Patrick Collison: Philosophies for Running Stripe, Hiring, Decision Making, and Reading

KEYPOINTS

• 01

  Hire for rigor, appetite and warmth—and prefer correctness over cohesion.

  Many organizations prize smooth interactions and unruffled feathers, effectively choosing cohesion over correctness. Stripe looks for people who seek correctness, don’t mind being wrong, and will entertain ideas that seem improbable. The second trait is hunger and willfulness. A startup’s default outcome is near-term non-existence, so you need people who enjoy pushing against that, not merely tolerate it. Third is interpersonal warmth and a desire to make others better. You spend so much of your life at work that it’s worth finding people you actively enjoy. Collison adds that his answer would have differed two or three years earlier.

• 02

  When an obvious idea doesn’t exist, look for the hidden force—then check how big the problem really is.

  By almost every sane analysis Stripe was a bad idea: a crowded market, regulation, no domain experience, two young non-citizens, and no viral distribution. What drove them was that it seemed impossible developer-friendly payments didn’t exist. They spent months googling for it. Its absence first worried them, because consumer banking had such a hidden force: rules effectively mandating products and almost no new charters after the crisis. Payments had no such barrier. Large companies listed the same wished-for features as hobbyists, and only about 2% of global consumer spending happened online. The ‘lake’ was an ocean, so they dropped out.

• 03

  Scaling is easy to diagnose and hard to keep up with.

  Problems are usually visible; failing to see them is subjective blindness, so the question is what your biases hide and what corrects for them. Solutions rarely need breakthroughs, since others have done it before. The hard part is that you don’t control the clock. A startup is like a tower-defense game: you control only how fast you build defenses, and problems can arrive faster. Some days the pile is so unreasonable it feels like someone is secretly messing with you. Leaders also have to build management skill on a timeline set by the company’s growth.

• 04

  Make faster decisions, sort them by type, and make fewer yourself.

  Collison names four changes from five years earlier. First, value speed: twice as many decisions at half the precision is often better, because extra deliberation quickly hits diminishing returns. Second, categorize by reversibility and magnitude, and deliberate only on irreversible, high-impact calls so that process doesn’t seep into the rest. Third, ask why you’re deciding at all—needing the CEO often signals something broken. Fourth, when you disagree, find the difference in models. Teams now write down missions, key long-term metrics and customers, so disputes are about means, not hidden disagreements about ends.

• 05

  Split effort between exploiting and exploring, and let teams differ.

  Roughly 70–80% of effort goes to optimizing what already works, or the company ‘flies into the side of a hill.’ About 20% goes to speculative bets, with the mix varying by team to avoid diseconomies of scale. Good bets are judged on their overall return profile. Atlas, which lets founders anywhere incorporate, had limited downside, low fixed cost to learn, large upside and useful spillover skills. Collison thinks bets are rare for sociological reasons more than cost. Optimizers and explorers must disagree structurally yet ‘have dinner together’—organizational pace layering, after Stewart Brand.

• 06

  Organizations are biology, not portfolios of decisions.

  An investment fund makes binary buy-or-sell calls. A company is fluid, so the real levers are incentive structures, mindsets, goal definitions and feedback loops from outcomes back to the work. The right unit of analysis is the cell and the organ. Stripe passed 500 people only about a year before, and claiming to have it figured out would be hubris. It studies other organizations, noting how tiny WhatsApp and Instagram teams operated at enormous scale.

• 07

  Admire cultures that are strong, perfectionist and long-lived.

  A strong culture is one whose members can praise it for half an hour and then list what they dislike; most are ‘milquetoast averaging.’ The Economist and Apple keep authors and designers anonymous because the work speaks for itself. Sequoia has been a top-three VC firm for about four decades. Koch Industries grew from about $20 million to an estimated $100 billion in revenue over roughly five decades with no single breakthrough. Koch and Berkshire share an obsession with structuring doubt and accounting for bias.

• 08

  Read actively, ruthlessly and by following what admired people admired.

  Collison starts about half the books he gets and finishes about a third of those, or one or two a week. Among books worth reading, he prioritizes the enjoyable ones, which still outnumber a lifetime. He opens mid-book to test whether he wants to be there, follows his brother John’s rule of always reading the best book you know of, and switches freely. Used hardbacks left around the house create collisions. He annotates margins, lists key page numbers inside the back cover, and writes reviews for friends, such as one on Joel Mokyr’s A Culture of Growth.