Adriel Frederick on Facebook Growth, Algorithmic Products, and the Marginal User

Adriel Frederick on Facebook Growth, Algorithmic Products, and the Marginal User

transcriptlennys-podcastgrowthproductalgorithmsdiversityfacebooklyft

Adriel Frederick on Facebook Growth, Algorithmic Products, and the Marginal User

Source: Lenny’s Podcast Speaker: Adriel Frederick Date: ~2022 Link: https://www.lennysnewsletter.com/p/humanizing-product-development-adriel

Key ideas

  • Humans in the loop for algorithmic products. PMs on ML-heavy products must design the interface between humans and algorithms. Algorithms optimise well for a given objective at scale but cannot see strategic context, long-term effects, or competitive signals (e.g. a snowstorm, a competitor price move, regulatory change). The “techno utopian” failure mode: assume you can feed all data to an algorithm and it will do the right thing. The PM’s job is deciding what the algorithm is responsible for, what humans are responsible for, and the framework for making those calls.
  • Marginal user as product design north star. The marginal user is the person just on the cusp of taking the action you want — highest traffic, lowest conversion. Go further to the worst case (feature phone, slow connection, distant data centre) to see everything that’s wrong with the experience at once. Data tells you how bad; watching and talking to users tells you why. Don’t try to read answers out of a funnel — problems are often orthogonal to the funnel you’re staring at.
  • Growth is lead bullets, not silver bullets — plus occasional cannonballs. Facebook growth was not primarily clever tricks. The fundamentals were simple and ground on repeatedly: make the product easy to find, easy to get into, easy to find your friends. “Growth hacks” are table stakes — good product communication. Cannonballs (phone number sign-up, reliable SMS delivery, seeding the friend graph) required months of hard, unglamorous engineering work. A portfolio discipline matters: roughly 80% of effort on cannonballs, 20% on lead bullets.
  • Diversity as business value. Diverse teams act as built-in proxy panels for global users — saving research cycles and surfacing blind spots instantly. The unlock is recognising concrete business value (faster iteration, fewer user research cycles), not just responding to social pressure. Retention follows when culture genuinely rewards and utilises diverse perspectives.
  • R&D team organ rejection. Incubation teams face organisational “organ rejection” when the rest of the company sees them as unrelated to the core mission, competing for resources unfairly. Three countermeasures: make the work core to the company mission; make it everyone’s success, not just the incubation team’s; signal that innovation is not monopolised by the R&D corner.

Overview

Adriel Frederick — first Black PM at Facebook (growth/user acquisition), five-year director at Lyft (marketplace, pricing), VP Product at Reddit’s incubation team — covers lessons from across his career: the design problem of keeping humans in the loop with algorithms (from Lyft marketplace pricing), the marginal-user heuristic for product optimisation (Facebook growth), what Facebook growth actually was (grinding on fundamentals, not hacks), and practical observations on building and retaining diverse teams.

The thread throughout: good product management is about designing the right interface between human judgment and machine power, at every scale.