Template — print to PDF or copy into your doc tool
Quarterly review: cost per accepted change
A single-page review structure for the recurring leadership conversation. Replace the bracketed placeholders with your numbers. Print to PDF, or paste the structure into Notion / Confluence / Google Docs.
Print this page Download tracker (XLSX) →
[Team or organization name] — Quarterly review
| Quarter | [Q1 2026] |
| Reporting date | [YYYY-MM-DD] |
| Owner | [Name, role] |
| Sponsor | [Name, role] |
Executive summary
- [One sentence on the headline number and trend direction.]
- [One sentence on the most important driver of the move.]
- [One sentence on the proposed next-quarter focus or ask.]
The number
| This quarter | Prior quarter | One year ago | |
|---|---|---|---|
| Cost per accepted change | [$XXX] | [$XXX] | [$XXX] |
| Δ vs prior quarter | [±X%] | — | — |
| Δ vs one year ago | [±X%] | — | — |
[One sentence interpreting the trend: improving, stable, or degrading. Avoid drawing strong conclusions from a single window.]
Component breakdown
| Cost component | This quarter | % of total | Δ vs prior |
|---|---|---|---|
| Model cost | [$XXX] | [XX%] | [±X%] |
| Infrastructure cost | [$XXX] | [XX%] | [±X%] |
| Engineering time | [$XXX] | [XX%] | [±X%] |
| Review cost | [$XXX] | [XX%] | [±X%] |
| Rework cost | [$XXX] | [XX%] | [±X%] |
| Total cost | [$XXX] | 100% | [±X%] |
| Accepted change units | [NN] | — | [±X%] |
What changed
[Identify the one or two cost components that drove the movement in the bottom line. Keep this descriptive, not interpretive. Example phrasing: "Rework cost rose 38% quarter over quarter, driven by reverts on the agent-generated migrations in service X. Model cost rose 12%; all other components were flat."]
Leading indicators
| Indicator | This quarter | Prior quarter | What it says |
|---|---|---|---|
| Change failure rate (DORA) | [X%] | [X%] | [One sentence — is the "stayed there" denominator holding?] |
| AI tool acceptance rate | [X%] | [X%] | [One sentence — are developers still finding the tooling useful?] |
| DevEx pulse (or equivalent) | [score] | [score] | [One sentence — is verification load degrading the team?] |
Hypothesis and next-quarter focus
Hypothesis: [One paragraph describing the most likely explanation for the trend. Be explicit that this is a hypothesis, not a conclusion.]
Proposed focus for next quarter: [One concrete commitment — an experiment, a tooling change, a review-process intervention, or "hold steady and re-measure." Avoid setting a CPAC target; targets invite gaming.]
What we are not doing: [One sentence on a tempting intervention you are deliberately deferring, and why.]
Methodology notes
| Window length | [Quarter — same as all prior windows] |
| Window dates | [YYYY-MM-DD to YYYY-MM-DD] |
| Hourly rate used (engineering) | [$XXX/hr, fully loaded] |
| Hourly rate used (review) | [$XXX/hr, fully loaded] |
| "Change" unit | Merged PR to production branch, size-normalized via the 500-LOC rule |
| Exclusions | Vendored code, generated code, lockfiles, [additional team-specific exclusions] |
| Model cost attribution method | [Aggregate billing / per-team tagging / per-commit attribution via tooling] |
| Rework attribution method | [Revert commits + follow-up fixes identified in git history] |
| Changes to accounting this quarter | [None — or note the change and the resulting one-time effect on comparability] |
Appendix: backup data and links
- Tracker spreadsheet: [link]
- Source data / queries: [link]
- Prior quarter's review: [link]
- Definition reference: costperacceptedchange.org
How to fill this in
- Pull this quarter's numbers from your tracker spreadsheet.
- Compute the prior-quarter and one-year-ago comparisons from the same tracker.
- Fill in the component breakdown with percentages.
- Describe what changed in one paragraph. Stay descriptive; save interpretation for the hypothesis section.
- Pull leading indicators from your existing instruments (DORA dashboard, AI tool admin panel, DevEx survey).
- Write the hypothesis and the next-quarter commitment. One paragraph each. Be specific.
- Update the methodology notes only if something actually changed; consistency across quarters is what makes the trend comparable.
- Print to PDF or paste the structure into your team's doc tool of choice.
The discipline that matters
- Stay descriptive in the "what changed" section. Reserve interpretation for the explicit hypothesis section. This keeps the conversation honest.
- Always include "what we are not doing." The temptation to react to a single moving number is the most common failure mode of metric-driven leadership.
- Do not set CPAC targets. Targets are gamed. Track the trend; investigate when it moves; commit to interventions, not to numbers.
- Note methodology changes explicitly. Any change to hourly rates, exclusions, or unit definitions creates a one-time discontinuity. Document it so future readers can interpret the trend correctly.
See the how-to-use page for the full reasoning behind these constraints.