Batch size
Also known as Work batch size, Change batch size
Definition
Batch size is the amount of work grouped into one processing, review, release, or handoff unit. The unit may be a change, pull request, story, deployment, or set of requests.
Grouping changes the queue
Batch size affects how much work must be understood, tested, approved, and recovered together. A large batch can make each handoff expensive and can keep downstream capacity occupied for longer. A smaller batch can reduce waiting, but splitting work without preserving a coherent behavior can add coordination overhead.
A concrete example
Imagine a team that reviews one 40-file change every Friday. The review queue is empty on Monday but overloaded at the end of the week. If the same behavior can be delivered as four coherent changes during the week, reviewers may get earlier feedback and a defect may have a narrower search area.
Choose the unit deliberately
Lines changed are only one possible measure. A batch can be counted as commits, pull requests, deployable slices, or customer-visible outcomes. State whether generated files and formatting changes are included. Compare like with like when looking at trend data.
Limitations
Small batches are not automatically safer. Database migrations, interface changes, and coordinated launches may require a larger unit. Measure quality, rework, and operational outcomes alongside flow time.
How this relates to Weave
Weave helps teams inspect change scope, review activity, and delivery timing so they can study how batches behave in practice. Use those observations to evaluate a process experiment, not to impose a universal size quota on individual engineers.
Explore Engineering intelligence