Architecture decision latency
Also known as Architecture decision latency, Architecture decision latency in software engineering
Definition
The elapsed time between an architecture question being raised and a decision becoming usable for implementation.
What it means
Architecture decision latency measures waiting in the decision process, including discovery, review, approval, and communication. Long latency can block work or cause teams to make local choices that later conflict. The start and end events need to be observable and agreed upon.
Example
A team waits nine business days for an API boundary decision, then spends a day reworking a parallel implementation. Tracking the queue helps leaders improve decision availability without bypassing necessary review.
Limitations
A short decision is not necessarily a good decision, and complex choices deserve time. Ticket timestamps can miss informal discussions or decisions made outside the recorded process. Combine the measure with reversal and rework evidence.
How to use the signal
Record when the question became actionable and when the decision was communicated, not only when a meeting occurred.
Weave can help inspect decision-related queue time and follow-on rework when those events are represented in engineering work. Keep the decision system as the source of truth.
How this relates to Weave
Weave can help inspect decision-related queue time and follow-on rework when those events are represented in engineering work. Keep the decision system as the source of truth.
Explore Engineering intelligence