Breaking the rules of growth: Why Shopify bans KPIs, optimizes for churn, prioritizes intuition, and builds toward a 100-year vision

The way we think about churn is really going back to Shopify as a mission of what we want to do, which is to increase the amount of entrepreneurship on the internet. As a business, we want to make it as easy as possible to get started with your online store, with your business. But most businesses do ultimately fail

If we do that, again, many of those businesses, many of those folks will maybe on their first attempt not be as successful, but we're going to have a set of merchants who go on to become extremely big businesses, the Allbirds of the world, Figs, etc. And the way the Shopify business model works is we do charge a subscription, but most of our revenue comes from payments, which is tied directly to a merchant's success

I would encourage everyone, if you can, look at some of the experiments that you thought were your biggest winners. Look at the down tree metrics for a year, two years on that experiment. And I bet you'd be surprised how many times the metric is different than what you thought it would be after a year

The most common is actually there isn't a long-term lift from a lot of things that you might think of in the short term. You get a short lift on the metric upfront, a more short-term metric, number of people who become a paying shop or number of people who make their first sale on Shopify. And then you look a year later, and there's actually no incremental lift on GMV from that cohort

Typically when you can lower the barriers to monetary friction in some form, that could be all sorts of monetary friction early, the common belief is that we'll usually get lower quality folks coming in the door. But in a business case, if I give you a little monetary boost and reduce that monetary friction, I can actually causally change your ability to become successful

We ship the winner to 100%, but we're looking at the cohort of folks who was assigned to the experiment. We're going back and looking at those people who were assigned a year later. So it allows us to still ship, get stuff out, but we've kind of held the experiment in a way that allows us to see those long-term effects just for the cohort that was exposed

So it's interesting. Most people use cohort retention curves. You're using cohort because you don't look at retention. You're looking at for GMV over time. So that's really interesting. GMV over time, which correlates and profit better. And then really like the absolute number of merchants who are on the platform and then reach a certain GMV

There is an enormous amount, and we do see these with long-term effects, but just the nuts and bolts of sign-up, collecting the right information. And you usually want to collect more information than most people think you do in your sign-up flow if you can leverage that to personalize the guidance

Because we focus on that long-term GMV, number of merchants who are successful, orienting every team to think about the total number of people, not the rate, but the total number of people who got to the end of their part of the journey is a very powerful way to incentivize people to do the right thing

When you have teams naturally break up the world into different funnel stages or different points in the journey, it gets very seductive to look at my part of the funnel, and what's my conversion rate through that part of the funnel? And then the team starts to optimize for that conversion rate as their north star over a longer time period. I'm going to try to move my conversion rate from 10 to 12% or what have you. But in practice, it's actually almost always easier to just make it harder to do the thing right before your step in the funnel to increase your conversion rate.

So instead of, I'm trying to convert a bunch of people, a conversion rate, I just want more people to get activated. And then once you start thinking that way, you realize actually the best way to get more people to get to a step, sometimes and often is just get more people in the door in the first place

One good example is something around payment failure notifications. So we did a bunch of experimentation around, hey, how can we alert people that their credit card has failed. And at the typical kind of growth win, usually produces a lot of short-term impact. But you look back six, 12 months, there was really no long-term lift

We tend to ship neutral. It's like, it could be positive. And so let's like, let it go. If we have good intuition about it, and it will turn. So we've seen a bunch of these things go in very different directions

And in that conversation, so much of the conversation is about both the technical how, how are we building this in a way that allows for Shopify to have optionality in the technical decisions that we are making. And I think for Toby, one of the things I've learned in this over is that the how, the technical architecture, determines strategy in a technology company even more than the kind of what and who we're building for

It's certainly like in growth, you have the metrics, but they take a different form. And then in core, it truly is, do we have conviction that this is the right technical foundation to build the feature of commerce. And that is built through certainly looking at data, but it's not the overriding