Growth for developer products
With Martin Gontovnikas, Interim CMO at Vercel & Co-Founder at HyperGrowth Partners
It’s 2024 and everyone needs help with growth, but developers are deeply allergic to activities we tend to associate with growth marketing.
The standard advice is to get them into the product and get out of the way, or risk coming off as untrustworthy.
Enter Martin Gontovnikas – a.k.a Gonto – a software engineer who moved to the proverbial dark side to focus on marketing. He currently runs marketing at Vercel and co-founded HyperGrowth Partners, and was SVP Marketing + Growth at Auth0 before that.
In this interview, we share Gonto’s playbooks for:
Building your team: How DevRel + Growth should work together, and who to hire first
A three-pillar content marketing framework for developers
Building a developer influencer strategy
There’s a perception that developers hate marketing. When a devtools company comes to you for help with Growth, where do you start?
The first thing to understand about doing growth for dev products is that every developer is not the same.
I’m a big believer in starting with research, because how these different engineers relate to products is very different. Do research to understand the type of developer you’re speaking to, and how they interact with the problem your product solves.
For example, to understand front-end developers, I’d look for 8 of them on Twitter and offer them $200-300 for half an hour or an hour of their time. I’d tell them: “I won’t sell you a product – I just want to understand how you learn about and relate to things.”
Let’s say I was doing this for Vercel. I’d ask the participants when and how they want to learn about deployments and managed infrastructure; who they follow on Twitter; what communities they are part of; what events they go to. Based on that research, you can then build your growth strategy.
As a real example: when I worked at Auth0, we realized that developers didn’t give a shit about authentication – when we asked them, they said they just didn’t care.
What was interesting is that they told us: “I don’t want to learn about auth, but when I’m implementing it and I get stuck, I will Google [a solution].”
Our strategy became to write landing pages with content for potential errors developers could run into [when implementing authentication] – which was literally what they’d be googling.
Should startups work with developer influencers, or “tastemakers”?
Sure. There are two ways you can engage with influencers: organic or paid. Starting with organic is usually best, and I believe the best organic relationships are usually built in person, not online.
That’s why having a good DevRel strategy is key. Once you’ve identified the influencers in your space, I recommend going to conferences to speak, because that’s where the influencers are – you’re going to go to speaker dinners with them, hang with them and get to know them – and that’s how you build real relationships.
Once you build these authentic relationships and you write some good-quality content, they’re likely to share it for free because you’ve become friends.
Once we build these organic relationships with influencers, we will give them access to features in beta before they launch. We give them a lot of exclusive experiences; we add them into our Slack to give feedback on early products or features. It’s a closer relationship, and they end up sharing more about us because of that.
What about paid developer influencers? What have you seen work?
Sponsored YouTube videos can do really well. I’ve noticed that developers are watching a lot more YouTube than before.
For example, with Clerk– one of the companies we work with – we decided to try to sponsor a YouTuber, but it didn’t work. Nobody cared. So we decided to do some research to understand why they didn’t sign up.
People responded saying that they didn’t believe what the influencer said because they knew it was a paid video.
So Nick, the head of marketing there, had a great idea. He decided to sponsor videos that were less about our product specifically, and more about doing some other project. We would sponsor a different YouTuber for a video on “How to build a ChatGPT clone,” for example.
As part of building that clone, they used Clerk, but they didn’t really talk about it; they didn’t promote it specifically. It said it was a sponsored video in the end, but of course, most people don’t see the end. That worked to drive signups because it was more natural.
You can also take the approach of teaming up with other devtool companies on a video where someone is using multiple technologies in one sponsored post; it’s cheaper and can work really well.
Tell me about the financial side – how do you pay influencers?
We always try to pay by CPM (cost per thousand views) because if the content doesn’t get many impressions, we don’t pay as much.
In some cases they’ll want to be paid per video regardless of CPM. We try not to agree to those terms, because if they make the video and don’t get any views, it just doesn't make sense.
When paying by CPM, you also want to add a cap for the maximum you’ll pay per video.
Finally, we also try to negotiate on a minimum number of views, and if the video doesn’t hit that [minimum], they have to do two videos for us instead of just one, for example.
What are the most common growth mistakes that developer startups make?
The first mistake is when companies think that all developers are the same. Number one, you need to understand that every developer is different.
A second mistake is that when companies hire DevRel, they fail to think about what framework or programming language they want to focus on first.
When you hire DevRel, you want to hire people who are already part of the community that you want to target. Think about: what’s the community I want to bet on? Why?
Vercel, for instance, used NextJS as a way to drive acquisition. They started by saying, “anyone who uses NextJS should use Vercel because it works better that way,” and they expanded to other frameworks from there.
My advice for startups is to bet on two different frameworks to begin with: one that’s established and one that is up and coming.
The established framework has a big community but it’s hard to get in [and become known]; the up-and-coming framework may be a smaller community, but it’s easier to become known – and then if that framework grows, then it was a good bet.
A third growth mistake happens when devtool companies eventually want to go up market and sell to the Enterprise. They start hiring more salespeople, CS, etc. Companies that do this well do it in addition to their developer love and developer focus. The ones that fail at it forget about developers. New Relic is a good example of a company that this happened to.
On the topic of DevRel, there’s the debate about what metrics – if any – DevRel teams should be responsible for. What’s your take?
You’re going to use different metrics depending on what you want to do. One thing to consider is tracking first-touch attribution on content DevRel teams build.
The idea here is that, for smaller startups, nobody really knows about your product. People are discovering your company because of this DevRel content. This is one way you can measure effectiveness.
A second idea: because DevRel is a lot about relationship-building and influencers, you can look at what type of influencers are sharing content about your startup and how often they’re sharing it.
Maybe we have 15 people who are important influencers for our startup. A KPI here could be that we want 50% of those 15 people to talk about our product at least once per month. This helps us understand whether people are talking about us and what it means to the business.
There are other metrics people look at, like reach for conference talks, but in my opinion those are more vanity metrics than anything else.
For new startups, the founder is often the biggest growth driver for the business. But it’s hard to justify a founder spending their time tweeting or posting on LinkedIn. Should founders spend their time this way?
It can definitely be worth it, but it depends. Everything in a startup is a conversation about trade-offs and priorities.
You can potentially get someone to own the social media of the founder. Perhaps you pay a DevRel person to co-write, and eventually they understand the founder’s viewpoints enough to write some content themselves.
I worked with a company that tried a similar strategy. This company had a customer base that lived on LinkedIn. The founder didn’t have time to write the content, but we had marketers start writing content for the founder.
They had around 100-150 employees at the time. They weren’t really known, but we wanted them to be. So a second thing we did was ask people in the company, for a period of time, to like the founder’s posts and reply to them so that the LinkedIn algorithm would recognize it. This engagement allowed us to start getting more traction on future posts.
When should a devtools founder hire a dedicated growth person?
The first hire that a devtool founder should make around marketing is DevRel. Once you have one or two DevRel, you should hire someone for Growth.
When hiring for growth, you don’t need to hire someone who has done Growth in the past. Prioritize hiring people with experience in DevRel, engineering or selling to engineers over those with previous roles in Growth. The ability to understand engineers is much more important than growth experience to do Growth for devtools.
This tends to be a good combination because DevRel people are great at content, community and relationships, but they typically suck at project management and thinking about things as experiments. DevRel can then do their own thing, and Growth can layer on marketing programs with the help of DevRel.
When is it a good idea to hire a growth advisor versus a full-time hire?
First, you should hire someone for Devrel, then hire someone for growth who understands engineers and has good project management skills.
Then you can consider layering on a growth advisor who can teach your growth hire what it means to do Growth really well.
Growth advisors can also bring creativity to your programs as a person from the outside. One of my mantras is that “it’s always better to be different than better.”
If you’re the same as the rest but only slightly better, nobody will care. If you do something very different from everyone else, even if it’s shitty, it will get more attention.
How should founders approach content strategy for a developer audience?
The most important thing is understanding what your audience wants to read. I follow a 50-30-20 rule with content.
50% focuses on content that you think is important to your target market based on your research, what you’re reading on Twitter, the founder’s views, and stuff like that.
30% focuses on SEO content for driving traffic. The idea here is that you want more unique visitors and more people clicking, which will improve your Google click-through rate, which will then drive reads of your “conversion content” – the 50% bucket that you really care about.
20% is what I call “jump the wave” content. This is content that capitalizes on trends that people are talking about to see if you can capture virality from it. This 20% isn’t about Google click-through rates, it’s more about virality and acting fast on timely trends that are happening.
You also have to remember that distribution of content is equally, if not more, critical than the content itself. Deeply thinking about how you’ll distribute the content – what communities, what influencers – is key.
Any devtools you admire for their content?
Honestly, Vercel has been doing a great job on video content lately. PlanetScale does a good job of deeply technical written content. They’re not just writing about their product; they’re deeply focused on how THEY build their service and think about scalability.
I love failure stories. What are some examples of growth experiments that failed or surprised you?
Here’s two.
In the early days of Auth0, we did a paid ad campaign to drive signups. The ads ended up succeeding in driving a lot of signups, which was our main KPI, but we failed to look at the quality of those signups and how many of them activated. Once we did, we realized that none of the signups we drove were activating. We ended up spending $200K to drive a bunch of signups that did nothing. Some of them were from countries outside of our target market, like India, Indonesia and Vietnam.
Another story: at Auth0 we built a series of onboarding emails for new signups. Our open rates and click through rates were constantly improving over the course of a year. We decided to do a test to check activation rates for people who received the onboarding emails versus those who didn’t through a control and test group.
And what we saw was surprising: The people who didn’t receive the emails activated 2x better than people who received them!
We realized that, while we did a great job at improving open and click-through rates by changing the subject lines and the content, the emails themselves were very confusing.
They were promoting the wrong features too early in the user’s lifecycle, and negatively impacting retention.
Unfortunately it took us almost a year to realize this. We killed all emailed instantly and slowly started to re-add them with a new focus on activation.
Both of these stories are related: In both cases we picked the wrong KPI. We thought we were doing a great job optimizing for that KPI, but the problem was that the KPIs were a vanity metric rather than things that actually helped improve the bottom line.
What’s your spiciest growth take?
Everyone in Growth talks constantly about data. But if all we care about is quantitative data, we’ll only come up with stuff that is only slightly better than what we had before – faster horses, basically.
I personally believe that to be a good growth person, gut feeling and taste matter a lot more than your quantitative data.
Similarly, qualitative data – talking to people, watching recordings– matters a lot more than quantitative data.
If you ask me the difference between a great growth person and a shitty one, it really comes down to taste, creativity and gut feeling. A shitty one would just focus on data. A great growth person deeply understands the customer, and comes up with ideas based off of that understanding.
Thanks for reading. If you haven’t already, subscribe to get our next interview directly to your inbox.