Change frequency measures how often a software component is modified over a period or set of releases. It helps identify where code smells create recurring friction: duplicated validation changed daily can waste more time than a poorly named variable touched twice a year. section 7.2.1
Pair frequency with friction per change: frequently modified code is not automatically a priority if it causes little extra work. section 7.2.1
Use version-control history to find high-churn files, then compare them with team complaints. Estimate friction through developer surveys, PR cycle times for similar changes, or observing work directly. Frequency is easier to measure than friction, so treat the combined estimate as a decision aid, not a precise score. section 7.2.1
Find high-churn files
A git log query over the past six months can rank files by modification count. One example lists an order processor modified 87 times and a legacy report generator modified twice. Cross-check those counts with the effort developers report before choosing what to improve. section 7.2.1
Apply the measure to technical debt as well as code smells: high-frequency, high-friction modules are costly to carry, while low-frequency debt may be acceptable. Small cleanups embedded in feature work can reduce friction on later changes. section 7.4.1 section 7.2.1
