Review Readiness Criteria
A stated bar a design must meet before review, so the session is spent on judgement rather than on discovering that the work is not ready.
Architecture reviews fail in predictable ways: the material arrives in the meeting, the reviewers have no shared basis for judgement, and the outcome depends on who spoke loudest.
Readiness criteria fix the first failure. A design should not be reviewed until it states: the problem and its evidence, the quality attributes that drive it as concrete scenarios, the options considered, the recommendation with its consequences, and the open questions. Material circulated at least a day in advance.
The second failure is fixed by reviewing against something: the prioritised quality-attribute scenarios, the published architecture principles, and the fitness functions. A review with no criteria is a discussion of preferences, and it will be won by seniority.
Three practices that change the character of these sessions:
Advisory rather than approving, wherever possible. Review as a service teams want because it improves their design, rather than a gate they must pass, produces earlier engagement and better outcomes.
Early and small. A review at 20% completion can change the design. A review at 90% can only approve it or waste the work, and everyone knows it.
A written outcome with decisions, actions and owners — including the concerns raised and deliberately accepted, which is what stops the same points recurring at the next review.