I invented, but I don’t think I’m the only one who has invented that one, which is, “Remember that in 100 years we’ll all be dead and forgotten. So don’t take yourself too seriously. You’re not that important. You’re probably not that important, as important as you think for the people that you are arguing with.”
That’s also the beauty in all things, which is it has an end. In the end it probably doesn’t matter so might as well give it your best shot. That’s what the existentialists back in France many generations ago were talking about. This point is close to Albert Camus’ thesis, right? “If it doesn’t make sense might as well go for it.”
Lenny Rachitsky: Really good points, really good lessons, really good mottos. Final question, you were ranked number four worldwide in a form of free diving. First of all, can you briefly describe what is free diving and then can you share one thing that might surprise someone about the sport and the skill?
Tanguy Crusson: Yeah. So yes, that’s my claim to fame. I was ranked number four in basically, the distance you can swim in a swimming pool without fins with 167 meters, which is 550 feet.
Lenny Rachitsky: And underwater?
Tanguy Crusson: Underwater breaststroke. I went further with the monofin, the thing you see back there, but breaststroke was one meter away from the French record and ranking number four worldwide that year. The one I’m most proud of is I actually went to 300 feet underwater deep, came back in one piece and I really enjoyed the whole thing.
Lenny Rachitsky: 300 feet. That’s like-
Tanguy Crusson: 92 meters.
Lenny Rachitsky: Oh, geez. I’m trying to imagine what that is like compared to a Statue of Liberty or something like that, but I’ll look that up later.
Tanguy Crusson: I look at buildings one. Buildings can be as little as 20, 30 meters and can go super higher, but 90 meters is pretty high, yeah. Last time I looked at a building like that, I can’t remember which one it was, but it’s pretty high. So it’s take one breath, go down, touch the bottom plate on the rope and then come back up.
One thing that may surprise people with this is that everyone is much more gifted at it than they think. So I used to give courses on the weekend for free diving and most people when I ask them, “Hey, how long do you think you can be underwater?” And they tell me something like 30 seconds, maybe a minute. “How deep you think can you go? Ah, maybe five meters.”
At the end of the weekend most people were able to hold their breath for two to three minutes and to go to 20 meters deep. So it’s one of those things where it looks absolutely impressive and crazy and amazing and in fact we’re naturally gifted to it. There’s a lot of physiological changes that happen when we dive that make it possible, but I think it’s fascinating how little we know about our bodies.
Lenny Rachitsky: I’m looking up what is 300 feet compared to real-life objects on perplexity. So what is 300 feet compared to real-life objects? Here we go, 300 feet. The length of a football field. Okay, yeah. I should have realized that. Close to the length of a Boeing 737.
Tanguy Crusson: Yeah. You need to look at it this way to get some perspective.
Lenny Rachitsky: Yeah, and a football field-
Tanguy Crusson: You go underwater in that-
Lenny Rachitsky: … Jesus Christ. Oh, my God. But I love your advice that people can do much better at the skill than they think.
Tanguy Crusson: Yeah. Everyone from kids to adults are really amazing at it actually.
Lenny Rachitsky: Tanguy, we’ve covered so much ground. I think this is going to help a lot of people that are trying to get better at free diving, but also at building zero to one within large companies. I am so thankful that you made time for this and that you shared some into this real talk and as you called it, therapy.
Two final questions. Where can folks find you online if they want to reach out, maybe follow up on some of the stuff you talked about and how can listeners be useful to you?
Tanguy Crusson: Yeah, so you can find me on LinkedIn if you know how to … You don’t need to know how to pronounce my name. You just need to know how to write it. I’m not super active on social media to be honest. I’m very much in the trenches building this product so I don’t do much in terms of networking to be fair, but if you are using this product for example, and you have ideas for how we can improve it for you or you’d like to share some stories about how that’s helped you or how it’s not helping you, I’d love to connect because I spend a large part of my days talking to users, responding to support tickets and stuff like that. So, yeah.
Lenny Rachitsky: Amazing. Tanguy, thank you so much for being here.
Tanguy Crusson: Thank you, Lenny. Honestly, it’s been amazing and I hope some of my gibberish is useful to people.
Lenny Rachitsky: Definitely not gibberish. I think it’s going to help a lot of people. With that, I’ll let you go. Bye, everyone.
Tanguy Crusson: Bye-bye.
Lenny Rachitsky: 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 |
|---|---|
| acqui-hire | acqui-hire |
| Atlassian | Atlassian |
| Ben Weiss | Ben Weiss |
| competitive myopia | 竞争短视(competitive myopia) |
| CSAT | 客户满意度(CSAT) |
| dogfooding | 狗粮测试(dogfooding) |
| don’t eat your own bullshit | 不要吃你自己的废话(don’t eat your own bullshit) |
| eat your own dog food | 吃你自己的狗粮(eat your own dog food) |
| founder market fit | 创始人市场契合度(founder market fit) |
| frankenstack | frankenstack |
| GA | 正式发布(GA) |
| GTM | 走向市场(GTM) |
| Hakim | Hakim |
| IC | 独立贡献者(IC) |
| Inria | Inria |
| Jira service management | Jira service management |
| Lenny Rachitsky | Lenny Rachitsky |
| Lighthouse Users Program | 灯塔用户计划(Lighthouse Users Program) |
| Megan | Megan |
| Mihika | Mihika |
| Mike | Mike |
| Noah Weiss | Noah Weiss |
| NPS | 净推荐值(NPS) |
| open company, no bullshit | 开放公司,没有废话(open company, no bullshit) |
| Point A | Point A |
| portfolio fit | 产品组合契合度(portfolio fit) |
| Product Discovery | 产品发现 |
| product market fit | 产品市场契合度(product market fit) |
| Safety Funnel | 安全漏斗(Safety Funnel) |
| Scott Farquhar | Scott Farquhar |
| squad | squad(小队) |
| Tanguy Crusson | Tanguy Crusson |
| tech-in | tech-in |
| Zuck | Zuck |
Reformatted by reformat_english.py
在大型企业内部实现从0到1的创新,往往比创立一家新公司更为艰难。本文是对Atlassian资深产品负责人Tanguy Crusson的深度访谈,真实剖析了巨头内部创新的隐秘角落。Tanguy基于十年间主导多款产品成败的经验指出,大公司极易陷入过度投资的陷阱,因此必须人为制造“稀缺性”。他深入分享了构建极度孤立团队以获取自主权、甚至违背企业既有设计规范等反直觉策略,并强调用真实数据与客户故事营造“高速行驶”的推进感,才是保护脆弱项目免遭夭折的关键。对于身处大型组织并试图推动实质变革的产品人而言,本文提供了一套沉稳且极具实操价值的内部创业破局指南。
在 Atlassian 内部从0到1构建的来之不易的经验 | Tanguy Crusson(Jira 产品发现(Product Discovery)负责人)
访谈文字稿
Tanguy Crusson: 我在 Atlassian 的产品管理团队待了差不多有 10 年了。我做过 HipChat 和 Stride,最近则启动了 Jira 产品发现(Product Discovery)。
Lenny Rachitsky: 为什么在大公司内部启动新产品、实现从0到1会这么难?
Tanguy Crusson: 公司往往容易过度投资。初创公司有挨饿的优势,因此你需要制造稀缺性。我们努力做的就是提醒所有人,事情大概率会失败,我们别把公司其他人都拉下水。
Lenny Rachitsky: 听起来最大的经验教训之一就是建立极度孤立的团队。
Tanguy Crusson: 我需要公司其他部门别来打扰我们,这样我们才能获得自主权去测试我们需要测试的东西,但这无法规模化。这也必然会违背一些设计规范。
Lenny Rachitsky: 我认为很多公司面临的最大挑战就是,搞了六个月,没人想要这东西,然后就要把它砍掉。你如何保护这种情况下的项目?
Tanguy Crusson: 要非常清楚我们在测试什么,用数据说话,用真实的客户故事说话,给人们一种速度感和推进感。没人想去挡一列高速行驶的火车。
嘉宾介绍
Lenny Rachitsky: 今天的嘉宾是 Tanguy Crusson。这是一期非常独特且重要的节目,因为我们将深入探讨你在类似播客中很少听到的话题:在像 Atlassian 这样的大公司中尝试创新和从0到1构建产品的真实挑战。Tanguy 在 Atlassian 工作了 10 多年,参与了许多内部的重大押注,有成功的也有失败的,包括像我当年超级喜欢的 HipChat,还有一个叫 Status Page 的产品,以及最近的 Jira 产品发现(Product Discovery),这是 Atlassian 历史上增长最快的产品之一,由 Tanguy 从想法一直领导到发布。我们将逐一回顾这些故事,Tanguy 会分享哪些做错了,哪些做对了,以及他在更大的组织内为创新创造空间所学到的一切,包括他们如何构建名为 Point A 的内部孵化项目。这期节目里有大量的干货和一堆非常有趣的故事,这也是节目时间这么长的部分原因。这是我做过的最长的一期节目。如果你正试图在你的组织中推动变革并促进更多创新,这期节目值得你花时间。如果你喜欢这档播客,别忘了在你最喜欢的播客应用或 YouTube 上订阅和关注,这是避免错过精彩单集的最好方式,也能极大地帮助到这档播客。言归正传,有请 Tanguy Crusson。非常感谢你来到这里,欢迎来到播客。
Tanguy Crusson: 非常感谢你邀请我,Lenny。说实话,能上这档播客我超级自豪。我一直是你的超级粉丝。只要有机会,我开车去哪里的时候都会听你的节目,是的。
Lenny Rachitsky: 我们这期节目要聊的是在大公司内部构建新产品和实现从0到1,特别是随之而来的痛苦和挑战,以及你从多次亲自操刀和观察中获得的教训。你在 Atlassian 见证了大量这类事情的发生,到目前为止你在那里待了超过 10 年,而 Atlassian 目前我不知道有十几条不同的产品线之类的。我知道很多人会来找你请教如何在大公司内部从0到1构建产品。那我就从一个宽泛的问题开始吧,你能分享一下你在 Atlassian 内部从0到1构建的经历,以及你看到的从0到1发生的过程吗?
在 Atlassian 的从0到1历程
Tanguy Crusson: 是的,就像你说的,我在 Atlassian 的产品管理团队待了差不多有 10 年了,我主要的工作就是为新业务做启动。最初我加入是为了启动云开发者生态系统,让开发者可以在 Atlassian 平台上构建应用并在 Atlassian 市场上销售。我做过 HipChat 和 Stride。HipChat 很有名,Stride 知名度则低一些。我们试图在 Slack 和 Microsoft Teams 出现之前拿下企业通信市场。我确实牵头做过一个加大对 IT 运营投资的商业计划,但毫无进展,后来我们收购了 Statuspage、Opsgenie,还有一些我试图做但没真正起步的事情。最近我启动了 Jira 产品发现(Product Discovery),它在我们的内部孵化器里待了两三年,一年前脱离孵化器正式发布了。我在 Atlassian 的战绩最多算五五开。Jira 产品发现实际上可以说是我在这里第一个取得重大成功的产品。那个产品做成了,但过程很艰难,而我之前做的所有产品都极其艰难,这种事情困扰了我超级长的一段时间。好在因为我做的是一款面向产品经理的产品,在过去三四年里我有机会和许多不同行业的产品经理交流,我意识到五五开其实已经算不错了。
Lenny Rachitsky: 太棒了,我知道这次对话会有很多干货。我很期待分享和聆听所有这些故事。
大公司内部从0到1的挑战
Lenny Rachitsky: 那么从宏观来看,根据你的经验,为什么在大公司内部启动新产品实现从0到1会这么难?你看到的通常最大的挑战和障碍是什么?
大公司做新产品的天然优势
Tanguy Crusson: 是的,所以在机会方面,Atlassian 拥有 30 万客户。我们涉足许多不同的市场,涵盖协作领域的方方面面。我们参与竞争的市场很多,这意味着我们有很多竞争对手。但基本上,当我们审视可以进入的领域时,可进入、可参与并有望获胜的领域清单是无穷无尽的。如果你是一家初创公司,做到这一点要困难得多。我们所拥有的广度使得寻找可以扩展到的领域变得更容易。我们不像较小的公司那样缺乏资源,所以我们实际上有资本尝试在某个领域发力,下一些赌注,允许其中一些失败,一些成功,这非常棒。
我们的客户,我说过有 30 万。我最欣赏 Atlassian 商业模式的一点是它非常广泛。它横跨中小企业,有初创公司在使用我们的产品,同时也有企业和大型企业在使用相同的产品,这意味着我们在很多领域都能找到利基市场并去追求,然后逐步扩展到所有这些领域,因此伴随着巨大的分发潜力。当我负责 Jira 产品发现(Product Discovery)时,我并不是从“好吧,我需要开始寻找产品经理,以及如何找到他们”开始的。不,他们早已经在使用 Jira 了。Atlassian 是一家拥有相对较深的组织层级但决策相对扁平的公司。所以它更像是一个,想象一下遍布整个组织的关键决策者网络。你的职位头衔是什么,或者你是否是管理人员,其实并不重要。决策是由推动变革的人做出的。
因此这带来了很大的赋权,但我认为这也是一种自上而下和自下而上混合的愉悦感。所以一开始它可能会让人感觉非常混乱,但一旦你知道如何驾驭它,实际上很容易去追求你关心的东西。当然,我们是一家大公司,所以有很多方式可以获得帮助。企业发展部门、研究分析师,只要我们想探索什么就可以和他们交谈,还有成千上万的客户,我只需要在可触达的社区群组里发个东西,隔天就能得到数百人申请和我交谈。所以这非常棒,任何初创公司都会想要这个。
优势为何无法转化为成功
Lenny Rachitsky: 这听起来就像是,当你发布一个新产品时,怎么可能会不成功呢?你有 30 万潜在客户可以发布。你拥有构建它的所有资源。听起来决策相对高效,是扁平化的。你有所有这些不同的客户细分群体,他们使用 Atlassian 所有这些不同版本的产品线。这简直就像是拥有发布新产品的所有可能机会,然而,很多事情还是没能成功,所以我认为这是一个非常重要的观点,而且我认为许多大公司都处于这种境地。我们拥有如此多的机会。我们要构建的一切都会,它将会疯狂增长,因为我们拥有所需的一切,但最终还是没能成功。这就是为什么我认为我们将要讨论的内容会如此重要。
Tanguy Crusson: 对我来说这会有点像心理治疗。希望外面的一些人会说,“好吧,不只是我一个人在吃苦头。”这在像 Atlassian 这样的公司也会发生。
Lenny Rachitsky: 太棒了。
Tanguy Crusson: 是的,那我们就来谈谈挑战吧。
Lenny Rachitsky: 开始吧。
成功门槛与评估指标的错位
Tanguy Crusson: 是的,为了那个。所以你想开始一个新事物。这个事物需要时间,你需要能够拥有这段时间,直到你能够证明这个事物到底存不存在。在 Atlassian 内部的问题是,对于一个新的赌注来说,成功的门槛非常高。如果你进来创造了一个产品,它有 100 个客户,这看起来会挺可爱。记住,我们服务于初创公司和企业,我们有自助服务和销售,我们所有这些为更大产品建立的机制都已经就位。基本上,一个 1 亿美元的业务只是一个好的开始。在外面的大多数公司,1 亿美元的业务是本垒打。对我们来说不是这样。我们试图建立真正庞大并能随着时间的推移持续庞大的业务。
现在,评估成功在早期阶段产品和成熟产品之间可能看起来非常不同。在 Atlassian 的很长一段时间里,我们对所有事物的对待都有些一视同仁,因为成功的指标是相同的。例如,像月活跃用户这样的指标,很长一段时间我们都会去看,那个产品在这方面成功了吗?而如果你正在构建一个“内部初创公司”,你的月活跃用户数量在相当长的一段时间内应该看起来非常低,直到你知道你的产品已经准备好服务于你想要推向的绝大多数客户,这样他们就不会只是看一眼然后说,“它还没准备好为我服务。”
他们会去尝试,然后他们会流失,而且把他们挽回过来会花掉无尽的时间,所以这使得尝试开始新事物变得非常具有挑战性,除非我们在内部拥有正确的指标和流程以及一切,能够为内部赌注的成功提供空间和呼吸余地。而在很多年里,我们并没有做到这一点。我们最近随着 Point A 的启动开始达到那个状态,那是我们的内部孵化器项目,这也是我最近一次成功的尝试。之前的那些没有这个,我真的为此挣扎过,我看到许多公司也在这个确切方面挣扎。
失败案例:HipChat
Lenny Rachitsky: 太棒了。让我们深入一个真实的故事。我有三个想谈的。有你提到的 HipChat,有 Status Page,然后还有你现在正在做的 Jira 产品发现产品。说到 HipChat,有个有趣的故事,我很喜欢 HipChat。当年在我的初创公司里我是 HipChat 的超级用户。我永远忘不了你们为了推广 HipChat 而挂出的那个广告牌,上面有一个火柴人表情包的家伙,上面只写着,“为什么要用 HipChat?”我觉得那是最搞笑的事情,而且那个产品太让人愉悦了。里面就是有各种各样的小表情,而 HipChat 对于 Atlassian 来说基本上就是要成为 Slack 杀手,那就是愿景。
Tanguy Crusson: 你真是要我的命。我们比 Slack 早多了。
Lenny Rachitsky: 好的,所以先发优势,惊艳的产品。
Tanguy Crusson: 是的,行得通。
Lenny Rachitsky: 而且它是 Atlassian 收购来的。
Tanguy Crusson: 是一次收购。
Lenny Rachitsky: 太棒了。那我们来谈谈 HipChat 哪里出了问题。你学到了什么?
Tanguy Crusson: 大家都知道 Atlassian 这家公司,作为这个工具,几乎可以有一个,好吧。[听不清]。
Lenny Rachitsky: 是的,这就是心理治疗。治疗环节现在开始。
HipChat 的早期崛起
Tanguy Crusson: 是的,就从那里开始了。我可以告诉你,我和 HipChat 团队的其他人都是这样。好吧,是的,HipChat 是一次收购,团队大概 20 人左右。它是 Slack 出现之前的 Slack。表现出色,很受初创公司的青睐。在当时,这是一种全新的协作方式。有几个这样的小应用试图做这件事。我记得刚加入 Atlassian 的时候,在那之前,我在金融服务、银行之类的行业工作,我们要么开大会讨论事情,要么走到别人的办公桌前去讨论。我加入了这家公司,我的同事和我坐在同一层楼、同一张桌子旁,我们却通过电脑聊天来交流。起初我经常觉得很奇怪,转头看着正在和我交谈的人,我们在争论,但却是在通过文字争论。无论如何,对于出生在 Slack 世界里的人来说,这可能看起来有点可爱,但在当时这是一个重大的改变。
Lenny Rachitsky: 是的,我记得。我记得我和我的团队在同一个办公室,我们使用 HipChat 聊天,感觉有点奇怪。现在这已经完全正常了。
HipChat Go Big 与重写的代价
Tanguy Crusson: 这就是很正常,而且 HipChat 是最早进入这个领域的之一。Slack 不知从哪里冒了出来,这家公司最初其实更专注于游戏,然后他们真正席卷了市场。当我们看他们的增长数字时,令人眼花缭乱。所以在某个阶段,HipChat 在 Atlassian 内部相对被冷落了一段时间,你们做得不错,继续推进吧。但面对 Slack,我们现在必须尝试做得更大,所以我们启动了一个叫做 HipChat Go Big 的项目,团队,他们招募了——
Lenny Rachitsky: 这是项目的名字?HipChat Go Big?
Tanguy Crusson: 是的,HipChat Go Big。然后是 HipChat Next Gen,有几个不同的版本。无论如何,是的。
Lenny Rachitsky: 非常清晰。HipChat Go Big,我喜欢。
Tanguy Crusson: Go big。但这真的是一次 go big,大量的新开发者、新产品经理、设计师,整个公司都在支持这个产品。对于一个之前变化没那么快、已经达到了很好的产品市场契合度(product market fit)、每天都有几十万用户的产品,我们突然有了一大群人想要对它进行修改,以对抗这个新的威胁,试图非常激进地扩大它的规模。平台还没准备好让这么多人同时在上面工作。于是我们不可避免地走到了那一步,好吧,技术债太多了,我们无能为力,所以我们做出了重写的决定。关于是否应该重写,有很多文献。如果你现在问我,我会告诉你永远不要。相信我。
Lenny Rachitsky: 大多数建议都是这样的,永远不要重写,但人们还是会这么做。
微软 Teams 的冲击
Tanguy Crusson: 永远不要重写,这有很好的理由。但基本上我们还是做了,结果实际上诞生了一个叫 Stride 的新产品,它最初是“HipChat next gen”。问题是这个产品很棒,但等我们做完的时候,Slack 已经把我们甩开几条街了。就在那时,微软推出了 Teams。我不知道你是否记得那一刻,Slack 在《纽约时报》上发了一个广告,模仿很久以前苹果对战微软的那一套,说什么“欢迎来到游戏,欢迎来到派对,我们欢迎你这样的竞争者”之类的,然后 Slack 几乎被 Teams 摧毁了。我们开始意识到这是因为微软的分发优势。每个 Office 用户都会得到它,他们把它作为 Office 的一部分免费提供,我想几个月前才把它拆分出来。
Lenny Rachitsky: 这很讽刺,因为理论上 Atlassian 也有同样的优势,对吧?你们有所有这些产品,你可以把它捆绑销售。
Tanguy Crusson: 我们实际上稍后就会谈到这个。
Lenny Rachitsky: 好的,太棒了。
Tanguy Crusson: 因为那曾是我们认为能赢的原因,而我认为那也是我们输的原因。
Lenny Rachitsky: 太精彩了。我们深入聊聊。
退出市场与阵痛
Tanguy Crusson: 总之,这一切都发生了,最后我们退出了市场。我们把 HipChat 和 Stride 卖给了 Slack,基本上退出了企业通信市场。
Lenny Rachitsky: 补充确认一下,你把它卖给了 Slack,现在基本上就是 Salesforce 了,这个故事就结束了。我不知道大家是否知道这件事。
Tanguy Crusson: 是的,实际上,市场注意到了这一点,我们一这么做,股价就上涨了。我记得当时大概是 60 美元,然后涨到了 70 美元。我的意思是,这对我们这些在这个团队里工作的人来说真的感觉很糟。首先,没人告诉你,失败,大家都告诉你失败是很棒的,因为你能学到很多东西,但一开始面对失败的感觉真的很糟。我们团队里没有人对此感到高兴,因为我们投入了数年时间,就我而言是三年,但在那之前做 HipChat 的人时间更长。为此着迷,为每一个细节、每一次客户对话、每一个解决方案着迷,我们要不要重写,我们要不要做这个,我们要不要做那个?
Tanguy Crusson: 无数激烈的对话,然后一夜之间,这些都不再重要了。我们一直在做的东西,就这样,结束了。我相信很多初创公司以前都经历过这种情况。对我来说,这是第一次感觉如此切身相关,而第二天市场反应是:“哦,你们停止你们在做的事了?太棒了。是的,给你们股价加 10 美元。”所以,是的,所有这些故事都有个人的一面。对我们来说,在关闭 HipChat 之后,有点像经历了悲伤的七个阶段。
Lenny Rachitsky: 对你和团队来说,那段哀悼和经历各阶段的时期持续了多久?
Tanguy Crusson: 持续了几个月。我不是决定关闭 HipChat 的决策团队的一员,我是团队里的产品经理之一,负责领导三个支柱之一。我被叫进去,我想是在宣布团队解散前一两个月吧,他们对我说:“嘿,Tanguy,顺便告诉你,HipChat 不复存在了,你现在的新任务是在这之后为团队寻找一个新的使命。”
Tanguy Crusson: 我们基本上在接下来的两三个月里,都在努力确保建立起能够完全掌控其业务的 squad(小队),确保是他们去和客户交谈,是他们去制定策略,去寻找解决方案以及该领域的所有事情。只要涉及人们在其他界面上与 Atlassian 产品互动,一切都在考虑范围内,所以经历了很多起伏之类的。我想大概花了两个月或三个月,我们才回到一个新的节奏。而有些人,当我们谈论这件事时,基本上仍然心有余悸。
不要吃你自己的废话
Lenny Rachitsky: 从那次经历中得到了哪些教训?
Tanguy Crusson: 是的,我个人从中得到的主要教训,要回到你谈到的那个假设,即我们拥有所有这些成功的产品,我们可以扩展到这个领域,我引用一下,这只是我个人的说法,不要吃你自己的废话(don’t eat your own bullshit),这是两件事的结合。我们有一个公司价值观叫做“开放公司,没有废话”(open company, no bullshit)。所以我们需要能够实事求是地谈论事情,我们不试图让事情听起来比实际上更聪明。我们不试图隐藏真相,我们追求真相,我们彼此不隐瞒。我们与每个人分享,我们默认是开放的。这是其中一项价值观,而且我们做了大量的狗粮测试(dogfooding),也就是吃你自己的狗粮(eat your own dog food)。我们确实大量测试我们自己的软件。
成功套路在新市场的失效
Tanguy Crusson: 我注意到,我们有时会倾向于相信某些东西,仅仅因为它过去对我们有效,并且我们会假设它将永远有效。创始人不断告诉我们,带你走到这里的,不会带你到达那里。这是我们反复听到的一句话。但当团队看到某件事成功时,很容易认为它是因为某个原因而成功,而这个原因并未经过验证。这就回到了我们今天的话题,为什么在成功的公司里做这些事情有时更难。Atlassian 凭借一套打法获得了成功,这套打法就是我们在开发者或技术、IT部门有用户。他们选择了 Atlassian 的应用并喜欢它们。他们开始向业务部门的人推荐,然后我们就会看到应用在公司内被采用,在人们决定全面标准化使用 Atlassian 之前,出现了自下而上的采用。我们打赌可以把这套打法应用到这个市场,基本上就是,我们可以从使用 Jira 的人入手引入 HipChat,然后技术团队的人会首先进入 HipChat,接着它会扩展到业务团队,基本上就是这样从一个圈层扩展到另一个圈层。问题在于,在我看来,我们没有足够早地去验证这个假设。我们做了大量的工作,甚至在面对这可能行不通的信号时,还在其他事情上做了大量工作。我确实记得和很多客户交谈过,他们会说,“嗯,我们的 IT 在用 HipChat,但业务部门更喜欢 Slack。” 然后我们开始看到这些企业选择 Slack,这最初就像开发者尝试了之后喜欢上了,然后大家开始采用。在那种情况下,Slack 成功在非技术和 IT 角色中建立了一个非常强大的拥趸基础。那时正是应用消费化趋势开始变得高涨的时刻。Slack 真正乘上了这股浪潮,他们将体验中的一切都集中在捕捉这一趋势上。他们对新手引导进行了游戏化,在外观和感觉上投入了更多精力。他们试图让工具变得令人愉悦、现代且实用。从那以后,我们从他们的做法中学到了很多,而我们当时错失了那部分。我个人从中得到的一点是,让我们成功的因素中包含了大量假设,这并不意味着它会一直有效。
购买者与使用者的影响力差异
Lenny Rachitsky: 我理解你的意思,这非常有趣,Atlassian 在向组织内的买家也就是 IT 团队销售方面非常成功,因为他们拥有所需的一切。他们满足了所有的条件,但结果在 Slack 的案例中,最终是用户对他们选择哪种工具产生了最大的影响力。
Tanguy Crusson: 我实际上更倾向于将其表述为双方都在争取用户。Atlassian 争取的是技术团队中的用户,Slack 争取的是业务团队中的用户。在这两种情况下,发生的都是自下而上的采用。而另一边的人,业务部门更喜欢 Slack,开发者更喜欢 HipChat。在我负责的动态流方面,我们做了大量工作。我们与市面上的每一个开发者工具进行集成,确保他们使用的每一个工具的信息都能流入 HipChat,并且从 HipChat 回流到那些工具中。基本上,他们可以通过查看 HipChat 中的活动流来完成大量工作。业务部门呢?对此并不怎么兴奋。表情符号,还有很多其他的东西,我记得在某个时刻我们认为那些东西是微不足道的。不,它们并不微不足道。这只是一群我们没有充分沟通的不同用户群体,在使用工具时采取的一种不同方式。
验证假设的重要性
Lenny Rachitsky: 所以这里的教训是,不要低估说服一个新的细分市场购买你的产品所面临的挑战。你可能认为他们很接近或相似,但很可能并非如此。
Tanguy Crusson: 是的,这是其中之一。另一个教训是,带你走到这里的,不会带你到达那里。所以,回过头去试着解释你今天为什么成功,然后如果你认为可以在下一件事上使用同样的方法,去找到验证它的途径,找到测试它的途径。不要仅仅基于这些假设就去构建。这基本上是我从这次磨难中得到的最重要的东西,也应用到了我之后的工作中。
Lenny Rachitsky: 你会怎么做?你会如何去测试它?是通过用户研究吗?是产品经理去和潜在用户交谈吗?你在那里会怎么做?
产品发现的早期验证
Tanguy Crusson: 例如,当我们开始做 Jira for Discovery 时,也许我应该先简单介绍一下这个产品。这是一个为产品经理打造的产品,主要用于优先级排序和路线图规划。人们使用 Jira 在工作确认后进行计划和跟踪。我们想在工作确认之前创建一个空间,让所有人可以就优先级进行讨论,无论他们是应该参与优先级排序过程的开发者、设计师等产品团队内部人员,还是客户成功、销售、支持、领导等外部人员。当我们开始时,我们认为,好吧,产品经理已经在使用 Jira 了。你知道吗?我们可以触达他们。所以我们创造这个工具,然后通过 Jira 进行分发。我们本可以走上先构建然后再开始分发的道路。相反,我们做了一些这样的事情,在写下第一行代码之前,我们在 Jira 的简报里投放了一则广告,写着:“嘿,我们即将推出一款面向产品经理的产品发现(Product Discovery)工具。” 然后我们建了一个网站,在写下任何一行代码之前就写着:“嘿,产品经理,你的工作很艰难。我们想提供帮助,如果你想与我们同行,请在这里留下你的名字”之类的话。也就是在那时我们看到,我想是在两周内,有超过 3000 人注册了候补名单。我们觉得:“好吧,很酷。需求得到了验证。” 我们正在与感兴趣的人交谈,而且我们能够触达他们。这就是我后来尝试做的事情的一个例子,基本上只是为了确保我们在游戏更早的阶段就去验证我们的假设,而不是等到为时已晚。
Lenny Rachitsky: 太棒了。这是一个非常棒的、极具战术性的例子。
Tanguy Crusson: 是的,大多数都是如此,我今天要谈的没有什么是革命性的。很多时候只是试图去应用,在正确的时间提出正确的问题,并真正尝试通过各种手段去回答它。
Lenny Rachitsky: 在我们转向讨论其他产品之前,HipChat 的经历还有什么其他教训吗?
竞争短视(competitive myopia)
Tanguy Crusson: 我有两个,我会快速讲一下。第一个是竞争短视(competitive myopia),不要陷入其中。在某个时刻,Slack 真的越来越受欢迎,赢得了越来越多的心智份额。Twitter 上的每个人都在赞美他们。甚至当他们的服务出现中断时,他们都能得到祝贺。我当时想:“这简直疯了。” 但这种喜爱如此强烈,而我们倾向于退回到大脑的功能性思维,觉得好吧,我们只需要再加这一个功能。我们只需要这一个功能,我们只需要这一个功能。我们最终只是在被动应对一切。
Tanguy Crusson: ……再多一个功能。我们只需要再多一个功能。我们最终只是在被动应对竞争对手的任何举动,我认为这非常非常糟糕,因为那时我们基本上失去了迄今为止让 HipChat 成功的因素,也就是把某些用户服务得非常好。相反,我们最终变成了基于竞争对手的举动而快速跟进,这超级糟糕,因为如果你把竞争对手所做的事情想象成一座冰山,露在水面上的顶端是他们发布的那些功能,但这建立在他们在研究和理解客户群等所有方面所构建的大量基础之上。因此,根据他们现在发布的东西,你看到的可能只是他们一年前想法的具象化。但我们却陷入了这种境地。
现在无论我做什么工作,我都倾向于尽量忽略竞争,除了大概每三个月看一次,看看他们发布了什么,有没有什么我们需要担心、害怕的东西之类的。但真正的重点是,尽量将所有创造过程和研究过程与竞争对手的所作所为脱钩,因为这是无法比较的。市场是巨大的。外面有数十万家公司。不是每个人都有相同的需求。我们服务的是特定的细分市场。比起关注竞争对手在做什么,我们从这些用户身上学习,然后再扩展到其他群体,会做得更好。
Lenny Rachitsky: 这就是你给团队的建议吗,就是忽略竞争,也许在重大发布之类的事情上关注一下就好?
Tanguy Crusson: 我们总是看到在 Slack Channels 团队里大家会分享,好吧,他们刚在 AI 上做了这个,刚做了那个。所以这些信息不断涌现,因此我们经常会进行讨论,回到原点,好吧,让我们再看一遍过去三周我们做的用户访谈。让我们现在一起看这些访谈,记住我们是在为谁构建这个产品。所以有很多尝试是为了重新锚定在我们已知的事物上,以及基本上如何在此基础上构建我们自己的旅程。我认为让每个人都参与进来会丰富得多。
Lenny Rachitsky: 我很喜欢这个。所以你发现,当有重大发布,大家都说“天哪,看看 Slack 做了什么”或者“看看这家公司做了什么”时,你的反应就像是,“好吧,现在让我们花点时间提醒自己,我们的客户一直要求我们做什么,让我们看几个用户访谈。”
Tanguy Crusson: 甚至当竞争对手做对了某些事情时,请记住,我们在打持久战,我们并不总是需要成为第一个、最闪亮的那个。我们需要确保人们有一个问题,我们解决这个问题。他们告诉我们,我们为他们解决了这个问题。当我们为他们解决这些问题时,他们会感到高兴。所以这些才是我们应该痴迷的东西。
Lenny Rachitsky: 我很喜欢。你说过你从这个经历中还有一个教训。
Tanguy Crusson: 是的,最后一个教训是,初创公司有饥饿的好处。对吧?我们是一家大公司,我们可以把大量资源投入到我们感兴趣的事情上。所以这种重塑 HipChat 的想法,伴随着我们要在微服务等新平台上重塑 HipChat,我们构建的一切都可以在所有其他产品中复用。所以你看到的聊天文本框,它是一个可以全屏打开的编辑器,它是一个 Confluence 编辑器,你可以做你在那里能做的一切。它是作为平台组件构建的,当你在开始构建产品时平台就已经存在,这非常棒。真正困难的是当你试图同时做这两件事时。所以我认为那部分,如果我要再次参与 HipChat 并说我负责这件事,我会在那部分说,好吧,我们需要先赢得问题空间,如果平台在那里,我们就使用它。如果它不在,我们就先用简陋的方法凑合,与客户一起测试、迭代,然后不管那里有什么好东西,我们以后再将其平台化。
但这正是你看到让我们作为大公司变得超级强大的东西,也会拖慢我们的速度,让我们把注意力集中在错误的假设上。所以在那种情况下,我认为我们以为我们会赢。有趣的是,我认为我们确信自己在这个市场上大有可为。同时我们认为自己可以搞定重写,也可以搞定平台化。所有这些事情都是必要的,但同时做所有这些可能有点太贪大了。现在我这么说,但今天你在 Confluence 中看到的编辑器就是从我们当时在 HipChat 中做的事情开始的。所以我想说,就代码而言,纯粹就代码而言,我们为 HipChat 编写的代码可能有 70% 今天仍然在 Atlassian 平台中。但对于一个新赌注、local Optima 来说,那真的很糟糕。
Statuspage 的经历
Lenny Rachitsky: 是的。想想如果这事成了,Atlassian 本可以拥有这个价值 300 亿美元的业务,我能理解为什么人们会因为它没成而感到沮丧。所以谢谢你分享这个故事。让我们来谈谈 Statuspage,你参与的另一段旅程,它也没有完全按照人们希望的方式发展。我们的心理治疗继续。谈谈那是一个什么产品,以及那里发生了什么。
Tanguy Crusson: 是的,所以这个其实是一个成功的故事,但它在我离开之后才成为成功的故事[听不清 00:34:08]。
Lenny Rachitsky: 好的。
Tanguy Crusson: 这更像是一个大公司可以打持久战的故事,而对于在这个过程中的你个人来说,它可能看起来像是一场失败,你可能会觉得自己毫无进展,然而公司在这个机会上坚持了足够长的时间使其实现。所以我们要在这里谈论的是,我在其中工作时面临的个人挑战,这个东西最终非常成功,但我当时在个人层面上感觉是个失败。所以 Statuspage 在某个时候……Jira 是开发者使用的,对吧?在那个时候,我想是在 2016 年甚至更早的 2015 年,每个人都在向云端迁移,每个人都在采用 DevOps,你构建它,你运行它。所以基本上开发者会实现软件,然后将其部署到生产环境。在部署到生产环境之后,他们不会把它扔过墙交给运维人员。是同一批开发者在生产环境中操作这个软件。他们随时待命,等等。
所以那时我做了市场调研,看看 Atlassian 是否应该进入这个领域,我们是否有机会去追求 IT 运维周围的所有工作。所以 Jira 基本上是用于构建软件的 Jira 软件。我们能有一个用于运维的 Jira 吗?也就是用于操作这些软件。因此我做了市场调研,发现了几家在那里做着超级有趣事情的公司,比如 [听不清 00:35:37]、Opsgenie、New Relic。BigPanda 是一家在那里做很多酷东西的小型初创公司,还有 Statuspage。后来我发现了 Statuspage,发现他们的产品非常有趣,因为 Atlassian 不是一个运维……我们并不真正构建非常深度的运维工具,我们非常关注的是围绕一切事物的协作方面。当我在观察处理故障事件的团队时,我意识到存在大量的无头鸡综合征,人们像一群无头鸡一样跑来跑去。
Tanguy Crusson: 事情闹得天翻地覆,接下来你会看到团队手忙脚乱,所有人都搅和在一起,急于立刻解决问题,但同时又在对发生的事情提出质疑,争论为什么会走到这一步。你的老板会发消息问:“嘿,怎么回事?我听说应用挂了。我们正在亏钱。”客户会发邮件问你发生了什么,你的支持渠道会被挤爆。销售人员很焦虑,因为他们的客户在给他们打电话。所以基本上,这对每个人来说最终都是一种压力极大的体验。Statuspage 提供的东西看似非常简单,那就是,你应该为你的服务建立一个状态页,把它告诉你的客户,客户可以订阅这个状态页。在那里,你有你的各项服务,当这些服务出现故障时,你可以发布一个事件公告,你的客户会收到通知,这意味着他们不会再联系客服,因为他们知道你们正在处理。这会与客户建立信任,因为基本上你对他们保持了公开的沟通。他们更有可能对你的处境产生共鸣并给予支持,而不是抱怨“哦,这破玩意儿又罢工了”。所以这其中有着非常多的好处。
透明的力量
Lenny Rachitsky: 既然谈到这个话题,我想分享一个小故事。
Tanguy Crusson: 当然可以。
Lenny Rachitsky: 巧合的是,十年前,我曾在一家网站性能监控公司工作,当时我开了一个叫 Transparent Uptime 的博客。我的整个博客都是关于在宕机时保持透明的力量,告诉人们现在什么东西坏了,什么时候能恢复。我当时对这件事非常着迷,深入参与了那个领域。所以当我在当年看到 Statuspage 以及 Atlassian 的运作时,我心想:“我太喜欢这个了。”实际上我还和创始人们聊过几次,因为他们当时也很喜欢我做的那项工作。这和我生命中做过的其他任何事情都毫无关系,但我当时对这个非常奇怪的话题充满热情,我很高兴看到公司们开始接纳它。
发现与收购 Statuspage
Tanguy Crusson: 我真的非常享受那几年对这个话题的深入钻研,这是一个了不起的话题。总之,这个特定领域有很多迷人的地方,但在那个时候……所以我们发现了 Statuspage,我们邀请了几家这样的公司和我们共事一周,意思是,嘿,我们在 Atlassian,基本上是开发团队所有协作活动的中心。如果把一切整合在一起,什么样的体验才能将正确的信息传递给正确的团队以便他们采取行动呢?所以我们在旧金山一起搞了一周的黑客马拉松,有来自 New Relic、Statuspage 和其他几家公司的成员。从中产生了一些非常有趣的概念,确实有东西在里面,我当时想,“好的,我应该开始为 Atlassian 起草一个关于 IT 工具/运营的商业案例了。”现在,像 Atlassian 这样的大公司,我们有钱。所以在我们的每一项战略中,我们都有银行里的现金,我们总是会看应该自建、购买还是合作?收购可以非常强大地加速你。我们有很多成功案例,我不记得我们做了多少次收购,但 Trello 就是一个很好的例子。所以我们决定买下 Statuspage,而我负责 Statuspage 业务在 Atlassian 业务内部的整合工作。
并购的痛点与教训
Lenny Rachitsky: 感谢你分享所有这些背景。从这次经历中你学到了什么?听起来基本上当你身处其中时非常痛苦,但最终却非常成功。从这种痛苦中得出了哪些教训?
Tanguy Crusson: 我从中得到的教训是在收购方面。所以大公司,我们有钱,我们可以买下一家公司,这会让我们走得更快,但情况并非总是如此。在我的案例中,有很多事情并没有我想象的那么容易。第一个就是加入你们公司的初创公司所经历的文化冲击。想象一下你是一家初创公司,我不知道,大概有二三十名员工之类的,事情进展顺利,你完全掌控着自己的命运。然后一家公司买下你为了加速你,但你却不再拥有所有的决策权。所以 CEO,也许你变成了一个产品人员。那个负责走向市场(GTM)但同时也做一堆产品工作,可能还做一些工程工作的人,突然之间就只在营销部门工作了。
有很多决策是在你头顶上做出的。例如,产品组合契合度(portfolio fit),哪些应该是你产品的一部分,哪些应该是使用我们现有平台的东西。所以有很多我们能够在日常基础上做出的决策,开始脱离你的掌控。大公司看得更远,所以 Atlassian 会着眼于长期博弈。所以我还记得 Statuspage 刚加入时,我们询问他们的路线图,他们回答说,嗯,未来三个月我们在做这个,接下来的三个月我们在考虑可能做 X、Y 或 Z。所以当我们问,“好的,那么三年计划是什么?”他们会说,“你说的三年计划是什么意思?我不知道,我想我们会活下去吧。”这正是初创公司的做法。他们对未来有非常粗略的想法,但远未达到大公司期望的程度。所以这是一个非常大的文化冲击。
你最终面临的,还有那些希望被收购或正在收购其他公司的公司,在此之前,你是一家初创公司,所有人都是一个团队。取决于收购公司的组织方式,你的团队内部可能会出现孤岛。Atlassian 的组织方式就像许多 Atlassian 公司一样,我们有一个产品组织,一个工程组织,一个营销组织,一个设计组织,它们各自向不同的领导者汇报,这些领导者可能然后再向同一个人汇报,但大体上仍然是不同的组织,然后被组装成 squad(小队),这些 squad(小队)一起运作。但是有些仪式是属于 squad(小队)的,而有些仪式则是根据你所处的专业领域向上延伸的。所以我确实记得 Statuspage 团队刚加入时,理解如何在这种结构中游刃有余是多么令人生畏。
如果我想雇佣一名新设计师,我不再和 Statuspage 的 CEO 谈话了,我需要和这个区域的设计负责人谈话,而这个负责人在收购前和 Statuspage 业务几乎没有关系。所以这就是当我开始做整合工作时人们告诉我的话,也就是,嘿,记住,你目前还不知道,但整合主要关乎人。它们并不那么关乎技术或产品愿景。所有那些东西都是容易的部分,难的部分是人。起初我不太理解,但到最后我真正明白了。
因为我们告诉这些公司的是,我们可以通过收购来加速你们,但实际上他们将面临更多关于如何管理员工、绩效评估、工程资源分配这个概念、收入预测、OKR、长期路线图等内部流程,所有这些东西对他们来说都变得非常陌生。所以公司的一部分自上而下地说,我们买下你们是因为你们很成功,继续做你们在做的事。而同时,许多其他团队虽然没有恶意,但基本上最终会不断与你们互动,以确保遵循了正确的流程,以确保做了正确的事情。
**Lenny Rachitsky:**这是一个非常有趣的观点,即一个群体说继续做你们在做的事。我们不会干涉你们,你们是专家。而另一部分实际在一线构建它的人则会说,“嘿,用这个组件来构建。嘿,我们需要这个流程,我们需要这份文档。”
**Tanguy Crusson:**你过去能够将90%的时间专注在打造产品上,突然之间,所有那些看似寄生般的东西涌进来,不断打断你。而且通常新加入的人不知道,他们基本上不仅是被收购了,他们是被这家公司雇佣了。所以他们基本上是加入了一家有着不同仪式、不同文化的不同公司。基本上所有这一切都非常不同。每次收购的情况都不会一样。就像我说的,我们有一些非常成功的收购。Statuspage 最终成为了 Atlassian 内部的巨大成功,但当我在那里时,我基本上是在经历它最困难的部分,那时我觉得,“这并不像人们可能想的那么简单。”所以一个警告是,如果你计划进行收购,确保你把这些都考虑进去,并思考整合计划来弥补这些方面。所以你不要指望业务加入后还能以完全相同的节奏运转,因为在再次加速之前,会出现一个巨大的放缓。
**Lenny Rachitsky:**那么也许作为这里的最后一个问题,假设你要再次进行一次收购,我不知道你是否还经历过更多,你会改变的一件事是什么?你会推荐的,让我们确保以非常不同的方式来做这件事的一点是什么?
技术栈与产品整合
**Tanguy Crusson:**有一个我打算谈到的方面,以此引出那个话题,但有一个我们没谈到的方面是我们买了一个产品,这个产品如何与剩下的部分契合?所以我们做过不同类型的收购。一种是我们买下公司并保持产品运行,然后尝试将其与我们的技术栈整合。然后另一种收购是我们买下公司,然后我们有点像在我们的平台上重建,或者我们买下它是因为它与我们的平台有巨大的协同效应。否则我们会称之为 frankenstack。也就是,你这里有了一个技术栈,那里有了一个技术栈,它们不能很好地互相通信,身份验证不同,整合方式也不同。所以它看起来就像一个拼凑起来的产品,这不是我们的客户希望从我们这里得到的。
所以下次我个人尝试的话,我会把它当作招聘来对待,而不是仅仅当作购买业务。因此会有很大一部分工作是既要教育也要引出将团队真正雇佣到公司内部意味着什么。同时我这么说是因为我不想做一次大规模的收购。我下次想做的是相对较小的。找到一家拥有惊艳产品的公司,我们可以将其作为 tech-in 引入,关闭产品,在我们的平台上重建它,所以这基本上等同于 acqui-hire,因为基本上我们买到的是我们路线图的加速。
我们可以尝试组建一个团队来做他们做过的事,但他们已经成功了,所以他们知道自己在做什么。我们大概能提前一年达到那个目标。在 Atlassian 这样的规模下,如果我们能提前一年进入一个市场,这对我们的收入意味着什么?这很可能自己就能把收购成本赚回来。所以我从中得到的经验基本上是它需要被当作招聘来对待,而我下次想做的基本上就是更多地像那样去做。
**Lenny Rachitsky:**那么 HipChat 的情况呢,感觉在那里重建一个东西是个错误。对于这种初创公司,是不是对于这种初创公司就不要重建了?重新开始。你对于如何区分这两种情况有什么想法吗?
**Tanguy Crusson:**实际上有很大的区别。有很大的区别。也就是在一种情况下,你试图做的是……所以当你重建时,你有一个成功的业务,例如你有几十万客户,而你正试图重建同样的东西或不同的东西,但面对的是对你的现有产品有期望的同一批客户。在另一种情况下,你买下了一家有一些初步势头、但不是太多的公司,所以你可以关闭业务,然后在你的平台上重建,以触达你的客户群。所以这不等于试图在飞行中重建飞机,这是在推迟起飞。这就像,“是的,我们做了一次试飞,飞机降落了。好的,很酷。换一架飞机,再次起飞。”
**Lenny Rachitsky:**明白了。我很喜欢这个比喻。在我们进入你的产品发现之前,从这次经历中还有其他关键的见解和教训吗?
**Tanguy Crusson:**最后一个,然后我就要停止这次心理咨询了,非常感谢你的提供。
**Lenny Rachitsky:**不,不,我们得继续下去。我得把这次心理咨询的账单累积起来。
“为什么是现在”的紧迫感
**Tanguy Crusson:**也就是,记得我提到过我当时正在做一个商业案例,关于……基本上是在 IT 运维领域做大。所以其中有 Statuspage 作为一部分,但我还有很多想能够在 Jira 和一大堆其他产品里做的事情,基本上是为了在那个领域、IT 运维领域做大,在 Jira service management 真正进入那部分市场之前,Jira service management 是我们在那方面非常成功的一款产品。每个人都很兴奋。那是在我们内部有孵化器之前。所以我试图推销它,比如让我们构建一个围绕那个东西的新产品。它在 Jira 上,并且与我们讨论过的这些工具集成,我们可以把 Statuspage 放在那里,这实际上也是我们能够加速它们的方式,等等。每个人都很兴奋,认为这很有道理,有很多鼓励。我把它推销给了组织的每一个层级,从业务领导者,到我们当时的 CTO,再到 CEOs。没有人说不,没有人说是。几个月都是这样。
几个月来,我处于这种悬而未决的状态中,我觉得每个人对此都很兴奋,每个人都希望这发生,但它就是没有发生。所以我记得当时和我的老板谈话,问,到底怎么回事?我什么时候该停止推进?因为在某个时刻我确信我会开始惹人烦。然后他就像,“嗯,基本上就是当你对它失去热情的时候。继续推进,直到你觉得不再值得推进为止。”我记得看着这个建议然后想,哇,基于我当时所处的处境,这根本一点帮助都没有。但基本上在某个时刻我放弃了。然而后来发生了什么,而且我是在后来才明白的,因为公司恰好在一年后进入了那个领域,是我误判了对那个话题的意愿和紧迫感,以及 Atlassian 就是 Atlassian 这个事实,我们在那么多市场进行投资,我们有很多像这样搁置在架子上的机会。
有人做了分析,有人创建了商业案例,那东西说得通。可能没有一个关于“为什么是现在”的触发因素。所以我们需要一个非常强烈的“为什么是现在”的触发因素去着手处理它。而我没能很好地阐明这一点。为什么是现在?为什么我们必须现在做,而不是在一年后,两年后?而后来进来的下一个团队在那方面做得好得多。但对我来说,那是一个很好的学习,也就是出色的成果可以被搁置并且[听不清]。是的,这就是会发生的事情,因为我们有如此多的机会,而且外面有很多公司可能面临着这种情况,但这并不意味着我们必须去追求所有这些机会。但探索它们确实是有道理的,然后决定何时扣动扳机,而那种紧迫感的理由需要存在。它必须在商业层面上显得无懈可击。
**Lenny Rachitsky:**这让我想起了我和 Mihika 的聊天,她在 Figma 做类似的工作,主要负责很多从 0 到 1 的事情。她将这种工作描述为,如果你想让所有人接受一个新想法,你的工作就是保持这团火焰不灭,并帮助它蔓延到整个业务中。我很喜欢你在这里分享的这个非常具有战术性的建议,关于如何做到这一点,那就是弄清楚为什么这……我把它理解为,为什么它会过期?为什么这个机会会过期?为什么如果我们现在不采取行动,它就会消失?你可以把它看作是,为什么我们现在必须做这件事?这不仅仅是“这是一个巨大的机会”,而是我们还需要现在就做,以此来给人们动力。
**Tanguy Crusson:**是的,这对我来说是一个巨大的教训。我在上面浪费了几个月的时间,但就像每一次失败一样,教训是以痛苦的方式得出的。所以产品经理通常偏向于行动,或者至少我是这样的。我想去做事,去构建东西,去和客户一起尝试。因此,处于空闲状态、等待确认并不是一个好的处境,但有时你需要认识到自己正处于这种状态,并且退一步是好的。
从零到一的经验总结
**Lenny Rachitsky:**在我们进入产品发现(Product Discovery)之前,让我们试着总结一下你到目前为止分享的一些最大教训。关于如何在大公司成功构建从 0 到 1 的产品,我记下了几点。第一点是要非常清楚,你要为这个新产品构建的用户,是否真的是你已经在向其销售的用户?这感觉上可能他们足够接近,但在 HipChat 的案例中,你发现可能并非如此,并且这比你预期的要困难得多。
第二点是在重写时要谨慎。在某些情况下,关闭它,立即重写,在内部加速这个新想法。在其他情况下,正如你所分享的,如果你想了解在什么情况下选择哪种做法更合理,你可以倒回去听一下细节。有时重写并不合理,只需继续你正在做的事情,专注于用户问题,不要减速。我记下的另一个提示是忽略竞争,不要痴迷于他们在做什么,专注于你的用户要求你做什么。然后就是关注“为什么是现在”这个想法,这是一个非常好的观点,就像当你试图提出一个商业案例时,要弄清楚为什么它必须现在发生。在我试图总结这些建议时,你还有什么想到的吗?
**Tanguy Crusson:**不,这看起来总结得很好。我从中得出的主要一点是,如果你试图开始新的事情,这将来自于你的动力和你的热情,而这正是推动事物向前发展的动力。所以不要放弃。因为在得到一个可行的项目之前,我真的非常努力地尝试过。所以我想这也许证明了它是可能的。
**Lenny Rachitsky:**这也证明了你的毅力和让事情成功的渴望。所以在这个话题上,让我们谈谈你的产品发现(Product Discovery)。我知道这现在是一个成功的产品。我知道其中也伴随着很多痛苦,以及一些没有起作用的东西。所以我很想听听这两方面的经验,比如在让它落地时什么是最困难的,然后就是什么起作用了,是什么让你能够做成这件事?所以从你想开始的任何地方开始吧。
Point A 孵化器与 Jira Product Discovery 的诞生
**Tanguy Crusson:**是的,很酷。所以在我们回到心理治疗之前,让我们先从好的方面开始。这里的一些好消息是,Atlassian 在那时已经认识到,我们在那些大型成功产品上的创新都是通过收购来实现的,我们必须纠正这一点,并开始自己构建新产品。因此创始人方面有很大的推动力去说:“嘿,我们需要重新启动这一点。”由此产生了 Point A,这是一个旨在解决这个问题的内部孵化器项目。他们对此的表述是,创新就像一块肌肉。除非你锻炼它,否则它会变得虚弱,我们现在必须重新开始锻炼它。因此从那个项目中,Point A 和 Jira Product Discovery 是其中的……我认为它是通过那个创新项目的 100 个提案之一。有 100 个提案,最终产生了三个基本通过了所有不同阶段的产品。所以 Jira Product Discovery 的起步实际上是因为那个原因,也因为我们在 Atlassian 内部而成为可能。所以我们的起点是,我可以花时间专注于这些东西,而没有其他事情要做。因为有了这个孵化器,这是一份全职工作。因为分配了预算,我能够更容易地组建一个团队。我能够组建一个不担心失去工作的团队,因为该项目的设置在技术上讲,你是从其他部门借调人员的。如果那件事行不通,他们就回到你的部门。所以基本上没有失去工作的恐惧。
我们能够利用内部研究和洞察团队所做的所有研究。我们能够让企业发展团队与我们合作。在形成一个观点,即嘿,也许我们应该在那个领域发展之前,我大概会见了 20 家在产品管理周边领域运营的公司。所有这些团队都愿意与我们交谈,因为 Atlassian 是这个市场的重要参与者。所以总是有机会进行整合、被收购、合作之类的……整合、被收购、合作,诸如此类机会。我能够与分析师会面,问:“嘿,那么你们认为产品管理市场怎么样?”
**Lenny Rachitsky:**这些都是 Point A 结构的一部分吗?
**Tanguy Crusson:**Point A 结构嘛,并非所有内容都在 Point A 中正式确定,还记得 Point A 当时是与第一批通过它的赌注一起创建的,而我们的就是其中的一个赌注。所以我们为后来所有的项目铺平了道路。但基本上,它给了我们攀爬的支点去寻求每个人的帮助,每个人都知道这很重要。因为每个人都知道公司的优先事项,而新产品是公司的首要优先事项。所以玩长远游戏的 Atlassian 决定,可以投资这些赌注,并每三到六个月重新评估一次,以了解我们是否应该投入更多筹码。
因此,每个人工作的心理安全感都得到了保障,再加上能够获取公司的所有资源。那部分对于 JPD 今天的成功来说,简直是不可估量的。给你一点背景信息,Jira Product Discovery 四年前是从一些研究开始的。那时只有我一个人,后来变成了三个人。
第一行代码是在我们正式作为通用版本发布的前三年写下的。这意味着有三年的时间我们都在进行狗粮测试(dogfooding)alpha 或 beta 版本,并且能够在公司的全力支持下做到这一点。所以当我说玩长远游戏时,我是认真的。这在市面上大多数公司里是很难做到的。
内部孵化器的止损与定位
当人们问我如何停止内部孵化器时,就好比考虑一下什么时候能收回你的投资,然后暂时把它忘掉。你需要看到的是团队如何回答他们在沿途提出的正确问题,并在他们这样做的过程中看看你是否对这些赌注依然感到兴奋。无论如何,我们在一年前发布了。快进到今天,我们拥有了 8000 名客户,令人惊叹的 CSAT,良好的发展势头。
这是 Atlassian 历史上增长最快的产品之一,这很棒。但困难的部分在于,首先是提醒所有人失败是最可能的结果。我会死磕到底向人们解释,当他们想要在内部停止某些事情时,应该这样去构建框架,提醒每个人。有 70% 的概率,完全是我凭空捏造的,你正在做的任何事情在六个月内可能都不复存在了。
我们试图推出一款新产品,进入一个新市场。我们的目标是成为价值一亿美元的业务。所以这样的业务在外面并不多。我们尝试过,我们也失败过。其中有几个就属于 Atlassian。所以要记住这一点,记住这一点超级重要,否则公司有一种过度投资的倾向,不是公司自上而下的,而是公司的部分部门倾向于跑过来试图帮忙。
举个例子,“看,我们想构建这个。”“哦,如果你愿意,我们可以更改这个服务来为你做某事。”“不,我们是一个赌注,只有七个人。我们不要把公司其余的人拉进来。公司现在的胃口就是这七个人。我们倒要看看用这七个人能做出什么,”这就是我告诉每个人的话。我这么说的原因是,否则的话,是的,你得到了帮助,但这种帮助总是带有条件的,而条件通常是事情会变慢。
所以我们试图做的是提醒每个人,事情可能会失败,这样我们基本上就能买到一个机会去拼凑出一些乱七八糟的东西,那些无法扩展、不符合我们设计指南、不契合 JIRA 目标架构的东西。但我们要用这些去和客户测试,看看这些概念是否合理,这些原型是否合理,他们是否能从中获得价值,然后我们再告诉你,“嘿,你知道吗?那东西已经成型了,顺便说一句,现在我们是个正规业务了。我们应该把这个东西构建到平台里。”
**Lenny Rachitsky:**这真的很有趣,因为直觉上你会认为你应该把你的新赌注定位为,这极有可能会失败。这只是我们正在尝试的东西,别担心,现在还不要投入太多,不用担心给我们所有这些资源。你为什么认为这如此重要?因为我认为大多数公司不会这样定位新的孵化项目。你为什么认为这如此有效且更有效?
**Tanguy Crusson:**是的。所以我们在开始新产品时试图做的,基本上是在一个不饥饿、不挨饿的环境中模拟一家初创公司。因此你需要创造稀缺性。我希望我的团队能感受到紧迫感。那个东西需要推进。我也需要公司其余的人走开,这样我们才能获得自主权去测试那些我们需要知道这东西到底能不能行的事情。我们不能去参加一个为期六个月的规划会议,与平台服务协商某些事情,以便我们构建一个功能然后再拿去和用户测试。
我说,“现在,我们正在重建这个组件,下周我们就要和客户测试它。它不完美。它不完美。”所以那帮了我们大忙,否则的话总是有一种倾向,认为对其他所有东西都有效的流程对这个也会有效,我们确实需要不断提醒他们,“嘿,我们六个月后可能就不存在了。你现在真的那么在乎这个流程吗?这个产品可能都不复存在了。”通常流程就不存在了。
**Lenny Rachitsky:**所以这基本上是一个技巧,为了让组织内的其他所有人远离,并且不担心你在构建什么?因为你只是告诉他们,“别担心,这不会成功的。我们只是想尝试一下看看。”所以这并不是为了让你团队觉得,“这大概率不会成功。”我想象团队应该是这样的,“哦,我们必须让这东西行得通。这是个好主意。”这更多是一个防止组织把你吞没并对你指手画脚的技巧?
Point A 的四个阶段
**Tanguy Crusson:**所以有一部分是那样的,但确实也有一种需求……我们需要在这里尊重 Atlassian 的资金,如果我们不知道这东西到底能不能行……我不想把一个 50 人的团队拖进来。我想知道这东西是否值得一个 50 人团队的投资。所以实际上两者兼有。现在,这当然是说起来容易做起来难。所以这就是 Point A 再次发挥作用的地方。我们有四个阶段,叫做 Wonder、Explore、Make 和 Impact,在第一阶段,全部是为了证明有一个我们可以进入的问题领域,有一个市场。
我们可以非常清楚地回答,阐明为什么 Atlassian 应该进入那里。我们可以阐明为什么是现在,以及那些以前让我们挣扎的东西,并且有足够的数据来验证所有这些说法。Explore 是关于探索解决方案的,这并不意味着把它构建出来扔到那里看看什么能留下。它是关于如果你让一群客户提出一个问题,你能否让他们复述这个解决方案,解决那个问题?
所以在 Jira Product Discovery 的案例中,因为我们没有构建,我们不需要任何新技术,主要是新的用户体验和新的工作流,我们基本上在几十个 Zoom 会议中用 Figma 验证了其中的很多内容。但这基本上是回过头来说,“这些公司是如何构建这个框架的。这是他们的问题,这是这个东西将如何为他们解决问题。”所以那是 Explore 的一部分,也就是验证它在解决方案方面值得投资。我们不仅有正确的问题,我们还有正确的解决方案。
然后 Make 是关于分阶段让它发生,从 alpha 开始,然后是 beta,然后再推向市场。而 Impact 是那个东西实际上已经准备好出发了。现在让我们看看它对 Atlassian 业务产生的影响,并从那里继续监控它。
从那以后,它就转变成了真正的业务。但是 Atlassian 的每个人都知道这四个阶段:Wonder、Explore、Make、Impact。每当我们与团队交谈时,我们都会告诉他们我们目前处于 Explore。我们开始在整个 bet 本身上这样做。然后我们开始用这种方式来谈论我们正在开发的不同功能或我们想要解决的问题领域。这意味着现在每次我们与其他团队交谈并提到“我们在 Wonder”或“我们在 Explore”时,他们都知道能期待什么。当我们去找领导团队的某人并说“我们想从 Explore 进入 Make”时,他们知道他们将要[听不清],因为现在需要开发者了。
因此,为流程的每个阶段设定的所有这些词汇和明确的期望,确实帮助我们促进了与组织中所有人的所有对话。并且确实再次保护了我们免受那些基本上觉得他们必须插嘴的团队的影响。现在,我们在 Explore,我们没有任何已经验证过的准备好的东西。在我们验证客户想要某样东西之前,让我们不要对架构如何完成发表意见。
**Lenny Rachitsky:**这太酷了。我觉得我们可以专门做一期播客,聊聊 Point A 的结构以及你们是如何做到这一点的。但只有一个问题,当你从 Explore 转移到 Make,从 Make 转移到 Impact 时,关卡是什么样的?是否有一群人坐在一个房间里决定是赞成还是反对?这个决定是如何运作的?
评审关卡与决策机制
**Tanguy Crusson:**所以我们基本上会写一份六页纸的文档,审视我们要回答的所有问题的所有不同方面。然后我们与 Point A 的利益相关者和 Atlassian 的创始人开会。每个人阅读那几页大约 15 分钟,然后是提问、回答、评论,以及[听不清],在这个会议结束时,我们就知道我们是否被批准进入下一个阶段。有一次我们被退回去了,当时我们说,“嘿,作为 Make 的一部分,我们已经准备好从 alpha 进入 beta 了。”他们说,“不,你们还没有。”我说,“不,我们还没有。”所以我们留了下来,基本上获得了比最初分配给我们的更多时间。但基本上,创始人和 Point A 的领导层以及 Atlassian 不同业务线的负责人都参与了这些会议,这——
**Lenny Rachitsky:**那太酷了。
**Tanguy Crusson:**基本上,可见性一路直达顶层。
**Lenny Rachitsky:**他们也可能就在这些会议中决定,“让我们砍掉这个东西,它行不通。”
**Tanguy Crusson:**所以可能是“让我们砍掉这个东西”,这发生在我们做的许多 bet 上,或者可能是“让我们把这个东西并入其他东西中”。例如,我们的新白板产品。我说是产品,其实最初是 Confluence 的一部分的白板功能就是从 Point A 出来的。最终被并入 Confluence,因为[听不清]在那里更合理。
项目漏斗与高层参与度
**Lenny Rachitsky:**太有意思了。你说有 100 个项目经历了 Point A。所以存在一个漏斗。创始人是去参加所有这 100 个项目的会议,进行所有这些孵化,还是他们后来才在漏斗的下游加入?
**Tanguy Crusson:**不是全部 100 个。
**Lenny Rachitsky:**好的,很好。
**Tanguy Crusson:**我说 100 个,是在几个不同的季度中,团队不断进入和[听不清],但并不是在进入 Point A 的初始阶段。
**Lenny Rachitsky:**我明白了。
**Tanguy Crusson:**通常是在他们被接受之后。
**Lenny Rachitsky:**我打断了你,让我们偏离了主题,你刚才主要在分享哪些方面做得好,以及这一切是如何结合起来的。
成功经验与破局之道
**Tanguy Crusson:**因此,失败是最可能的结果。这是我会在所有事情中坚持的一点,因为我以前见过当我们对“它会成功”变得过于自满时会发生什么。这是之前我谈论的事情的教训。所以这一点非常关键。第二点,这一点要困难得多,也就是如果你开始做这样的事情,你的团队将需要打破许多既定的规则。但他们需要能够在不破坏组织中每个人信任的情况下做到这一点。这些规则是为了在业务成功的阶段支持业务而创建的,而恰巧它们可能不适用于新的 bet。所以对我来说,这里的技巧一直是……我想象的画面是,我有一堆筹码。自从我加入我的团队以来,我一直在积累筹码,而这些筹码是——
**Lenny Rachitsky:**像社会资本。
**Tanguy Crusson:**是的。这就像我与创始人、与不同的业务领导者建立的信任,通过在事情上取得成功,或者失败并解释原因并在下次尝试做得更好。所有这些给了我,就像你说的,资本。我有了这些筹码,并在这个 bet 上做出了决定,“你知道吗,我要全押了。”所以我总是说,“如果它成功了,那就成功了。如果不成功,我可能就会离开这里,但我要全押。”所以如果我看到行不通的东西,我就会说出来。我们不会去做。所以我知道我会让自己陷入艰难的对话中,因为人们在这里是为了保护需要保护的东西。它们只是对我在做的事情没有意义。所以举个例子,在不破坏信任的情况下打破规则,这是一个棘手的地方。
在像 Atlassian 这样的公司里我们有很多规则,工程部门就是这样运作的。我的意思是,工程师是我们员工中最大的部分。基本上,事情能做成是因为工程师在处理它。而我需要的是能够雇佣合适的团队,只有 principal 级别的人,那些拥有大量内部信用的人,这样他们就可以毫无疑问地提交到任何团队的 repo。那些不是在寻找下一次晋升,而是想要引起轰动的人。而当这不可能时,我希望能够雇佣承包商来填补空缺等等。我就像,“我需要打破的很多规则基本上都在工程部门。”所以我决定团队里不设工程负责人,由我自己来做。所以我是产品负责人和工程负责人。我的技术足够熟练,能够进行这些对话,但主要是我只是与那些能够自我驱动的优秀工程师一起工作。所以他们能够在不属于他们的领域进行更改。他们能够做出不符合我们任何标准的更改。他们能够通过变通的方式重建服务之类的东西。我们现在不需要做那样的事情了,但在那时我需要采取那样的立场,这样我们才能真正快速前进,基本上相当于成为 Atlassian 内部的一家初创公司。
做这样的事情并不舒服。我不会在许多环境中推荐这样做。Atlassian 在整个过程中非常宽容。这并不意味着我没有[听不清]那样做。那并不简单。例如,其中一条规则是,那时我们不希望在欧洲有更多工程团队的据点。今天情况不同了,但在那时就是那样。我就像,“糟糕。我驻扎在法国,我正试图开始做这些事情。我现在不想仅仅为了这个就搬到美国或搬回澳大利亚。”所以一开始我雇佣了承包商,很多承包商,或者说也不算很多,因为我们不是一个庞大的团队,而是雇佣承包商来构建这些东西。
**Tanguy Crusson:**再说一次,承包商并不受制于与其他工程人员相同的规则。所以基本上,所有关于你需要贡献于……的规则,比如,工程团队可能会说,“[听不清]到下一个阶段”,领导层可能会说,“你需要将 15% 的时间投入到可靠性上。”对我们来说就是,“我们还没到那个阶段。我们连原型都没拿出来。”是的。但这是公司规定,好吧。同样,没有工程经理加上承包商,砰,这些规则都不适用了。所以从外部看这件事的人就会觉得,“哇,伙计,你确定你能接受所有这些吗?”这让人感觉超级不舒服,但我们得到了领导层对此的支持。这只是 Atlassian 从“我们只投资大型收购目标”到“让我们投资于[听不清]”的一个艰难转变。
**Lenny Rachitsky:**你告诉我的这个故事太疯狂了。所以你在领导这个团队,你雇佣了一个承包商团队来构建这个产品。你基本上身处一个与 Atlassian 其他地方完全不同的国家。而这里整个想法就是去做那些在 Atlassian 不一定被允许的事情。他们不会让你这样工作的。而为了做出你需要做的东西,那个听起来像是你押上职业生涯的东西,你发现你只要在法国做一个这样干这事的“海盗”,而且这奏效了?
**Tanguy Crusson:**所以 Point A 这个表情符号在 Atlassian 是一个[听不清]。我不是唯一的一个。我们有相当多的人在致力于新的赌注,基本上都在那样运作。每个人都在质疑不同的规则。最终目标并不是质疑规则,最终目标是我们需要完成我们需要做的事情,也就是我们只需要清理这个空间,与用户合作测试原型,直到它起作用,并逐步达到,正如我所说,一个足以发布的产品。所以我们从未打算打破规则,而是我们要选择那些能够在这个使命上支持我们的规则,并对其他规则说不。
你提到了欧洲。当时,许多其他的 Point A 负责人都在为从他们位于悉尼的母公司基地运作而挣扎,因为他们仍然和其他所有人在一起。你在谈论做所有这些事情,但要记住我们有这个 OKR 的事情,我们有我们正在构建的 GR 愿景,你需要参与所有这些。所以我在欧洲就会说,“行,按你们的时间安排。”我想很多团队都会觉得,“我们对这个测试不够关心。他们告诉我们这东西六个月后就不存在了。我不够关心这件事,以至于不想每天熬夜或者超级早起去反驳他们。”所以我们早期取得的很多能够超级快速推进的成功,就是来自于距离如此之远,以至于人们根本就不会参与进来阻止我们。这就是为什么我们最初是速度最快的团队。然后,就有可能将这些事情制度化到 Point A 中,从而使其能够在 Atlassian 内部完成。但起初,我们只需要开辟出一条道路。所以我们就这么做了。
超级隔离团队的价值
**Lenny Rachitsky:**所以听起来,对于这样的事情,最大的经验之一和非常有效的方法就是建立一个超级孤立的团队,尽可能地将他们与核心业务断开连接,让他们只做你认为需要做的事情?
**Tanguy Crusson:**你之前在节目里采访过 Megan。
**Lenny Rachitsky:**是的。
**Tanguy Crusson:**我又重新听了一遍她推介 Point A 的方式,是的,那正是她所说的话得出的结论。也就是,我们为这些团队提供了空间,让他们能够以很大的自主权来做这件事。那基本上就是所有这些工作的成果,也就是随后建立了这个项目,让新团队可以这样做,并且基本上能够融入母公司中。
**Lenny Rachitsky:**我想人们听到这些会想,是的,好吧,当然。[听不清]孵化、孤立、分离,他们做自己的事情。人们听到了,也尝试那样做,但我认为这实际上并不是他们需要做的。而我听到的是,你为了实现这一目标走向了一个相当极端的地步,不基本上就是通过打破规则来让一个孤岛真正存在。我很喜欢这一点。让这一切成功运作的另一个关键是什么?
在成熟产品中安全地进行创新
**Tanguy Crusson:**我最充满热情的一点是,在一个超级早期的赌注中,我们与客户合作的方式与在像 Jira 这样非常成熟的产品中的做法是非常不同的。所以第一部分是,我们如何能够以一种不搞砸现有客户的方式进行创新?Jira 有 12 万客户,Atlassian 有 30 万客户。我们不能就这样冲进去,开始实验,弄坏数百万人体验到的一大堆东西,然后他们说,“你们在干什么,Atlassian?”所以我们需要创建这样一个区域,我们可以在其中进行实验,它远离 Jira 但同时又处于 Jira 内部。
那部分有点棘手。但这是可以做到的。事实上,我偶然看到了 Ben Weiss 的一篇文章。不,抱歉,是 Noah Weiss,这已经是很多年前的事了。他在里面基本上谈论了如何在成功的实体产品中进行创新,并谈到了孵化、迭代、整合这三个阶段。作为其中的一部分,他举了 Instagram、Twitter 和 Foursquare 的例子。他基本上在解释,比如在开始的时候,Instagram 是一个按时间顺序排列的信息流。然后团队开始在一个侧边的“流行”标签上进行实验,这个标签并没有集成到每个人使用的核心产品中。他们对此迭代了一段时间。那个标签变成了探索标签,然后探索标签变成了主信息流。信息流不再按时间顺序排列,而是基于你的兴趣。那部分真的让我产生了共鸣,我当时就想,“你知道吗?我们要尝试将此应用到 Jira 上,也就是我们将在 Jira 中构建它,但我们将把自己从核心底层的许多核心组件中抽离出来,以重建对我们追求的受众有效的 UX。”
我们的受众是产品经理。他们对加载图标没有耐心,事物必须是可视化的。他们需要能够快速移动东西,以可视化他们在路线图等方面拥有的潜在选项。它需要轻快,它需要不碍事,同时也要是他们可以自豪地展示给利益相关者并围绕其进行讨论的东西。它不能陷入细节中。它可以深入到非常、非常深[听不清]的地方。它需要是一个区域,一个让你感觉可以呼吸并进行创造性对话的空间。而 Jira 的 UI 并不完全以这一点闻名。事实上,我交谈过的大多数产品经理都会说,“我不想在 Jira 里做这个。它太死板了,有太多的工作流,它是由 IT 部门控制的”等等。所以我们说,“没关系。我们将尝试一种仍处于同一平台上但脱离了 Jira 的体验。”基本上我们就是这么做的。所以这种通过在网站上处理某件事来进行孵化、迭代直到做对、然后再将其整合回主产品的概念,是我绝对会再次采用的方法。我们目前正处于 Jira 产品发现的整合阶段。
**Lenny Rachitsky:**所以这里的经验是,起初不要强迫自己成为更广泛产品的一部分。首先,弄清楚如果它是一个独立的东西它可能会是什么样,然后最终你就可以将它整合?
**Tanguy Crusson:**这取决于你是为谁做以及他们拥有哪些项目。基本上,不要觉得受限于你的平台,以至于无法满足产品运作所需的核心条件。所以如果平台对我们来说足够好,如果用户体验对产品经理来说足够好,我们就不会那么做了。而且我们很难证明其合理性,但正因为事实并非如此,我们基本上是给了自己一个许可去做这件事。所以这仍然属于“不要搞砸现有客户”的范畴,这是我从一个网站上读到的东西,从 Reforge 那里读到的,它谈论了安全漏斗(Safety Funnel)。他们有这样一个概念,典型的增长漏斗从获取一直延伸到收入,中间有一系列不同的阶段。你的目标是最大化成功通过这个漏斗的人数。当时我正试图向人们解释我们想用 Jira 产品发现做什么,但我很难给它命名。我们的做法是,先与少数几个客户合作,直到它对他们来说是合适的,然后我们再逐渐向更多人开放。我知道这非常符合常理,但这基本上就是我们决定采用的策略,而不是 Atlassian 的惯常做法,后者更像是,“嘿,我们基于……来衡量成功。”例如,那时,我想应该是在所有那些项目之前,“……月活跃用户数”。如果那是衡量成功的标准,我们就会被非常快地推动去将它暴露给尽可能多的人。我们不想那么做,因为如果他们体验不佳,事后将很难把他们挽回。
所以 Reforge 里的某个人给它起名叫安全漏斗。安全漏斗这个概念太棒了。人们并不理解这一点。你基本上设置了一个硬性停止点,限制了产生糟糕体验的人数。你这样做一段时间,直到你能证明它非常棒,然后你邀请更多的人。所以我们在这方面做了很多。我们基本上在 Jira 之外创建了这个独立区域,重新构想 Jira 对于产品经理来说可能是什么样子。起初我们只向极少数客户开放了它。所以这就是“不要搞砸现有客户”的第一原则。
安全漏斗与业务风险
**Lenny Rachitsky:**这一点之所以重要,只是想强调一下,是因为如果你让很多客户感到不悦,业务方就会开始不满。
**Tanguy Crusson:**如果我们触发了一个导致数百万用户 Jira 宕机的事件,那就是这种下场了。第一次可能还没事。我的意思是,到第三次的话,我想我们就会被要求收拾东西走人了。
**Lenny Rachitsky:**明白了。所以那里的建议就是,限制接触你非常早期的产品的人数,即使这会损害你的数据。你不想让人们觉得,“你们到底在那边搞什么鬼?”
灯塔用户计划与衡量标准
**Tanguy Crusson:**所以损害你的数据现在成了有趣的一点。那么当你的目标基本上是在极长的时间内与少数客户合作时,你如何构建成功的框架,又如何定义指标?所以我最终为此定义的东西,是基于 Atlassian 内部使用的一个术语,我试图将其正式化为大家都能参考的东西,也就是灯塔用户计划(Lighthouse Users Program)。它的原则是它是一个计划,所以是一件正式的事情。我们在 Atlassian 有数十万客户,但我们只会为其中少数人构建体验。所以它是有阶段的。
第一阶段是我们与 10 个人合作,证明他们遇到的问题正是我们解决的问题,而对于最初的 10 个灯塔用户,我们花费大部分时间的地方是解释为什么他们是灯塔用户。这一切都是在解释,为什么我们相信他们是之后我们服务的所有客户的代表。然后我们进入了从 10 到 100 的阶段,在这个阶段我们招募更多客户,但我们仍然没有达到完全自助、不需要大量引导的程度。我们仍然在测试价值,但我们在测试人们面临的不同场景中的不同变化,以确保我们拥有的核心解决方案能够满足这些需求。因此有很多不同的选项和不同的安全需求。在 10 到 100 之间你会捕捉到很多细微之处。
然后我们到了一个阶段,我们会想,“你知道吗?它很好。它解决了人们的问题,但它不是自助的。我们只需要解释一些东西。我们需要在 Slack 上与人们合作。我们需要与他们一起工作,回答大量的支持工单,或者开大量的 Zoom 会议来解释事情。”现在我们需要从 100 达到 1,000,为了达到那个目标,我们基本上需要解决所有这些问题,然后我们就能毕业了。所以基本上通过详细说明这些,并解释中间所有的成功标准,它有助于在制造阶段定义成功,我们有 10、100、1,000,然后人们就能准确理解我们处于什么位置,以及我们试图回答的问题类型。在 10 的时候,我们通过与用户的一段视频通话片段来回答每一个问题,要么谈论他们的问题和解决方案,要么展示产品以及他们如何用它解决问题。
这就是我们希望利益相关者看待它的方式,它是非常定性的,但它是从用户的角度感受的。当我们从 10 到 100,从 100 到 1,000 时,情况是不同的,但对于每一个阶段,都有一个关于如何从一个阶段过渡到下一个阶段的操作指南。这帮助了我们,因为我们可以清理出空间来做正确的事,我们不会被要求去达到,例如,月活跃用户数、客户数或使用该功能的用户百分比等数字。因为当我们在那个阶段时,它仍然是非常定性的。
**Lenny Rachitsky:**这超级酷,也超级有趣。它符合关卡的概念。它与每个关卡中我们在寻找多少用户相一致。基本上,它为成功的定义设定了期望,让你不必过早地扩展规模。这是一个非常棒的想法。你叫它灯塔用户,比如灯塔用户计划?
灯塔用户计划的团队内部价值
**Tanguy Crusson:**灯塔用户,它有两个方面。我们刚刚谈论了面向利益相关者的一面。我更喜欢的是它面向团队的一面,我看到很多团队,在 Atlassian 和其他公司,随着产品的增长,你往往会更多地利用用户研究、正式的用户研究或客户满意度(CSAT)服务。我把它称为躲在研究的背后,没有足够贴近客户的实际情况。而我想在我工作的团队中做的,是与客户建立直接的反馈循环,但不是那种有人给你反馈你就去做的方式。我们与那些客户交谈,我们为他们构建东西。所以我看到了这样一个例子,就是气候变化。这是件大事。每个人都知道这是件大事。我们都读过关于它的报告。我们都会……每次我们读到 IPCC 的报告,我们都会想,“糟糕。”我们有改变什么吗?改变得不够。
是什么让人们想要改变,并真正给他们一个改变的触发点?我看到的是更多地与个人的经历产生共情。如果我认识一个受气候变化影响的人,这会让我与气候变化的概念产生更强烈的共鸣,并想要改变自己。抱歉在这里讲得有点沉重。我——
**Lenny Rachitsky:**不,这是一个非常好的例子。这是一种让那个观点变得非常清晰的好方法,也就是展示了与少数用户交谈的力量,相比于认为拥有越多的数据,查看数据、客户满意度分数、净推荐值(NPS)和留存率就能告诉你需要什么。
**Tanguy Crusson:**所以我们在那基础上所做的就是招募10个人,并把这些人的情况展示给整个团队,而不仅仅是产品经理、设计师和工程师。我们在Zoom上开会,聊天,在几个月的时间里与同一批人合作构建产品。他们也在Slack上和我们在一起,我们会有定期的交流。我们所看到的,我看到的能加快速度的情况是,工程师走进规划会议,产品经理会说:“所以我们应该做X”,然后工程师会说:“等一下。我们和这个客户谈过,他们在这方面很挣扎,所以我认为我们应该转而做那个,修复那部分体验”,诸如此类。
突然之间,你谈论的不再是一个产品经理,而是一个拥有产品工程师的产品团队。在Atlassian有这样一种说法,我们称一些工程师更像系统工程师,而另一些更像是产品工程师。在我看来,每个人都可以成为产品工程师。他们只需要被置于正确的用户上下文中。正确的用户上下文是什么?就是你叫得出名字的10个客户,你了解他们的上下文,了解他们的问题。你能对此产生共情。这种共情让你想要采取行动去改变你的产品以解决他们的问题,并且你会从中获得巨大的自豪感。
在像Atlassian这样的公司里,这看起来可能有点违反直觉。30万客户,对吧?我们应该为他们构建。好吧,你做不到,对吧?我们让改变适用于绝大多数人的能力是有限的,所以不如坦然接受这一点。接受我们确实存在偏见,接受我们确实是在基于感受做出反应,比如想帮忙或不想帮忙,或者对你所看到的产生强烈反应,但我们是在拥抱这一点,以尝试构建出最好的产品。实际上我看到的成果是,到目前为止在这个产品上它奏效了。
**Lenny Rachitsky:**如果你真的从初创公司会如何处理这个问题的角度来看待这件事?这正是刚刚起步的公司会采用的方式。所以这样想非常有道理。只是人们实际上并不会这么做。这很难做到。
度过“丑婴儿”阶段
**Lenny Rachitsky:**这正好引出我想问的一个问题。所以你们可能花了很长时间才展示出真正的成功、真正的进展、那种真正“这会行得通”的感觉。首先,那段“这可能行不通。哦哇,也许这能行得通”的时期有多长?顺着同样的思路,Atlassian 或者……你从如何保护这个东西上学到了什么,皮克斯的人称之为“丑婴儿(ugly baby)”?当一个想法刚诞生时,他们称之为丑婴儿。人们不想要它。就像,“把这东西弄走”。天哪,这么说听起来很刻薄,但这就是他们讨论它的方式。这是创造力领域的一个术语。
**Tanguy Crusson:**这个说法很好。
**Lenny Rachitsky:**你是如何保护它的,因为我认为很多公司面临的最大挑战就是,“已经六个月了。没人想要这个。我们要把它砍掉。”你是如何保护它的?所以它花了多长时间才显示出成功,关于如何保护它你学到了什么?
**Tanguy Crusson:**是的。所以基本上,内部沟通在那里就是一切。我把我的工作看作,很大一部分是在我们处于什么阶段、我们在学习什么、我们学习有多快、我们行动有多快这些方面保持非常清晰。在每一步都保持超级诚实,不试图胡编乱造,不试图把它夸大,不试图让它看起来比实际更成功。相反,要非常清楚我们在测试什么,在哪里测试。用数据来做,用个人的客户故事来做,因为我们都能不断与心灵和思想产生共鸣。所以这就是我总是试图编织进我随后每周发送的沟通内容中的东西。
所以我们有一款内部叫Atlas的产品,在Atlas上有一个项目。人们可以订阅这个项目,然后你每周发送一条推文大小的关于你项目的更新。我把它作为一个平台,在内部就这个产品进行沟通,实际上最终有几百个人订阅了这东西。每周,当我们处于试图将第一个版本交付给客户的阶段时,这是一次关于产品和我们所构建的所有东西的每周演示。尽可能多地展示势头。我甚至会把功能从一个星期留到下一个星期,只是为了……例如,当我知道我们要去度假时,只是为了有东西可以展示,展示这列不断前进的火车,它一直在前进。你不想挡在它前面。
但接下来我需要做的是,将这一切与所有客户方面的事情平衡起来。所以当我们开始把它展示给客户时,那就是客户对话的片段。没人会去读一份需要花30分钟读的研究报告。每个人都乐于观看一个三分钟的片段,里面有四个客户在谈论某件事。所以我每周都在大量使用这个方法,把它们发布出去,然后在Slack上广泛分享,说:“这是我们的发现。这个客户在Jira方面面临着这样的问题。我们是如何解决的。这是他们在谈论这件事,”诸如此类的东西。
所以我在那里发布这些,然后把它发到Slack的各个地方。这类东西就是给人们一种速度感和快节奏感,而且没人想去阻挡一辆高速行驶的火车。你不会挡在它前面。那无济于事,而且你这样做会显得很糟糕。所以基本上,这最终为我们买到了我们需要的时间,让我们走出这个,我不知道你怎么称呼它,丑婴儿阶段。
**Lenny Rachitsky:**丑婴儿阶段。
**Tanguy Crusson:**所以基本上它帮助我们走出了这个我们还拿不出任何成果的阶段,因为我们所展示的是基于我们试图获得的认知,而我们试图用“这里有一些数据,这里有一些演示,这里有一些用户片段”来做到这一点。每周都这样做,因为人们每周都能消化小块的内容。他们只是对那种三四个月后你才走进来说“这就是我们要展示的成果”的东西感到吃力。
**Lenny Rachitsky:**对。对你们来说,走出这个丑婴儿阶段之前那段时间有多长?人们什么时候开始觉得,“哦哇,这可能行得通”?
**Tanguy Crusson:**老实说,我们在两个月内就有东西在进行狗粮测试(dogfooding),在五个月内交到了第一个灯塔用户手里。我们在alpha版本上迭代了大概六个月。然后我们进入了一个持续近一年的beta阶段。所以我认为在每一步中,人们都能看到我们正在走向某个地方,而且我认为最初他们评判的,不是结果。他们评判团队的成分远远多于评判结果。
所以有那么多关于创始人市场契合度(founder market fit)的讨论,我认为这真的是你需要去梳理出来的东西,即,这是否是正确的事情,是否在追求正确的问题?然后,如果他们能够以你没想到的方式,连续几周回答“是的”,那就足够了。我认为那就是我们直到能够把东西交给客户之前的阶段,但是第一个灯塔用户是在五个月时介入的。所以它看起来好不好甚至都不重要了,因为人们可以看到,“好的,那就是……哦,那里有点东西,”对吧?
所以我不确定是否有一个确切的时刻发生了这种转变。然而,确实有这样一个时刻,我们觉得,“好了,就是这样。我们准备好正式发布了(GA)。”一位创始人 Mike 说,“不,你们还没准备好。那东西太丑了。我都不想看它。它需要提升到符合 Atlassian 其他设计标准的水平。我们的客户在这方面对我们有期望。你们的东西是可用的,但现在发布太早了。”所以我们去修改了它。花了我们两三个月,然后我们才准备好出发。但是,是的,因为我们有 Point A,因为我们设定了到达那里需要一段时间的预期,而且因为我们能够每周展示进展,我们基本上没有真正感受到那种因为太早而不知道该拿它怎么办的时刻。
**Lenny Rachitsky:**这个设计改进阶段,我脑海中的画面是你这个海盗被邀请去参加一个精致的晚宴,你不得不开始收拾打扮,让自己看起来体面,融入主流社会。这非常有道理。同时,你描述的这些让我再次想到这样一个概念,即你的工作是保持火种不灭,并帮助它在组织内部蔓延。所以你有这个想法,你就像是在保持势头,通过不断分享更新、分享进展,用小片段和视频让它非常容易被消化,从而让这团火越烧越旺。这里有很多很好的建议。那么从时间线来看,有趣的是我听到的是,从想法到 alpha 基本上花了一年左右,差不多是这样吗?
**Tanguy Crusson:**差不多七个月……不,五个月交到第一个 alpha1 客户手里。
**Lenny Rachitsky:**好的。
**Tanguy Crusson:**我们在 alpha 阶段停留了六个月。
**Lenny Rachitsky:**好的,明白了。所以度过 alpha 阶段总共大约一年。很多人听到这些。一方面他们会觉得,“当然,Atlassian 有这么多资源。他们当然愿意在一个不知道能不能成功的想法上花一年时间。这对他们来说太容易了。”我想象在这一路走来肯定有很多痛苦、折磨和挑战,使得事情并没有那么容易。关于真正把这件事做实所经历的挣扎,你有什么可以分享的吗,就当作是我们结束对话前重回一次心理治疗,如果你有什么想分享的话?
**Tanguy Crusson:**是的。基本上,这很大程度上就是我们之前谈到过的,有很多流程还没有改变,因为我们正处于 Point A 的最开始阶段,而人们仍然希望我们参与到对组织他们那部分很重要的事情中,而如果我们将来的项目成功了,我们之后会在其中发挥作用。所以这些东西都没有一劳永逸地解决。这只是一个不断与之周旋的持续过程,但我认为我们最终设法相当优雅地从这些事情中脱身了,我认为我们因此赢得了某些团队一定程度的钦佩,有些人觉得这有点……看到有人打破规则还能全身而退总是很有趣的。我认为这启发了一些其他创始人去尝试这样做,看到这一点非常酷。实际上,这是我希望能看到更多的事情,从我正在交谈的关于他们公司内部情况的客户那里得到的反馈来看,而不是被当前的流程扼杀,并且他们觉得无法打破枷锁。所以我真的希望它能启发更多的团队去试一试。
**Lenny Rachitsky:**关于 Point A 以及其他公司如何做这种孵化工作,显然还有更多我想了解的。所以这是深入挖掘这类项目的一个非常好的灵感。我们涵盖了很多内容。我们谈论了有效和无效的事情,各种心理治疗和痛苦,但也谈到了帮助你成功并构建出色产品的事情。在我们进入非常激动人心的闪电问答之前,你还有什么想分享或想留给听众的吗?
平衡推动改变与自我保护
**Tanguy Crusson:**是的。我想平衡一下我刚才提出的观点,即这很难,所以必须有人去尝试。要始终觉得你可以去推动。我想稍微平衡一下这一点,那就是也要注意保护自己,并确保你是在一个准备好接纳它的环境中做这件事。正如我所提到的,我在 Atlassian 工作了过去 10 年。在此之前,我曾在银行、受到高度监管的行业以及一大堆其他领域工作过,在那些地方这种事情是不被允许的。我本可以尝试去推动,但那绝对会进展得很糟糕。所以我深信,我们不需要那么多的自上而下的领导。我们需要的是许多自主的领导者,无论他们在组织中的职位如何,都能推动变革,为正确的事情而战,但你必须在一个对你来说安全的环境中这样做。如果不是,我认为考虑其他选择是可以的。换句话说,不要觉得你被困住了。对吧?你只有一次职业生涯。要在你真正相信可以做出出色工作的地方工作,并在让你感到兴奋的人包围下工作。否则,我自己也曾经历过那样的时候,在不可能做到的事情的重压下,可能会变得愤世嫉俗,并最终开始怀疑自己做事的能力。所以这有点像是在为正确的事情推动和保护自己之间寻找平衡,如果环境不对,改变一下也是可以的。
**Lenny Rachitsky:**这是一个在产品管理中非常热门的话题和常见问题,我想象在产品的其他职能中也是如此,那就是作为公司的一名独立贡献者(IC),到底能改变多少。我听到了所有这些关于做出改变、改变文化、孵化东西、创新的建议。你实际上能产生多大的影响,对比“什么都不会改变,我应该去别处工作”?你在那方面有什么建议吗?听起来你基本上是在说,很多时候你对业务和文化运作方式没有影响,你可能应该去找一家像 Atlassian 这样确实懂得如何孵化、创新和用不同方式思考的公司。
**Tanguy Crusson:**是的。这个是一个……我真的不确定该怎么回答,因为我共事过、咨询过、为其工作过的公司接近 50 家。Atlassian 是我加入的第一家从第一天起就让我有归属感的公司,以正确的方式受到挑战,但是妈的,我真的可以选择我的战斗并去追求它们。事情很难,但没关系。所以我加入了 Atlassian。[听不清] 就那样找到了蹦床,但是是的,我只想说,不要安于现状,因为你不能成为房间里唯一清醒的人。在某个时刻你会发疯的。环境会渗透到你身上。你不是一个对周围发生的一切绝对渗透不进的实体。你周围的一切都会影响你,并改变你这个人。我真的相信这一点。所以你需要让自己被那些能帮助你激发出最好一面的人和环绕,否则你可能会变坏。我自己过去就曾在愤世嫉俗的环境中变得愤世嫉俗,然后我决定那不是我。我不想让它成为我。那时候那是我。我不想让它成为我。所以是的,我唯一的建议就是,要有勇气问自己这些问题,否则它可能会以你不想看到的方式改变你。
闪电问答
**Lenny Rachitsky:**令人惊叹的建议。非常重要的建议。说到这里,我们已经到了非常激动人心的闪电问答环节。你准备好了吗?
**Tanguy Crusson:**我准备好了,很期待。
**Lenny Rachitsky:**好的,我们开始。第一个问题,你向其他人推荐最多的两三本书是什么?
推荐书籍
**Tanguy Crusson:**是的,第一本是 Atlassian 联合创始人 Scott Farquhar 推荐给我的。《Who: A Method for Hiring》基本上是关于如何做一件我曾一度非常不擅长的事的书,那就是通过面试去了解谁是一个好人,而不是一个适合加入你团队的好人。第二本,那是一本关于工作的书。另外两本则与工作无关。强烈推荐阅读《Hakim’s Odyssey》。这是一个叙利亚难民离开叙利亚,试图进入欧洲的故事。我们听过很多关于人们试图穿越地中海或在其他国家基本上成为难民的统计数据和数字,在你遇到一个个人故事之前,很容易自满地看待这些事。这位插画家非常优美地讲述的 Hakim 的个人故事,在这方面非常值得一读。最后一本是一本我不确定是否已从法语翻译过来的书。它叫《Vivre avec la Terre》(与地球共生),讲的是我们如何能够建立一个不同于今天的农业系统,可能有更好的方式在未来养活我们,而不需要大规模的单一种植蔬菜,那可能不那么可持续。他们在非常小的地块上设法创造了一个真正高效的结构,许多物种协同工作来控制并不需要使用任何杀虫剂之类的东西。他们周围总是有研究团队,我想叫 Inria。它是法国的研究中心之一,已经取得了非常惊人的结果,但我认为没有多少人知道这些东西的存在,也没有多少人知道它实际上可以在外面使用,我觉得这很遗憾。
**Lenny Rachitsky:**我想这是第一次听起来推荐了一本只有其他语言版本,可能还没被翻译过来的书。
**Tanguy Crusson:**它可能被翻译了。我希望如此。不过它相当厚,而且非常技术化。所以我把原理部分读了一遍,说实话,我会在所有种植食物的技术细节中迷失。
面试技巧
**Lenny Rachitsky:**好的,很喜欢。那么问问你最喜欢的面试问题。你最喜欢的面试问题是什么?
**Tanguy Crusson:**是的。所以我在面试上挣扎了很久,我读过外面所有标准的面试问题,我也在你的播客里听过。有一个问题来自这本书,Scott Farquhar 也告诉我这个问题对他非常有效,那就是当人们描述一段经历时,你问他们当时一起工作的人的名字。然后你问他们:“那么当我在我们通话后给这个人打电话时,你觉得他们会怎么说?”显然这会产生某种效果,我也见过这种情况,人们无法从另一个谈论他们的人的视角出发,当场投射和编造一些东西。所以他们也许能够谈论,“我做了这个,我做了那个,我做了那个。所以当我说……”“你的老板是谁?我的老板是……所以当我问这个人,你觉得他们会怎么说?”就像,“啊,好吧,其实我只领导了其中一部分,也许你不应该给他们打电话,因为某某某原因。”那一部分让人们突然脱离了他们给自己设定的脚本版本,在一瞬间变得更真实。你会得到一点真实性的流露,这对我来说在面试问题中通常很难做到。我见过这种情况发生。我现在觉得这样做很有趣,因为人们会有一瞬间变得非常不自在,然后你就会得到真实的东西。
**Lenny Rachitsky:**这太绝了。非常有道理。这让我想知道如何能把这个技巧整合到我的播客访谈中。非常好的建议。下一个问题,你最近有没有发现什么你非常喜欢的心爱产品?
最近的心爱产品
**Tanguy Crusson:**有,它跟科技毫无关系,跟科技完全无关。我最近经常玩风筝冲浪,我接触到了水翼板,基本上就是这些东西,在连接着板的桅杆下面有一个机翼。你带着风筝站在板上,你基本上就是在水面上飞行。这项发明简直太棒了。它相对容易学,我基本上可以花周末的一部分时间在水面上飞行,哪怕只有微风吹过,因为板和水之间绝对零摩擦力。所以,从技术上讲,我认为他们创造它的方式并从飞机中汲取灵感,是天才之作。
**Lenny Rachitsky:**就是那个 Zuck 涂着防晒霜带着旗子骑的那个东西吗,还是别的什么?我不知道你有没有看到那张照片。别管它了。
**Tanguy Crusson:**不,我没看到。
**Lenny Rachitsky:**我很喜欢你虽然说有个风筝在水里拉着你,而你在水面上这个东西上,但你说它相对容易学。我不知道该不该信你。
**Tanguy Crusson:**是的。我的意思是,有些运动和技巧很难学。这个你可以在 18 个月到两年内变得熟练,这真的非常短。它可能是一项你以前做过的运动。
**Lenny Rachitsky:**才几年而已。好的。我很喜欢你的标准……我本来想顺着这个方向问一个问题,但在那之前,你有没有什么最喜欢的人生座右铭,是你经常回想起来的,发现在工作或生活中很有用,也许还会和亲友分享的?
人生座右铭
**Tanguy Crusson:**我有两句,一句关于工作,一句关于生活中的一切。关于工作的那句最初是奥巴马的一句话。我不记得原话了。我会转述一下,但让我们把焦点放在工作上。所以在你的职业生涯中会有这样一些时刻,你觉得自己没有受到重视,你会想:“我做对了吗?我被认可了吗?我被重视了吗?我在对的地方吗?”每当这种事发生时,很多寄生想法可能会涌入。你可能会面临冒名顶替综合征之类的情况。我经常遇到这种情况。每当这种情况发生时,我会提醒自己,“把焦点放在工作上”,因为只要你把它放在工作上,总会有工作要做,而且总会从那项工作中浮现出一条道路。这有点像你在瑜伽里遇到的那种情况,一旦你全心投入,好事就会发生。嗯,这正是我所看到的,而且每次都对我有帮助。每当我处于周围动荡不安的时刻,我就会重读这句话。现在当我面对那些时刻时,我也会提醒自己另一句,这是我发明的,但我不认为我是唯一发明这句话的人,那就是:“记住,100年后我们都会死去并被遗忘。所以不要太把自己当回事。你没那么重要。你对于和你争吵的人来说,可能并不像你想象的那么重要。”这也是所有事物的美妙之处,那就是它都有尽头。到头来可能都不重要,所以不妨全力以赴。这也是许多代以前的法国存在主义者们在谈论的事情。这个观点接近阿尔贝·加缪的论点,对吧?“如果没有意义,不妨放手一搏。”
自由潜水经历
**Lenny Rachitsky:**非常好的观点,非常好的教训,非常好的座右铭。最后一个问题,你曾在一种自由潜水项目中排名世界第四。首先,你能简单描述一下什么是自由潜水吗,然后你能分享一件可能会让人们对这项运动和技能感到惊讶的事情吗?
**Tanguy Crusson:**是的。没错,那就是我出名的原因。我基本上是在无脚蹼游泳池潜水距离这个项目上排名第四,游了167米,也就是550英尺。
**Lenny Rachitsky:**是在水下吗?
**Tanguy Crusson:**水下蛙泳。我用单蹼游得更远,就是你后面看到的那种,但蛙泳离法国纪录只差一米,那一年的世界排名第四。我最骄傲的一次是我实际潜到了水下300英尺深,安然无恙地回来了,而且我非常享受整个过程。
**Lenny Rachitsky:**300英尺。那就像是——
**Tanguy Crusson:**92米。
**Lenny Rachitsky:**天哪。我试图想象一下和自由女神像之类的相比那是什么感觉,但我稍后再查一下。
**Tanguy Crusson:**我通常会看建筑物作参考。建筑物可能只有二三十米高,也可以超级高,但90米确实相当高了,是的。上次我看着那样的一栋建筑,我不记得是哪栋了,但那是相当高的。所以过程就是吸一口气,下潜,摸到绳子底部的底板,然后再上来。关于这件事可能会让人惊讶的一点是,每个人在这方面的天赋都比他们想象的要高得多。所以我以前周末会教自由潜水课程,当我问大多数人:“嘿,你觉得你能在水下待多久?”他们通常会说30秒左右,也许一分钟。“你觉得你能潜多深?啊,大概五米吧。”到了周末结束时,大多数人都能憋气两到三分钟,并能潜到20米深。所以这就是那种看起来绝对令人印象深刻、疯狂且不可思议的事情,而实际上我们天生就有这方面的天赋。当我们潜水时会发生很多生理变化使其成为可能,但我认为我们对身体了解之少这一点非常迷人。
**Lenny Rachitsky:**我正在 perplexity 上查300英尺相当于现实生活中的什么物体。那么300英尺相当于现实生活中的什么物体呢?来了,300英尺。一个足球场的长度。好吧,是的。我本该想到的。接近一架波音737的长度。
**Tanguy Crusson:**是的。你需要这样看才能有点概念。
**Lenny Rachitsky:**是的,一个足球场——
**Tanguy Crusson:**你要潜到水下那么深——
Lenny Rachitsky:……我的天哪。哦,我的上帝。但我很喜欢你的建议,即人们在这项技能上的表现可以比他们想象的要好得多。
**Tanguy Crusson:**是的。从小孩到成年人,其实都非常擅长这个。
结束语与联系方式
**Lenny Rachitsky:**Tanguy,我们聊了很多。我认为这会帮助很多试图在自由潜水方面变得更好的人,也会帮助那些在大公司内部从零到一做产品的人。非常感谢你抽出时间,并分享了这些真实的对话,就像你所说的,疗法。最后两个问题。如果人们想联系你,或者想就你谈到的一些内容进行跟进,他们可以在网上哪里找到你?听众怎样才能帮到你?
**Tanguy Crusson:**是的,如果你知道怎么……你可以在 LinkedIn 上找到我。你不需要知道怎么发音我的名字,你只需要知道怎么写就行。说实话,我在社交媒体上不是很活跃。我深深地扎根于构建这个产品之中,所以公平地说,我在拓展人脉方面做得不多,但是例如,如果你正在使用这个产品,并且你有关于我们如何为你改进它的想法,或者你想分享一些关于它是如何帮助你的、或者没有帮助到你的故事,我很乐意与你联系,因为我每天花很大一部分时间与用户交谈、回复支持工单之类的事情。所以,是的。
**Lenny Rachitsky:**太棒了。Tanguy,非常感谢你的到来。
**Tanguy Crusson:**谢谢你,Lenny。说实话,这太棒了,我希望我的一些胡言乱语对人们有用。
**Lenny Rachitsky:**绝对不是胡言乱语。我认为这会帮助到很多人。那么,我就放你走了。大家再见。
**Tanguy Crusson:**再见。
播客结语
**Lenny Rachitsky:**非常感谢大家的收听。如果你觉得这很有价值,你可以在 Apple Podcasts、Spotify 或你最喜欢的播客应用上订阅本节目。另外,请考虑给我们打分或留下评论,因为这真的能帮助其他听众找到这个播客。你可以在 lennyspodcast.com 找到所有往期节目或了解更多关于本节目的信息。下期再见。
术语表
| 原文 | 中文 |
|---|---|
| acqui-hire | acqui-hire |
| Atlassian | Atlassian |
| Ben Weiss | Ben Weiss |
| competitive myopia | 竞争短视(competitive myopia) |
| CSAT | 客户满意度(CSAT) |
| dogfooding | 狗粮测试(dogfooding) |
| don’t eat your own bullshit | 不要吃你自己的废话(don’t eat your own bullshit) |
| eat your own dog food | 吃你自己的狗粮(eat your own dog food) |
| founder market fit | 创始人市场契合度(founder market fit) |
| frankenstack | frankenstack |
| GA | 正式发布(GA) |
| GTM | 走向市场(GTM) |
| Hakim | Hakim |
| IC | 独立贡献者(IC) |
| Inria | Inria |
| Jira service management | Jira service management |
| Lenny Rachitsky | Lenny Rachitsky |
| Lighthouse Users Program | 灯塔用户计划(Lighthouse Users Program) |
| Megan | Megan |
| Mihika | Mihika |
| Mike | Mike |
| Noah Weiss | Noah Weiss |
| NPS | 净推荐值(NPS) |
| open company, no bullshit | 开放公司,没有废话(open company, no bullshit) |
| Point A | Point A |
| portfolio fit | 产品组合契合度(portfolio fit) |
| Product Discovery | 产品发现 |
| product market fit | 产品市场契合度(product market fit) |
| Safety Funnel | 安全漏斗(Safety Funnel) |
| Scott Farquhar | Scott Farquhar |
| squad | squad(小队) |
| Tanguy Crusson | Tanguy Crusson |
| tech-in | tech-in |
| Zuck | Zuck |
此文档由 AI 分片翻译(translate_long_document)