Before After Coverage
Before / After Coverage
A lightweight way to show impact. Capture a snapshot before you apply a technique or pattern, then an after snapshot once you’ve updated scenarios/cases/gates. Keep it to one page so reviewers and stakeholders can scan it in < 3 minutes.
How to use this page
- Pick a feature (e.g., checkout discount code) and note the baseline commit/branch.
- Run the relevant coverage checklists in
60-checklists/*
and record results. - Apply one or two techniques (e.g., Boundary & Equivalence + MAE flows).
- Update scenarios/cases and re-run gates.
- Fill in the After snapshot and summarize outcomes (defects found, gates passing, time saved).
Link this note from your PR description so reviewers see the improvement.
Snapshot Template
Copy below into a new file under
05-field-notes/before-after-coverage-<feature>.md
and fill the blanks.
# Before / After — <Feature/Flow>
**Owner:** <name/role>
**When:** <YYYY-MM-DD>
**Baseline:** <commit/branch/tag>
**After:** <commit/branch/tag>
**Scope:** <area, endpoints, UI screens, roles>
## Before (Baseline)
- **Scenarios:** <#> (Main: <#>, Alt: <#>, Exception: <#>)
- **Cases:** <#>
- **Checklists/Gates:**
- Functional: <X>/<Total> pass
- API: <X>/<Total> pass
- Non-functional (perf/security/compat): <brief summary>
- **Gaps observed:** <bullets>
## Change Applied
- Techniques/Patterns: <e.g., Boundary & Equivalence, Error Taxonomy → UX>
- Files updated: <links to 20-*/30-*/40-*/60-*/70-*/>
- Notes: <short rationale, assumptions>
## After
- **Scenarios:** <#> (Main: <#>, Alt: <#>, Exception: <#>)
- **Cases:** <#>
- **Checklists/Gates:**
- Functional: <X>/<Total> pass
- API: <X>/<Total> pass
- Non-functional: <brief summary>
- **Evidence:** <links to artifacts: logs, API transcripts, screenshots>
## Outcomes
- Defects found/avoided: <# + short description>
- Time saved in review/regression: <estimate>
- Remaining risks / Follow-ups: <bullets>
Example — Checkout Discount Code
Owner: QA (Solo)
When: 2025-09-15
Baseline: feature/discount-code @ 7a1b2c3
After: feature/discount-code @ 9d4e5f6
Scope: Web checkout (apply code), API /discount/verify
, roles: Guest, Signed-in
Before (Baseline)
- Scenarios: 4 (Main: 1, Alt: 1, Exception: 2)
- Cases: 11
- Checklists/Gates:
- Functional: 7/12 pass (missed: max length, charset mix, expired code)
- API: 4/9 pass (no idempotency check on verify; missing error taxonomy → UX mapping)
- Non-functional: perf budget undefined
- Gaps observed:
- No boundary tests for code length (min/max/overflow)
- No i18n (emoji/Chinese/accents) input variation
- Error codes not mapped to user messages; duplicates in support tickets
Change Applied
- Techniques/Patterns: Boundary & Equivalence, MAE flows, Error Taxonomy → UX
- Files updated:
20-techniques/boundary-and-equivalence.md
(worked example added)30-scenario-patterns/main-alt-exception.md
(exception cases expanded)40-api-and-data-contracts/error-taxonomy.md
(codes + messages)60-checklists/functional-coverage.md
(length/charset items)70-mini-projects/checkout-discount-code/*
(scenarios & cases)
- Notes: Added ±1 around length bound, mixed-charset values, and explicit expected results per error.
After
- Scenarios: 7 (Main: 1, Alt: 2, Exception: 4)
- Cases: 20
- Checklists/Gates:
- Functional: 12/12 pass
- API: 8/9 pass (remaining: pagination consistency on list codes)
- Non-functional: set p95 ≤ 500ms for verify; micro-check added in CI
- Evidence:
- API transcript for
VALIDATION.code.length.exceeds
→ “Enter a code of at most 16 characters.” - Logs with
correlation_id
anderror_code
per failed attempt - CI job link for perf check
- API transcript for
Outcomes
- 3 pre-release defects found (overflow, unsupported charset, wrong UX copy)
- Review time −15 min/PR (cases shared and reusable)
- Support tickets for “code not working” down ~25% in the next release
- Follow-ups: add cursor pagination for admin list; extend role matrix to Staff/Admin
Minimal CSV for tracking (optional)
You can paste this into a sheet or export it for dashboards.
date,feature,baseline,after,scenarios_before,cases_before,scenarios_after,cases_after,functional_pass,api_pass,notes
2025-09-15,checkout-discount,7a1b2c3,9d4e5f6,4,11,7,20,12/12,8/9,"added boundary & equivalence, error taxonomy→UX; set p95≤500ms"
Tips
- Keep “Before/After” diffs small and focused (one or two techniques).
- Prefer observable expected results (messages, codes, UI state) over internal assertions.
- Link artifacts (logs/metrics/traces) for faster review.
- Add a single metric (defects found, gate passes, time saved) to prove value.
If your note helps others, consider converting it into a longer field note or case study using _templates/field-note.md
or _templates/case-study.md
.