An organisation makes architectural decisions slowly and revisits them repeatedly. What structural changes improve decision quality and speed together?
Show the full answer Hide the answer
Why decisions are slow and repeatedly revisited
Usually three causes, and they compound:
1. Unclear decision rights. Nobody knows whose call it is, so everyone consults everyone and the decision waits for a consensus that is not required.
2. No recorded rationale. A decision made six months ago is re-argued because nobody can say what was considered, what was rejected, or under what constraints — so the same conversation recurs with the same outcome.
3. Every decision treated as equally weighty. A reversible choice receives the same deliberation as an irreversible one, which is expensive for the first and, worse, means the second gets no more attention than it deserves.
The changes
1. Classify by reversibility before deliberating. A choice that can be undone in a sprint deserves a decision and a measurement, not a review. A choice that would take a year to reverse deserves genuine analysis. Spending equal effort on both is the most common cause of slow decisions and of bad irreversible ones.
2. Name the decision-maker explicitly. Not the group. Someone accountable, who consults and then decides — which converts an unbounded consensus process into a bounded consultation.
3. Record decisions with rejected options and a revisit condition. The rejected options prevent re-litigation; the revisit condition turns a permanent decision into one with a checkable trigger.
4. Agree criteria before comparing options. Half of stalled debates end here, because participants are optimising for different things and neither has said so.
5. Make disagreement explicit and then commit. The person who did not prevail says so on the record and supports the decision. Suppressed disagreement resurfaces at the first difficulty.
The measure
Not the number of decisions, but how many are re-opened without new information. A decision revisited because a trigger fired is the system working. One revisited because nobody remembers why it was made is a documentation failure — and it is the dominant cost in most organisations.