Product Learning Lab
Skill · Product judgment

Prioritizing under uncertainty

Decide what to build next when the evidence is partial and the cost of waiting is real.

Waiting for certainty is a decision too — usually the most expensive one, because you pay for it without noticing.

The real problem

You will never have enough evidence

The seductive move under uncertainty is to gather more data. Sometimes that's right. Often it's a way to feel responsible while the window closes — a competitor ships, the quarter ends, the team idles waiting for a signal that a week of research won't sharpen.

Prioritizing under uncertainty is not about becoming certain. It's about asking a different question: given what I don't know, what's the cheapest move that teaches me the most, and what does it cost me to be wrong? Good PMs don't have better forecasts. They have a better sense of which bets are cheap to reverse.

The method

Deciding without certainty

  1. 01

    Separate the reversible from the one-way door

    Amazon's frame: reversible decisions should be made fast and cheap; irreversible ones earn the extra week. Most decisions are more reversible than they feel.

  2. 02

    Price the delay

    What does waiting a month actually cost — lost learning, a competitor's head start, team morale? If waiting is free, wait. It rarely is.

  3. 03

    Find the cheapest test of the riskiest assumption

    Not the whole feature — the one belief that, if false, kills it. A fake door, a concierge flow, five customer calls.

  4. 04

    Size the downside, not just the upside

    If this bet is wrong, what's the blast radius? Bets with capped downside deserve a lower evidence bar.

  5. 05

    Decide, and name what would change your mind

    Commit — but write down the signal that would reverse the call. That's what separates a decision from a guess.

The shift

Chasing certainty vs pricing the bet

Waiting for proof

Evidence as a gate.

  • "Let's research one more sprint."
  • Reversible and one-way doors treated the same
  • Cost of delay ignored
  • Confident, late, sometimes irrelevant

Pricing the bet

Evidence as a purchase.

  • "What's the cheapest thing that teaches us?"
  • Speed matched to reversibility
  • Delay is a line item
  • Roughly right, in time, and adjustable

Core insight

The question is never 'am I sure?' It's 'is this bet cheap to reverse, and do I know what would tell me I'm wrong?' Certainty is expensive; reversibility is usually cheaper than it looks.
Reflect

Recall something your team delayed to 'get more data.' When the data arrived, did it change the decision — or just confirm what a cheap test would have shown weeks earlier?

Takeaway

Match your speed to how reversible the decision is, buy information only where it's cheap and decisive, and always name the signal that would change your mind. Deciding well under uncertainty beats deciding late with confidence.

Keep going
Skill · Product judgment

Reading tradeoffs

Make the exchange behind a decision legible.

Concept

Opportunity sizing

Estimate the prize before you bet on it.

Skill · AI

Designing for model uncertainty

Ship when the system itself can't be certain.