Investing in Stellate
behind the scenes on one angel check
Thanks for the patience while I was radio silent for a bit! I think that's apt to happen occasionally. This week I'm sharing the more personal side behind the Stellate investment.
Next week will be about a few frameworks I use to assess how someone thinks about their career (they could be useful for you personally, or if you find yourself in a managerial or advice-giving role).
Thanks for reading Ellen’s Newsletter! Subscribe for free to receive new posts and support my work.
Last month Stellate (fka GraphCDN) announced their seed round led by Boldstart, as well as their Series A led by Tiger! There's an official post on the Boldstart blog, but like I did with Liveblocks, I wanted to add some personal color, too.
I no longer angel invest in developer tooling because of my role at Boldstart, but my first decision to invest in Stellate was as an angel. Max had DM'd me* about what they were building so we could schedule time to chat.
At the time, my personal angel policy was "take a 30 minute call, sleep on it, then decide." In this case I made an exception where I "slept on it first," and was fairly confident I'd commit during the meeting.
Investors talk a lot about needing to build up relationships and see lines (vs. dots) over time, so I thought it would be interesting to share how I came to such a fast conclusion for Stellate, in the hopes that it might help other founders get to more “yes” answers from angels.
First, I'd already built a scaffold of information about the GraphQL space. That's not to say I knew exactly where the world was going and looked for a company doing that. Quite the opposite. I see a lot of new things every day, and use that to start to think about spaces where something interesting might happen. I'm a big believer that founders (end early teammates) will tell me what the "something interesting" is and I just have to recognize it. I was not going to go out and start a GraphQL company.
In this case, at the time we started Dark (early 2017), GraphQL still felt very niche. It had only been open sourced and publicly available since 2015. We'd talked about it a lot, but decided scoping to only support REST was most appropriate at the time. Over the following three+ years, I never heard the end of that choice. Developers who were interested in new web backend languages were also interested in new backend API protocols, so it made sense that almost all Dark users wanted GraphQL support.
Given that, I already had real conviction that the world was moving towards GraphQL as an API format. There was definitely something going on, but just because I knew we'd need more support around GraphQL didn't mean I knew what that support should be.
Second, I knew that Tim and Max were the types of engineers who would be apt to know what could be interesting here. Tim was originally an early user of Prisma before becoming the first engineering hire (at the time it was still Graphcool). That makes the call really easy. If I think GraphQL is going somewhere, do I think Tim has the potential for a uniquely informed technical perspective on what that should be? Absolutely.
But even if I believed that 1) Something would happen in the GraphQL ecosystem and 2) Tim had the right background to have a perspective on what that was, that still wouldn't be quite enough background for me to want to commit in the meeting. I still needed 3) Can they build something?
While Twitter isn't perfect, it does allow you to get to know someone over a long period of time in a low-fidelity way. Max could DM me because we'd been Twitter mutuals for a long time. But more importantly, I'd seen him go through parts of building (and exiting) Spectrum, as well as knew about his long running projects like Bedrock and styled-components. These weren't just things he'd built - they were things he'd built, written about, and supported users in.
So without having had the initial conversation, I knew that I was looking at a deal that had three key components:
founders with expertise in the space
founders with track record of executing
For an angel investment that's definitely enough signal to get me very interested. As an angel I was also comfortable with the idea that Max and Tim might pivot around in the GraphQL space before landing on the use case they wanted to go after. (When we invested at Boldstart it was a few months later so there was more information to dive into around the initial caching use case, but you can read about that in the other post).
That meant the only thing I wanted to get at in our conversation was what was driving them personally and if I thought we'd like and work effectively with each other. That makes for a more fluid conversation. We talked a lot about interpersonal dynamics and I was happy to learn one of the things that Max and Tim valued from our chat was being open and direct (in fact, default to open is also one of the core values at Stellate).
As I'd expected, it was an easy yes by the end of the call!
If this has appealed to you, Stellate has a ton of open roles and I'd be happy to help connect you. If you just need to cache your GraphQL API, head right on over to https://stellate.io.
Across the board, I think this pattern is more common in angel investments than people like to talk about. Timelines for angel checks tend to be short. It’s much harder to get to a strong "yes" in half an hour when you spend most of it just trying to ramp up on a space.
Getting the easy “yes” is something that’s built over time from both sides. It doesn’t have to be a series of 1 on 1 meetings with investors. One way is to focus on investors who have already spent time in your space. Another way is by spending time on loose ties where people can see your work, even if you don’t quite know what the immediate benefit will be. Being able to get that sort of low hanging fruit out of the way can let you focus a 30 minute conversation on deeper topics.
DMs worked here because we're longtime Twitter mutuals. In the case that you’re reading this and you want to cold pitch me, please send an email, because otherwise I'll just either totally miss the message or ask you to email me anyway.