Flow predictability
Also known as Delivery predictability, Predictable flow
Definition
Flow predictability is the degree to which completed work matches a defined forecast, range, or service expectation over repeated observations. It concerns consistency, not simply speed.
Consistency is the point
A fast system that sometimes takes one day and sometimes takes three weeks may be difficult to plan around. Flow predictability asks whether a defined class of work tends to finish within a useful range. The range, population, and finish boundary must be explicit.
A concrete example
An illustrative team completes 85 of 100 routine changes within its stated service expectation. That result says more about predictability than the team's fastest change. The 15 exceptions deserve inspection by age, dependency, size, and failure mode.
Improve the distribution
Reduce avoidable variation by limiting WIP, clarifying entry rules, shrinking batches, or addressing bottlenecks. Do not narrow the range by excluding every difficult item. Segment genuinely different service classes instead.
Limitations
Predictability does not prove value or quality. A team can reliably deliver the wrong work. It also does not imply that all work should share one expectation.
How this relates to Weave
Weave can help teams inspect delivery distributions and the changes behind late or early outcomes. Use those patterns to improve system policies, not to score individual engineers, and keep the service expectation defined outside analytics.
Explore Engineering intelligence