Evidence-first product management.
Field notes on turning customer evidence into decisions you can defend, and ship. Cited synthesis is table stakes; this is about everything after it.
How to use interview data to build product roadmaps
A practical five-step system for turning customer interviews into a roadmap: collect traceable evidence, synthesize with a citation floor so weak themes drop, rank survivors into bets, hand off a build pack, and learn outcomes back into memory.
Read →Why most customer interviews fail to move roadmaps (and how to fix it)
Customer research rarely fails because teams collect too little evidence. It fails because n-of-1 themes get promoted to patterns, the loudest customer sets priorities, the readout goes stale in a deck, and nothing links the interview to what you actually shipped. Here is the structural fix.
Read →The citation floor: why two quotes beat one in customer interview synthesis
Cited themes are table stakes. NotebookLM and coding agents do it free. The real discipline in customer interview synthesis is the floor: drop any theme with fewer than two verified quotes, then act on what survives.
Read →How to write a PRD for a coding agent (and why the handoff is the moment that matters)
A PRD for a coding agent is a different artifact than the doc you wrote for humans. Here's what goes in a Build Pack, and why the paste into Cursor is the irreversible handoff that decides whether the build matches the evidence.
Read →Product decision memory: the graph an ephemeral agent can't keep
Teams measure everything and remember nothing. Product decision memory is the durable graph linking evidence to bets to ships to outcomes. The thing an ephemeral agent session structurally cannot hold.
Read →Feature Bets: feature prioritization with evidence, not scores
Most roadmaps are bets dressed up as decisions. Feature Bets make prioritization defensible by attaching cited evidence. Problem, audience, confidence, expected outcome, counter-evidence, and risks, to every item you rank.
Read →Evidence-based PRDs: keep the customer quote attached to every requirement
Requirements decay the moment they're separated from the customer evidence behind them. Here's how an evidence-based PRD keeps the "why". The verified quote. Attached to every requirement and task so the rationale survives.
Read →Continuous product discovery: the full Klyr loop, end to end
The full Klyr loop, end to end: customer interviews become cited themes, themes become a ranked Feature Bet, the bet becomes a Build Pack your coding agent can run, and the shipped outcome learns back into product memory. Continuous product discovery that actually compounds.
Read →