Stripe

Stripe plans a year out because its payments market is predictable. Payments behave much as they did before, so Stripe can model bull, base and bear cases and plan many months ahead. OpenAI cannot model its market that way and plans 60 to 90 days at a time.

Tara calls her Stripe plans theoretical, like a 'PhD thesis'. At OpenAI she found the switch to a sharp hypothesis and a fast user test jarring at first but necessary. The lesson she draws is to set cadence by "understanding like the pace of the market you're in."

Lenny's Podcast▶ 38:30

Share docs at 70% so people help finish them. A manager told Tara to bring rough drafts to the people whose buy-in she needed, have them poke holes, then repeat with the next person. She used it many times at Stripe, where her briefs could go viral internally.

Her reasoning: "very few great people want to interact with like a perfectly polished finished idea." Stripe's writing culture was polished overall, down to weighing every enum name.

Lenny's Podcast

Stripe plans a year out because its payments market is predictable. Payments behave much as they did before, so Stripe can model bull, base and bear cases and plan many months ahead. OpenAI cannot model its market that way and plans 60 to 90 days at a time.

Tara calls her Stripe plans theoretical, like a 'PhD thesis'. At OpenAI she found the switch to a sharp hypothesis and a fast user test jarring at first but necessary. The lesson she draws is to set cadence by "understanding like the pace of the market you're in."

Lenny's Podcast▶ 21:25

Share docs at 70% so people help finish them. A manager told Tara to bring rough drafts to the people whose buy-in she needed, have them poke holes, then repeat with the next person. She used it many times at Stripe, where her briefs could go viral internally.

Her reasoning: "very few great people want to interact with like a perfectly polished finished idea." Stripe's writing culture was polished overall, down to weighing every enum name.

Lenny's Podcast▶ 15:59

Let roles blur, but keep one DRI for the outcome. At Stripe there were no boundaries on what an engineer, PM or designer could do. Tara's rule is that work goes to whoever has the affinity and capability, while one person owns whether users adopt the product, want it and find it high quality.

"Someone is the DRI and then whatever work needs to be done to make that possible."

Lenny's Podcast▶ 9:33

Stripe data can tie organic video to revenue. SideShift connects Stripe, along with RevenueCat and Superwall, so brands can see revenue from UGC videos.

Lenny's Podcast

Pingo moves US users to Stripe web checkout for better margins. Because Stripe web checkout only works for Pingo in the US, it splits paywall tests between US and international users. The split showed 3-day and 7-day trials roughly equal in the US, while 7-day converted about 2% higher internationally.

Superwall▶ 35:25

Build fraud defenses before launching a generous free offer. Lenny's Product Pass offered a free year of Cursor, Lovable, Bolt, Replit and v0. Fraud rings, mostly in China, exploited the API and spread it through student networks, and Lenny had to work with Stripe and Substack to shut them down.

His engineer barely slept for a week patching holes, and Lenny calls it his most stressful business moment. "That was way too good an offer. People came for it, bad people."

App Masters▶ 16:10

Polish billing, cancellation and refunds before reading web funnel results. Off Apple, an app owns payments, cancellation, plan changes, refunds, support and chargebacks. Rise Science found payments infrastructure was its early pain on the web, before the funnel itself worked.

Without a way to cancel, trial conversion looks falsely high while chargebacks and refunds climb, reviews suffer and card systems may flag you. Leon Sasson: "The billing sort of work stream of the journey needs to be just super polished."

Superwall▶ 26:05