I took a job at Shopify

I started as an apprentice product manager at Shopify last month.

I've built four companies inside this ecosystem, sold one, sold another, and shut down a third when the platform deprecated the file it ran on. I've written a fair amount about what it's like to be a tenant. Now I work for the landlord, and I'd rather address the obvious questions directly than have them sit under everything I write from here.

Why I took it#

I've never seen this from the inside. Seven years of reading changelogs and inferring intent from the outside, building a whole risk framework out of guesses about how platform decisions get made. There is no substitute for being in the room where that happens, and there's no way to buy that access.

I did a version of this in 2017 — left running my own company to take a media buying seat, specifically to see the inside of an operation I'd only ever sold to. That year was the most informative of my career and it directly produced the thesis for the next company. I'm running the same play.

The scale of the problems. Everything I've built has been narrow by necessity. A decision here touches an enormous number of merchants at once, and that's a different kind of problem than I've had access to.

Curiosity about how the sausage is made. I built a company that died from a deprecation notice, and I've spent two years with a private model of how that decision got made. I'd like to find out how wrong it is.

The conflict question#

People have asked, reasonably, how this works alongside running Shopify apps. The arrangement is between me and my employer, and I'm not going to narrate it here.

What I will say is that I'm not going to write about the inside. Nothing in this job is going to show up in a post. If I have a view about how platforms make decisions, it'll be built from the same public material anyone else can read, and if I ever can't write about something honestly, I'll write about something else.

That's a real constraint on this site and it's the actual cost of the job. I've been writing about platform risk for years from a position where I could say anything I observed. I've traded some of that for the access.

What I don't think this changes#

I still think the tenant analysis was right. Building on a platform gives you distribution you cannot otherwise buy, and the price is that your revenue is worth less than the same revenue from ground you own, and almost nobody applies that discount. Joining the platform doesn't make that less true — if anything I expect it to make me more precise about it.

I also don't think this makes me a defender of anything. The most useful thing I could take from this job isn't loyalty in either direction. It's a better answer to a question I've been guessing at since 2018: how do you tell, in advance, which platform changes will end your company and which ones will cost you a quarter.

I have a rough test for that — whether your product uses the platform or is a behavior of it. It's held up across four companies and it's still a feeling rather than a method.

The thing I'm actually watching for#

Whether I stop being a builder. That's the risk that doesn't show up in a conflict-of-interest conversation.

I've watched founders take a role at a large company and quietly become someone who has opinions about products rather than someone who ships them. The salary is comfortable, the problems are interesting, and the years pass. I've got Internet Research Unit and three apps that need me to still be the other kind of person.

Ask me in a year.