Why does Peter Sellis believe most product managers are ineffective?

Peter argues that product management skills follow a power distribution, not a normal one. The most talented PMs exit the system to become founders or executives, while those remaining are highly incentivized to stay, leading to a median PM being 'pretty bad.' The profession's subjective definition and high compensation for non-technical roles also contribute to this dynamic.

What are the main reasons Snapchat struggled with monetization?

Snapchat's monetization challenges stem from three core issues: an extremely young user base (under 17) which is harder to target and measure for ads, a camera-first product design that doesn't naturally integrate ads into the core flow, and its utility as a messaging app, which historically has poor ad monetization models.

How can systems thinking be effectively applied in product development?

Peter suggests that true systems thinking goes beyond simply asking 'why' five times. It involves understanding and modeling the system mathematically, mapping out stocks, flows, variables, and distributions. Creating a spreadsheet model to clarify how different levers impact the system can be highly clarifying, even if not widely shared.

What is OpenAI's approach to advertising, and what are its challenges?

OpenAI is 'speedrunning' the development of ads infrastructure, adopting a mathematical auction mechanism similar to Meta's, which factors in organic value. The main challenge is to maximize trust over the long term through this mechanism and to innovate on ad formats within a chat or voice interface, without alienating users by offering an ad-free top tier too early.

What is the counterintuitive leadership philosophy of 'riding your best people into the ground'?

Instead of focusing on improving average performers, Peter aggressively challenges and empowers his best people with more responsibility until they reach their 'level of incompetence.' He believes in loading them with decisions, trusting their judgment, and defending them publicly, reinforcing their confidence and driving maximum output.

Why does Peter Sellis believe growth should come from a product's core business?

For network effect-oriented businesses, implicit growth comes from optimizing for daily active use (DAU). Focusing on existing users and making the core product faster, more performant, and more relevant (e.g., Snapchat's Android performance, Discord's focus on intentional multiplayer gaming) drives significant growth by increasing DAU/MAU ratios, which is a form of retention.

How does SpaceX avoid 'the innovator's dilemma'?

SpaceX demonstrates a culture of 'burning the ships' by willingly sunsetting highly successful products, like the Falcon 9, to invest in future, more ambitious projects like Starship. This commitment to its long-term, multi-planetary vision, even at the cost of current success, helps it avoid being disrupted by its own innovations.

What defines a great product manager (PM) through 'oxymorons'?

Great PMs embody three oxymoronic traits: they are both highly confident (bordering on arrogant) and deeply humble, giving credit where due; they are extremely organized and process-driven but also comfortable with ambiguity, even despising authority; and they combine long-term strategic thinking with a borderline annoying bias for immediate action.

What was the Ustream Shiba Inu puppy cam and what did it teach Peter Sellis?

The Shiba Inu puppy cam was a 24/7 live stream of puppies on Ustream that became incredibly popular, with many using it for background soothing. It taught Peter about unexpected consumer use cases, the emotional connection users form with content, and the potential of live streaming as an 'uncracked product category' despite its technical challenges.

Key Moments

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

Lenny's Podcast

People & Blogs7 min read97 min video

Sep 20, 2026|626 views|30|2

Save to Pod

Want to know something specific about what's covered?

We've already dissected every moment. Ask and we will deliver (with timestamps).

Key Moments

TL;DR

Product leaders are often net negative, and teams should be structured like terrorist organizations for hyper-efficiency, but beware of AI's tendency to only add, not subtract.

Key Insights

Teams can be designed like 'terrorist organizations' with a strong ideological culture and clear decision-making structure to avoid over-collaboration and ensure results.

Most product managers are 'pretty bad' and potentially net negative, with the best ones becoming founders and executives, creating a skewed distribution where median PMs are often less effective.

Snapchat's monetization struggles stem from its extremely young user base (under 17) with restrictive ad targeting rules and its camera-first, non-feed-based format, making ad integration difficult.

Focusing on improving the core product and DAU/MAU ratio for existing users, rather than solely on new user acquisition, can drive more sustainable and effective growth, as seen at Snap and Discord.

A key skill for PMs is developing 'taste' by knowing what to remove from a product, a crucial counterpoint to AI's current tendency to only add features.

OpenAI's approach to advertising in chat interfaces needs to prioritize 'trust maximization' over simply maximizing bids, drawing parallels to early Meta successes and potential pitfalls on Amazon search.

Designing high-performance teams like 'terrorist organizations'

Peter Sellis advocates for designing teams with the efficiency and focus of a 'terrorist organization,' a concept he clarifies is about structure and ideology, not violence. This involves cultivating a strong, shared ideological culture where everyone understands the 'why' behind their work, akin to a religious fervor. Crucially, it requires a clear organizational structure that defines who can be trusted with specific decisions, minimizing the need for excessive collaboration. Sellis believes that while collaboration has costs, a well-structured team with a clear mission can achieve results as effectively as, if not more so than, a highly collaborative but slower-moving group. This approach aims to empower individuals to make decisions that founders might have made years prior, particularly in high-growth environments. He emphasizes distilling the founder's vision into concise, repeatable language that acts as a shortcut for the entire team, fostering alignment and rapid execution.

The 'median product manager' problem

Sellis presents a contrarian view that the median product manager is often 'pretty bad' and can be a net negative for a company. He attributes this to the nature of the profession and market dynamics. Product management skills, when normally distributed, often form a power distribution curve at the top end, meaning the average skill level is skewed downwards. During the ZERP era, product management became a highly lucrative career for non-technical individuals, incentivizing many to stay in the profession even if their skills weren't exceptional. Furthermore, great product managers often transition to founding companies or executive roles, leaving a gap at the median. This creates a situation where those who are not top-tier may have strong incentives to remain in the role, potentially leading to mediocre outcomes. He notes that excellent engineers often don't complain about PMs because they can guide them or avoid bad ones, while truly struggling engineers may hide behind PMs' decisions.

Snapchat's monetization hurdles

Despite Snapchat's massive user base, Peter Sellis explains three key reasons it struggled with ad monetization. First, its user base is predominantly Gen Z and Gen Alpha, a demographic under 17, where advertising rules are highly restrictive due to privacy concerns, significantly limiting targeting and ROI. Advertisers get more value from users aged 18 and above. Second, Snapchat's core loop—opening the camera to create and share moments—is inherently hostile to advertising. Unlike feed-based platforms like Instagram or TikTok, there's no natural place for ads within the primary creation flow. Sellis admits his own failure to crack camera monetization at Snap. Third, Snapchat is fundamentally a messaging app, and while messaging apps are valuable utilities, they haven't proven to be exceptionally high-monetizing from an advertising perspective compared to content-consumption platforms. This combination of a less monetizable audience, a difficult ad format, and a utility-focused core made building a multi-billion dollar ad business a significant challenge.

Focusing growth on the core user base

Sellis argues that for network-effect driven businesses like Snap and Discord, sustainable growth often comes from deeply serving and improving the core user experience, rather than solely pursuing new user acquisition. He emphasizes the importance of optimizing for Daily Active Users (DAU) over Monthly Active Users (MAU) because DAU is a proxy for retention. At Snap, a disastrous redesign in 2018 led to flattened DAU, and the subsequent focus on performance, especially on Android, for existing users sparked a growth renaissance. Similarly, at Discord, despite external focus on ventures like Midjourney, the team doubled down on its core gamer audience, improving the experience 'before, during, and after playing games.' This laser focus on existing users, making the product faster and more relevant, drove rapid growth even in segments perceived as having 100% penetration. The lesson is that small improvements in the DAU/MAU ratio can be more impactful than acquiring many new, less engaged users.

Developing product 'taste' and the role of restraint

Sellis highlights 'taste' as a crucial, yet hard-to-define, attribute for product leaders and designers. He uses the analogy of a museum curator: a great curator doesn't display everything but selectively chooses what is most impactful for a given time and place. This implies the importance of restraint—knowing what *not* to build. He suggests that a good indicator of taste in a candidate is their ability to discuss high-quality ideas they developed but chose *not* to ship, rather than just ideas that failed testing. This concept of knowing when to stop or remove features is particularly relevant in the age of AI, which currently excels at adding functionality but struggles with the crucial task of strategic omission. This ability to say 'no' and exercise restraint is presented as one of the few tangible ways to exercise and evaluate taste in product development.

The future of advertising in AI interfaces

Regarding AI and advertising, Sellis subscribes to the idea that ads are the fundamental product-market fit for the internet. He believes OpenAI is 'speed-running' the development of ad infrastructure, learning from decades of online advertising evolution, particularly Meta's innovations. He posits that modern ad auctions, like those based on VCG mechanisms, are sophisticated enough to incorporate not just bids but also the 'organic value' or 'trust' an advertiser brings to a user. For chat interfaces like ChatGPT, trust is paramount. He cautions against the current trend of saturating search results with ads, using Amazon's search experience as an example where an overabundance of sponsored results can erode user trust. Sellis believes OpenAI's success will hinge on how mathematically they instantiate their auction mechanism to maximize trust, which will likely drive user retention and advertiser value. He expresses disappointment that OpenAI opted for an ad-free premium tier early on, suggesting ads could have been additive to the chat experience.

Counterintuitive leadership and management philosophies

Sellis shares several contrarian takes on leadership. Firstly, he advocates for focusing on developing 'best people' by giving them more rope to 'hang themselves,' rather than trying to uplift average performers. This means pushing high performers harder, believing that their success is self-reinforcing and that he will defend them publicly. Secondly, he views excessive collaboration as a significant drain on efficiency, leading to teams moving at the speed of the slowest member. Thirdly, he believes growth primarily stems from refining the core product and user experience, citing examples like Snap's performance improvements and Discord's focus on gamers. Finally, he discusses the 'oxymorons' of a great PM: confidence balanced with humility, extreme organization coupled with comfort in ambiguity, and long-term vision paired with an urgent bias for action. These traits, he believes, are essential for navigating the complexities of product development.

Defining core product value and long-term greed

Sellis emphasizes the importance of defining 'core product value' using concise, memorable phrases that are tied to actionable metrics. For Snap, it was 'the fastest way to share a moment with people you care about,' broken down into specific metrics around speed, sharing, and focus on best friends. For Discord, it was 'the best way to talk and hang out with your friends before, during, and after playing games.' This ideological language provides a framework for decision-making and onboarding. He also champions the concept of being 'long-term greedy' in advertising. This means focusing on delivering advertiser ROI today, even at the potential cost of short-term revenue, because it builds trust and retention, leading to greater long-term compounding growth. This philosophy requires conviction, especially during tough public company periods, and acts as a guiding principle for sustainable business building.

Mentioned in This Episode

●Software & Apps

●Companies

●Organizations

●People Referenced

Common Questions

Peter Sellis suggests designing teams like a 'terrorist organization,' focusing on two key elements: a strong ideological culture where everyone understands 'why' they're doing what they do, and a clear organizational structure defining who can be trusted with which decisions. This approach minimizes coordination costs and allows for aggressive internalization of collaboration costs, empowering individuals to make decisions founders would have made.

Topics

Technology & Innovation Business & Entrepreneurship Leadership & Management Leadership Philosophy Product Management Consumer Products Systems Thinking AI Product Development Monetization Strategies Startup Wisdom Team Design Product Growth

Mentioned in this video

Companies

Anthropic

A company mentioned as using Work OS.

Snapchat

Peter Sellis was the first PM at Snapchat, where he spent seven years. He discusses its struggles to build a massive business despite a billion monthly active users, attributing this to its young user base, camera-first design, and messaging app core.

OpenAI

A company mentioned as using Work OS and exploring ads for its AI tools. Peter also suggests OpenAI needs his expertise in ads.

Sierra

A company mentioned as using Work OS.

Stripe

Used as a comparison for Work OS, highlighting its effectiveness in providing enterprise features.

Mentioned as an example of a platform with a more valuable user base for advertising compared to Snapchat.

Roblox

Mentioned as a platform from which Gen Alpha users graduate, eventually moving to platforms like Snapchat.

TikTok

Mentioned as the number one content consumption app, posing a challenge for Snapchat's ad strategy.

Mercury

Sponsor of the podcast, providing banking services for startups, including spend management features and individual cards with limits.

Mastercard

Mentioned as the issuer of the IO card through Patriot Bank.

Amazon

Mentioned as an example where sponsored ads can subtly influence user trust in search results.

Toyota

Credited as the origin of the 'five whys' method for systems thinking.

SpaceX

Praised as the 'greatest American company' for its ability to guard against future disruption by sunsetting successful products like Falcon 9 to focus on Starship, embodying a 'burn the ships' culture.

Instagram

Mentioned as a platform with a more valuable user base for advertising compared to Snapchat. Noted for copying Snapchat's features, sometimes even the less successful ones, but also for evolving the 'Stories' format effectively.

Twitter

Mentioned as a platform with a more valuable user base for advertising compared to Snapchat. Described as a global town hall with valuable users.

Ask anything from this episode.

Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.

Get Started Free