Why most product reviews fail
Two failure modes dominate. The first is the demo showcase: the team shows what it built, everyone nods, and nobody asks whether the metric moved. The second is the metrics dump: forty charts, no decision, and a shared sense that the meeting was thorough. Both feel productive and neither changes what happens next month.
The fix is structural. Fix the blocks, fix the order, and require that the meeting produce written decisions.
The six blocks
- Outcomes since last review (15 min). What shipped, the metric it was supposed to move, and what happened. One line each. No demos in this block.
- Adoption and retention read (15 min). Activation, adoption, and retention by segment. Only numbers that could change a decision.
- Customer evidence (15 min). What was learned from customers this month and which roadmap assumption it strengthened or killed.
- Prioritization changes (20 min). What entered the roadmap, what left, and the trade-off said out loud. Nothing enters without something leaving or a date moving.
- Risks and dependencies (15 min). Technical, commercial, and — where relevant — AI feasibility risks, each with an owner and a next check date.
- Decisions and owners (10 min). Written decisions, owners, review dates. Circulated within a day.
The two blocks teams cut first
Under pressure, customer evidence and the decisions block are the first to go — and they are the two that make the meeting worth holding. Cutting evidence turns the roadmap back into internal opinion. Cutting the decisions block means the discussion evaporates and the same topics return next month.
If time is genuinely short, shorten blocks two and five instead. Never delete blocks three or six.
Running it for the first time
The first review is usually uncomfortable, because block one exposes that several releases had no agreed metric. That is the finding, not a failure of the meeting. Fill the gap forward: from this month on, nothing significant enters the roadmap without a primary metric and a review date attached.
Keep the invite small — product, engineering, and one commercial leader. Reviews that grow past eight people stop producing decisions and start producing performances.