Discover counterintuitive product management strategies from Snap and Discord's ex-leader Peter Sellis on team design, PM quality, and consumer product success.
Product Leadership Lessons: Building Teams and Making Better Decisions
Key Insights
- Organize teams around clear ideology and decision rights, not maximum collaboration, to reduce coordination costs and move faster
- Most product managers are statistically mediocre due to power-law distribution in the field and misaligned incentives during the ZIRP era
- Focus growth on your core product and improving retention (DAU/MAU ratio) rather than expanding into new markets or use cases
- Great PMs require contradictory traits: confidence paired with humility, process-orientation mixed with comfort in ambiguity, and long-term vision combined with daily bias for action
- Taste is exercised through restraint—saying no to good ideas and knowing what not to ship matters more than adding features
Designing Organizations Like Mission-Driven Systems
Peter Sellis describes effective team structure using a provocative analogy: organizing like a mission-driven organization with clear hierarchies. The core insight is that collaboration carries hidden costs—coordination overhead that slows teams to the pace of their slowest member. Instead, successful teams need two things: an ideological culture where everyone understands why they're doing something, and a clear organizational structure defining who can make which decisions.
At Snap and Discord, Sellis worked with founders to absorb and internalize their vision, then created shorthand language and repeating phrases that acted as cultural shortcuts. This approach reduces over-collaboration while achieving results as if the team had collaborated deeply. The framework helps product managers operate at the decision-making level founders would have used months or years prior, without constant alignment meetings.
The Product Manager Problem: Math and Incentives
The median product manager is statistically poor—not because of bad individuals, but because of power-law distribution in the field. When skill is normally distributed across a population, selecting only those who become PMs creates a power distribution among practitioners. This shifts the median far below the average.
The problem worsened during the low-interest-rate era (ZIRP) when product management became the most lucrative non-technical career in tech. With high compensation and subjective role definition, mediocre PMs stayed in the field while great ones exited—becoming founders, executives, or moving into other leadership roles. This created a self-reinforcing cycle: the distribution became increasingly weighted toward the lower end.
Great PMs are leaving the system. The solution isn't hiring more PMs; it's raising the bar and ensuring PMs create measurable net positive value.
Growth Lives in Your Core
One of Sellis's most impactful lessons: growth comes from perfecting your core product, not from expanding into adjacent markets. This applies especially to network-effect businesses and consumer products that depend on daily active usage.
At Snap, after the disastrous 2018 redesign, the company's laser focus shifted to performance and retention—making the app faster, especially on Android. This focus on existing users drove years of sustained growth. At Discord, despite massive hype around Midjourney and crypto communities, the core insight was that people use Discord to talk and play games with friends. By intensely optimizing that experience—improving performance, reducing latency, making gaming easier—Discord achieved some of its fastest growth post-pandemic, even as many assumed the gamer market was already saturated.
The math is simple: moving your DAU/MAU ratio (daily active users to monthly active users) by a few percentage points drives more growth than acquiring millions of new users with poor retention. It's the difference between frequency and acquisition—and frequency almost always wins.
The Paradoxes of Great Product Leadership
Truly excellent PMs embody contradictions. Sellis identifies three key pairs:
Confidence + Humility: You must believe deeply in your vision and your ability to shape it, yet genuinely credit designers, engineers, and others whose contributions changed outcomes. The confidence is earned partly because great PMs know they'll get too much credit anyway.
Process-Driven + Comfortable with Ambiguity: Great PMs maintain systems and checklists but don't panic when they must throw them out and rebuild. They respect good process without being enslaved to it.
Long-Term Vision + Daily Bias for Action: They know what the world should look like in 6 months or 2 years and patiently build foundational pieces, yet ship with urgency and set a relentless pace for their teams.
These aren't learned skills in the traditional sense—they're mindsets that either fit a person or they don't.
Taste: The Art of Saying No
In an era where AI adds features automatically, knowing what to remove becomes a rare and valuable skill. Sellis saw Evan Spiegel (Snap's founder) reject better ideas than most consumer startups ever shipped. This is taste in action—the museum curator who leaves most of the collection off the walls.
One way to assess this trait in candidates: ask them about prototypes they built, held in their hands, loved—but never shipped. People with great taste can articulate why the world or product wasn't ready. They stop before things become "cringe" or jump the shark. This restraint muscle is nearly impossible for AI to replicate, making it increasingly valuable for product leaders.
Conclusion
Peter Sellis's insights challenge conventional startup wisdom: organize around mission and decision clarity rather than maximum collaboration; focus relentlessly on your core and your existing users rather than new adjacencies; recognize that most PMs are statistically mediocre and that great ones often leave the field; and cultivate taste through disciplined restraint. These principles, applied at Snap and Discord, shaped two of the most beloved consumer products of the past decade.
Original source: 90 minutes of unfiltered product advice from Snap and Discord’s product chief | Peter Sellis
powered by osmu.app