Roadmaps as bets with review dates

What it is

A roadmap usually reads as a list of features with quarters attached. Written as bets, each line states three things: the payout (what is true for users or the business if it works), the appetite (how much time the team is willing to spend, as a cap rather than an estimate), and a review date at which the bet is judged. Shape Up describes the framing: bets have a payout, bets are commitments for the duration of a cycle with no interruptions, and a smart bet has a cap on the downside, so the most that can be lost is the cycle. Its circuit breaker rule adds that work which does not ship within the bet is not extended by default; it goes back to shaping. The review date generalises this to teams that do not run six-week cycles.

Why it matters

An open-ended roadmap item is never wrong: it is always "in progress" or "next quarter". A bet with a review date forces an explicit continue, re-scope or stop decision, which is where roadmaps create value. Framing work as a bet also states the uncertainty honestly to stakeholders instead of implying a delivery promise.

How to apply

• Keep the list short: a few bets per team per period, and only one period ahead; Shape Up argues that betting only one cycle ahead keeps options open for whatever comes up.
• For each bet write: the problem, the payout you expect and how you would notice it, the appetite, the review date, and what would make you stop early.
• Put the review on the calendar when the bet is placed, with the people who placed it. At the review, compare against the written payout, not against the current enthusiasm.
• Record the outcome next to the bet: continued, re-scoped (how), stopped (why). The history is the team's calibration data.
• Treat requests that arrive mid-bet as candidates for the next betting round, not as interruptions to the current one, unless they are genuine crises.

Pitfalls

A "bet" with no cap is a commitment in disguise. Review dates that are silently moved defeat the mechanism; move them only with a recorded reason. Payouts phrased as outputs ("ship the export feature") cannot be judged; phrase them as effects ("support tickets about exports drop", "three pilot customers use it weekly"). Shape Up's specifics (six-week cycles, a betting table with the CEO and CTO at it) are one company's practice, not a requirement of the framing.