Eeke de Milliano on Stripe, Innovation Culture, and Process as Variance Reduction

Eeke de Milliano on Stripe, Innovation Culture, and Process as Variance Reduction

transcriptlennyproduct-managementstriperetoolinnovationcultureprocess

Eeke de Milliano on Stripe, Innovation Culture, and Process as Variance Reduction

Source: Lenny’s Podcast Speaker: Eeke de Milliano Date: ~2023 Link: Episode

Eeke de Milliano is Head of Product at Retool, previously one of Stripe’s first product managers where she built Stripe Connect and Stripe Radar. This conversation covers what makes Stripe’s culture unusually generative, why most teams fail to innovate (and how to fix it), and how process can be a double-edged tool that raises the floor while lowering the ceiling.

Key ideas

  1. Process as variance reduction. Process reduces variance — it raises the floor by getting weaker performers closer to a standard. The cost is that it also reduces the ceiling, constraining the highest performers and most creative thinkers who do not need it. Companies introduce more process as they scale not because they want to, but because variance has become intolerable. Understanding this trade-off is the precondition for managing it well.

  2. Stripe’s culture: think rigorously, write clearly. Stripe’s operating principles (not “values” — you can’t tell people what to value, but you can tell them how to operate) include think rigorously, move with urgency and focus, and micro-pessimists, macro-optimists (critical about daily decisions, optimistic about long-run trajectory). The culture was operationalised through a strong writing norm: you cannot write clearly unless you think clearly. First-principles questioning of the status quo (“why is it done this way?”) was the default stance, founder-modelled and propagated through hiring.

  3. When companies need PMs. Stripe had no PMs until ~100 employees, because the product was built for developers and the engineers were the customers. The need for PMs emerges when: (a) the customer base expands beyond the builders; (b) organisational complexity grows — multiple products, multiple geographies, legal/compliance matrix. PMs bring the whole story together across teams that don’t naturally see it.

  4. Why teams are not innovative. Three root causes: (a) fear of failure — leaders want the upside of innovation but not the cost; fix by normalising failure publicly (retrospectives, shipped-to-org emails, celebrating learning); (b) urgency trap — tech debt and incidents crowd out strategic thinking; (c) thinking big is genuinely hard and takes time — “give people permission to think” as an explicit leadership act. Structural tools: “Think Bigger” section in team charters (“with 20% more time, what would you do?”), annual Crazy Ideas doc (90% chance it makes no sense, 10% chance it makes a 10x difference).

  5. Startup-within-startup for new products. New product bets start at 1–2 people (one engineer and one designer, or one PM) and get no additional funding until there is clear evidence of something. Retool functions as the VC; the existing customer base is the unfair advantage. This model produced three launched products in one year with a 300-person company.

Context

The Crazy Ideas tradition was in use at both Stripe and Retool: at the start of each year, the founder sends a blank doc to the org with the prompt to list crazy ideas that are probably wrong but would be transformative if right. Consistently 3–8 ideas from each year’s doc end up being executed. Trapdoor decisions (one-way vs. two-way door, borrowing from Amazon) were evaluated rigorously at Stripe — pricing is often not a trapdoor decision (existing users can be grandfathered); titles are (you cannot revoke a title once given).