Notes RSS feed

Analytics excel at describing behavior at scale. They are nearly useless at explaining motive. A retention chart that moves eight points is a signal that something changed, not a story about what broke or delighted users. Diagnosis still lives in conversation: interviews, support, session replay with context, the messy qualitative layer dashboards cannot replace. Teams that optimize only on metrics get very good at moving numbers and very bad at understanding people. Pair the chart with a human sentence, or you are treating symptoms while the underlying problem mutates.

Tech debt is rarely an engineering surprise. It is the residue of tradeoffs someone accepted under pressure: ship now, cut scope, defer the hard path, say yes to a partner deadline. Product owns those choices even when they were implicit or invisible on a slide. Pretending debt is “their problem” is how teams slow down and blame each other instead of the original bargain. Own the tradeoff that created the shortcut, and you can sequence paydown like any other bet. Ignore that lineage, and you will fight the same fire every quarter.

The MVP is the smallest artifact that lets you test whether a specific assumption holds. It is not a stripped-down version of the product you already decided to build. If you ship something and learn nothing measurable about the bet you were trying to validate, you did not run an experiment—you shipped unfinished work and called it discipline. The bar is learning, not launch. Name the assumption before you name the build; otherwise you are optimizing for output, not for truth.

A roadmap should read like a portfolio of bets, not a scroll of promises. Each item is a guess about what will move the outcome you care about, backed by whatever evidence you have today. The instant stakeholders treat the roadmap as a contract, you stop using it to learn and start using it to negotiate. Strategy needs room to change when reality disagrees with the deck. Treat dates and themes as working hypotheses, and you keep the tool honest. Treat them as guarantees, and you have traded clarity for politics.

The most important product decisions are the ones you refuse to make. Every yes is a calendar line, a headcount, and a maintenance burden you cannot unwind without pain. The roadmap that impresses executives is often the one that looks busy; the one that ships great product is the one that stayed narrow. Great PMs earn their reputation for what they kept off the roadmap, not for how long the list became. If you cannot say no, you are not prioritizing—you are collecting requests. Practice the sentence until it feels boring: not now, not us, not this way.