Kickstarting a Community on Top of Product: James from PostHog

In this podcast series, I'll be talking to a few community builders to learn more about their journey and hopefully help others grow their own communities.

Podcast - Kickstarting a community on top of your product
Podcast - Kickstarting a community on top of your product

Welcome to our podcast series on how to kick start a community on top of a product.

Every week or so, we explore the long journey of community building by talking to founders and community builders.

I'm very excited to chat with James Hawkins. He's the co-founder and CEO of PostHog, an open-source product analytics platform.

They've been through YC in the winter of 2020 and had exponential growth since then. Obviously, community building has been part of their DNA since day one, as an open-source product, and I'm so interested to learn more about it.

Enjoy reading the interview or listen to our original conversation.

👉 How, as an open-source project, the community was part of their DNA since Day 1
👉 Why Slack is the right platform to go from 0 to 1
👉 What are developers and product teams looking for in the PostHog community
👉 The biggest mistake they made at the beginning
👉 How do they measure the health of the community
👉 Why community is your moat

Louise: So first, can you tell us a little bit more about yourself, what are you doing and how did you end up building a community?

James: I'm the CEO and co-founder of PostHog. We're an open-source platform for product analytics. We help you measure your performance and your products. We help you diagnose why your numbers are good or bad. And we help you release changes, and run experiments to kind of iterate and improve.

Then from the community perspective, we started off with an open-source project. We spent the first year and a half of our company life really focusing on our free, open-source thing.

We're now at the point where we have about 30,000 users using the product or in the community itself. About 10,000 companies have also adopted PostHog at this point.

🧬 Community as part of DNA since Day 1

Louise: You've started the community right away when starting your products? Was it at a similar time that you started to get creators?

James: Yeah, absolutely, we realized within a week of having the idea. Originally, at PostHog, we wanted to do product analytics that you could self-host. We have realized, actually, that open-source is the strategy for having a large community that will be a massive part of how we grow.

It's been quite intentional that we've enabled a large community to be around our product.

And later on, like a year and a half later now, it's wonderful. All of our sales are inbound, all of our growth is completely word of mouth driven, and it lets you kind of double down on community rather than Google ads, for example.

Louise: How should early state startup founders assess whether this is a good move for them or not to kickstart a community?


So I think a really basic question to ask yourself before you get started is, does this make sense for your end users?

We're selling a product to a very technical audience. To put it really bluntly, we didn't think that developers would want to be sold to and that's who we're building for. We anticipated whether our growth and reputation would be more important.

And secondly, it felt it'd be a really good fit. It's a well-trodden path in open source projects to build a large community and then kind of get a lot of great off the back of that, looking at GitLab, HashiCorp, Mattermost.

Louise: Talking about the end-users, how did you get the first people to join your community, and who was responsible for taking care of them? How did you structure your company around this?

James: It was very important to be proactive with encouraging people to join. We were open source from the start so it was always possible to contribute code. We had things like contribution guidelines laid out.

One of the things we've really learned is, that if you make it simple to know how to take part in the community, you've got have much more chance of success.

For example, we have GitHub issues where we plan out things, like features that you want to build for the platform. If we explain in more detail what we want to do: if we were just like, build a feature that lets you do X, it's kind of vague and high level. No one will take part and build it because they're not sure if their work is going to get accepted or not.

But if you make it more granular and you're clear on what you're looking for, we have almost 100% uptake of good first issues that are well-scoped.

So being really clear about how people can take part and kind of guiding them has been one of the less obvious lessons that were taken from the early days.

🧑‍💻 The platform to take the community from 0 to 1

Louise: On which platform do your community members interact?

James: From the day we launched, we had Slack. We created this community group, and we didn't advertise it particularly hard. We just mentioned it in our project's "Read me".

Today we have 2500 to 2700-ish people in our Slack group.

Slack's been excellent for getting responsiveness from people. To give you a statistic, if we email a user versus we drop them a message on Slack, we have probably 20 times the response rate on Slack because people are going to open it for work.

As we're getting bigger, now we got a couple of thousand people in there. I don't think it's the best place for everyone to be.

It's great when we're trying to talk to large enterprise customers to act as a single team with them, but we're starting to build out some tooling to enable people to ask and answer questions directly. So we're starting to move the community on a little bit from Slack as a platform.

But for getting from 0 to 1, it was absolute, the right choice for us.

Louise: In the Slack community, the people who are there, are they only developers, or do you also have, potentially end users, like product teams? Who are the people in your community?

James: Developers principally, we're probably 70% very technical people. We got a couple of product managers as well.

You get different questions from people. Developers will generally ask us about the setup, deployment, implementation of the products, and how to contribute.

Product teams ask questions about how to use the product once it's already running, so quite different styles of questions come from different people, but we have a really wide range.

🔎 What is the community looking for

Louise: What are people looking for in your community? Because I guess, developers are looking for stuff different than for a team. How have you decided to gather them all at the same place? Or do you think it would make sense to have a community specifically for developers and a community for more product people?

James: I would heavily push to make sure everyone's in the same place because you need enough saturation that the community is vibrant, and it doesn't feel like a ghost town.

What we're trying to teach users is that if they ask a question or they interact in the community, there are people there. No one has shown up to an empty party.

I just say, lean towards, just make sure the community appears, even if it's quite mixed in terms of the profile of the person that might be joining, and just be really open.

I think what we've learned in terms of what people are actually looking for there, is something we've made some mistakes with.

It's support. Product support has been a huge part of that. That's the overwhelming majority of interactions is people asking questions.

One of the things I think I would have done differently, is, what we're now doing but it's something we would've done from early on, is encourage people to answer each other's questions.

💡 The biggest mistake at the beginning

Louise: But how can you do that? How do you think we can kickstart this kind of initiative, having people chatting with one another directly within your community? That's the most difficult thing I think.

James: It's very hard. One of the things that we learned was, initially, you have 1000 people in the community but a really small handful that you could go very deep with. So the mistake we made was we created this kind of paradigm where our engineering team would just answer all of the questions almost immediately as they were coming up.

And I kind of told people, okay, if I just ask a question, some random person will most likely answer in like two minutes, which doesn't leave space for other people to ask.

It's why we're starting to build tooling around questions and answers. We're about to open source a cool product to do this.

The thing we looked at are, okay, let's get deeper relationships with the people that are most heavily engaged, they're using our products intensively. They're submitting code and pull requests and so on. And let's just say, hey, you want to take a stab at this one?

Just to demonstrate to other people in the community that that's how this community works, where people answer each other's questions.

That is still a major work in progress for us even at this point.

But I think that's kind of like the second step.

Step one, you just need a community to start growing and being obvious.

Step two is to start working on it being self-sustaining growth if it's quite support-oriented because otherwise, you end up with a huge amount of support workload. That can be good. It can improve our products. Our engineers talk to people and generally fix things very quickly, which improves the product the user is going to experience. But it is extremely demanding from an engineering workload perspective.

I love this idea of reaching out to the power users and letting them have the space to answer questions.

🩺 Measuring community health

Louise: You mentioned you have very active users in your community, other people are, I guess, less active. How do you personally measure the health of your community? What are the metrics you're looking at?

James: So we track some numbers, but this is something we could be more sophisticated with.

One of the things that we found as a company: we have the best product-market fit for people that have themselves a product-market fit, so great stage enterprise companies. And so a community kind of popped up around the product almost despite our core focus.

We've actually said the use of the product is the thing we should focus on, even to build a community.

If we have no users, no one's going to shop in the community later, so it's irrational to prioritize having a Slack group until we've got a good active daily user number.

The top tip tooling-wise would be, and there are a couple of the products that are similar. It lets you track pretty much anything you want with your community so it gives you a lot of flexibility.

It gets quite tempting to track metrics that are sexy from an investor perspective. We see people competing over Slack group size.

I think the reality for us is like, why do we care about the community? It leads to lots of word-of-mouth growth, it leads to product enhancements; those are the kind of things that I actually care about, things like how many people are submitting code. That's meaningful. How many people are building plugins for PostHog. These are really strong properties for very deeply engaged teams, so that's more our angle so far.

It can become a vanity competition. It could be quite distracting. Just the sheer volume of people in a group. The level of engagement is almost more important.

Louise: Especially at the beginning, you prefer to have like 20 people but super engaged rather than 100 but who are not contributing anything.

And you mentioned something already that it was super hard to have people engage with one another within the community. Do you think of anything else that has been the hardest thing with regards to your community?

James: We've actually found some things harder than we expected. For example, we tried running meetups and found it quite difficult, even with the very large community size, to get a lot of attendance at things that were running. We think Covid actually impacted this quite significantly.

Realistically, though, I think it's if we get people to answer each other's questions and help each other out, that's always just being the top challenge we've had.

Things like getting people to write code for themselves and for their companies are pretty easy to do. And then it's quite easy to get people to say, Hey I'm willing to open source this contribution I just made.

So yea, everything else has been kind of easy.

There is an interesting challenge with it all as well, which is the leap of faith the community-led companies require.

As a start-up, you're looking at things like your revenue, how many customers you have, what your runway looks like, and so on. So kind of saying, Okay, I believe that if I encourage people to answer questions, this is somehow going to lead to a ton of growth, a ton of revenue and there is a huge number of steps to it.

It's a bit like investing in a brand. It is the kind of thing that many companies do not do properly because they don't really believe in it because it's hard to demonstrate that return.

So yea, it's long term. That's the reason why this is unbelievably great if you can pull it off. Because no one else is doing it.

Obviously, some companies are but the vast majority are not. So if you believe in this properly, you just throw yourself into it, you need to set yourself up for success to pull this off.

You need financing and the ability to take a long-run approach.

There are some really small things we've done that have enabled this. This has been quite aggressive fundraising very early, so we could just have a very large runway so we can just do what we think is the best thing in the long-term interest.

I wish there'd been heavier investors in it earlier on, but that's very easy to say with hindsight.

🎢 Why community can be your moat

Louise: It's really great that you mentioned this because I'm pretty confident that a lot of people actually give up on their community after three months. But you need to really work hard and just continue and it will pay off one day.

James: We've seen that the whole company's been like this because we have an open-source project. It's free, we spent a year and a half just working on that.

There are many open-source projects that are run by companies that have spent 4 to 5 years with no revenue, just believing that if we keep doing the right thing for this large userbase, some of these users going to work for big companies and that will turn into revenue later on.

Since we released our paid product last summer, it's been extremely quick to generate a bunch of revenue. But that journey took a big leap of faith.

I think the thing that's changing is there are more and more examples of companies creating this kind of playbook. So it's sort of de-risking this approach.

But it's a tough mental challenge to sustain that long-term belief that this is just the best experience from a user perspective.

Louise: It's great chatting with people you like yourself who have managed to build that. Thank you so much, James.

I hope you enjoyed the interview!

You can find the recordings of all the episodes in this series on Spotify.

Thanks 👋