Every product roadmap is a theory of causation — a bet that action A will produce outcome B. The philosopher of science Imre Lakatos noticed that scientific research programs protect their core assumptions by surrounding them with a 'protective belt' of auxiliary hypotheses. When a prediction fails, practitioners adjust the belt — tweaking timelines, blaming implementation, questioning the measurement — rather than touching the core. Product teams do exactly this. The core belief ('users want X,' 'retention is driven by Y') stays frozen while every adjacent explanation rotates around it. Lakatos called programs that do this 'degenerative': they absorb every failure without ever updating the fundamental bet. The discipline is to periodically declare the core hypothesis falsifiable in advance — to specify, before shipping, what result would actually force you to abandon the central belief, not just its accessories. A roadmap without that specification isn't a strategy. It's a commitment to never being wrong.

In the last quarter, what result did your team explain away — and what would have had to happen for that same result to count as evidence against the core assumption?

Drawing from Philosophy of Science / Lakatosian Research Programs — Imre Lakatos (The Methodology of Scientific Research Programmes, 1978)