Archie Abrams on Shopify's Growth Model, Long-Term Holdouts, and the 100-Year Vision

Archie Abrams on Shopify's Growth Model, Long-Term Holdouts, and the 100-Year Vision

transcriptlennys-podcastgrowthshopifyexperimentationmetricsproduct-managementmarketplaces

Archie Abrams on Shopify’s Growth Model, Long-Term Holdouts, and the 100-Year Vision

Source: Lenny’s Podcast Speaker: Archie Abrams Date: ~2024 Link: https://www.lennysnewsletter.com/p/shopifys-growth-archie-abrams

Key ideas

  • Optimise for churn (Shopify’s contrarian retention model). Most SaaS obsesses over retaining every user. Shopify’s mission is increasing entrepreneurship, so it removes barriers to starting. Many merchants will fail, but the power-law winners (e.g. Allbirds, FIGS) generate enough GMV-linked payments revenue to make the entire cohort profitable. The unit of success is total cohort GMV over three to five years, not per-user retention. Monetary friction (trial length, price, incentives) is the primary lever: reducing it can unlock a class of merchants who would otherwise give up too early.
  • Long-term holdout experiments: 30–40% of short-term wins disappear. Shopify runs automated re-reports on every experiment at 3, 6, 9, and 12 months — no one can hide from long-run results. Roughly 30–40% of experiments showing short-term metric lift have no long-term GMV impact, usually because the lift was a pull-forward effect. Conversely, neutral short-term experiments sometimes show large downstream impact (e.g., pre-configured store sections had zero lift on conversions to paid but drove significant long-term GMV by helping merchants build better stores). Practical heuristic: if something showed no short-term lift, it won’t have long-term lift; if it showed lift, still ship it, but don’t overestimate the impact.
  • Absolute numbers over conversion rates. Teams that own a funnel stage and optimise its local conversion rate create a perverse incentive: the easiest way to lift conversion is to make it harder for users to reach that stage. Orienting every team to maximise the absolute number of people through their stage removes this incentive. An experiment that reduces conversion rate but brings more people through the door can be the right call; CAC falls proportionally, leaving more budget to invest.
  • Taste-based accountability vs. metrics-based accountability. Shopify’s core product team (under Glen Coates) has no KPIs. Decisions are driven by taste, product sense, and founder intuition (Tobi Lütke). Every shipped project requires “okay-to” approval from the relevant group lead, enforced through Shopify’s internal project management tool GSD. Tobi and all R&D group leads review every active project every six weeks, often spending 30 minutes on technical architecture decisions (e.g., how to build a CSV importer). This model only works with an extremely opinionated founder or leadership group with strong, consistent taste — without it, it devolves into undirected building.
  • The “how” determines strategy in technology companies. Tobi’s view: technical architecture decisions matter more than the “what” or “who.” Building with long-term optionality takes longer and forgoes short-term wins, but creates a platform that can adapt. Shopify’s 100-year orientation — “build the best product, make money to do more of one; never reverse one and two” — keeps the SMB/entrepreneur segment primary even when enterprise revenue looks attractive short-term.

Overview

Archie Abrams — VP of Product and Head of Growth at Shopify, overseeing 600+ people across product, design, engineering, data, and growth marketing — offers a detailed account of how Shopify breaks conventional growth rules. Key topics: why Shopify optimises for churn, long-term holdout experiment infrastructure (with 30-month experiment re-report automation), the absolute number vs. conversion rate distinction, Shopify’s org structure (growth R&D: growth product, enable, customer support; growth marketing: paid, affiliate, email, SEO), embedding sales into a historically PLG motion, and Tobi’s 100-year technical architecture philosophy. Also covers discounting as a value-signalling lever (from Archie’s time at Udemy) and why Shopify does not have a CMO.