Engineering change classification
Also known as Software change classification, Engineering work classification
Definition
Engineering change classification is the practice of assigning software changes to defined work categories such as feature, defect, maintenance, incident recovery, security, or enablement. It helps teams interpret flow, quality, capacity, and outcomes in the context of why the change was made.
Categories explain the demand on the factory
Feature work, defect fixes, maintenance, incident recovery, security work, and platform enablement can all produce code. They do not have the same purpose, urgency, risk, or outcome. Blending them into one throughput number can make a team look slower during a migration or make recovery demand disappear inside a feature count.
Choose categories that support a real decision. Keep the taxonomy small enough to apply consistently. Define how a mixed change is handled, which source owns the label, and whether classification is assigned at the pull request, work item, deployment, or outcome level.
Validate the classification
Automated inference can provide a useful starting point, but labels should be sampled and corrected. A change that looks like maintenance may be a reliability investment. A feature branch may contain a defect fix. Record confidence and preserve the underlying examples when the category drives planning or comparison.
Read work mix with throughput, quality, rework, reliability, and outcomes. Do not treat one category as intrinsically more valuable than another.
How Weave can help
Weave can make classifications visible alongside the changes and workflow evidence they summarize. Teams can inspect whether a work-mix shift reflects real demand, a labeling change, or a different delivery boundary.
How this relates to Weave
Weave can help classify the work represented in pull requests and inspect the code output, review, quality, and delivery context for each category. Classification should be checked against product, incident, and team records when the label affects a factory decision.
Explore Engineering intelligence