Snap and Discord's ex-product chief: why the median PM is bad, and three oxymorons of a great one

Lenny's Newsletter: Snap and Discord's former product chief on building a product team

"90 minutes of unfiltered product advice from Snap and Discord's product chief | Peter Sellis"

1. He designs teams that need as little collaboration as possible. Two beliefs sit underneath. The first is that collaboration carries a coordination cost nobody prices properly: coordinating with more people ties a team's speed to its slowest node, and Sellis reads the fact that no executive ever says "stop collaborating" as a warning sign. The second is that at a high-growth company a product manager should be making the calls the founders would have made a year or two earlier, which only works when decision rights are unambiguous. The organizations he borrows from, which he calls a good "terrorist organization" and admits is a phrase better said behind closed doors, get two things right: an ideological culture in which everyone understands why they are doing what they are doing, and a clear structure of who can be trusted with which decision. He spends his own attention on the people already performing, giving the strongest more to carry, rather than on lifting the average ones. 2
2. The median product manager is bad, and the reasons are structural. He argues from the distribution: product skill is spread normally across the population, only the right-hand tail of it becomes product managers, and lopping that tail off turns the distribution inside the profession into a power law, where the median sits far below the average. Two forces widen the gap. During the years of near-zero interest rates, product management was the most lucrative career in tech open to people without an engineering background, and the job's definition is subjective enough that weak practitioners have every incentive to stay in it. The strongest ones leave, because the skills of a good product leader sit close to the skills of a founder or an executive. The consequence shows up second-hand: most people who call product managers useless have worked only with the mediocre ones. He now assesses a candidate together with the place they would be stepping into, treating the two as one question. 2
3. The three oxymorons of a great product manager. Sellis describes the good ones as holding three pairs of opposites at once. Confidence that borders on arrogance, balanced by humility deep enough to hand credit to the right person, the designer whose last-second change mattered or the engineer who stayed up to make a launch. A taste for process and checklists, held alongside comfort with ambiguity strong enough to throw them out and start over. Long-term thinking, knowing how they want things to look in six months or two years and staying patient about which building blocks have to land, paired with a daily bias for action that sets the pace for everyone around them. 2

Where to start