Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Design, audit, and improve analytics tracking systems that produce reliable, decision-ready data. Use when the user wants to set up, fix, or evaluate analytics tracking (GA4, GTM, product analytics, events, conversions, UTMs). This skill focuses on measurement strategy, signal quality, and validation— not just firing events.
.claude/skills/dokhacgiakhoa-analytics-tracking/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 16 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 49% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 36% | 0% |
You are an expert in analytics implementation and measurement design. Your goal is to ensure tracking produces trustworthy signals that directly support decisions across marketing, product, and growth.
You do not track everything. You do not optimize dashboards without fixing instrumentation. You do not treat GA4 numbers as truth unless validated.
Before adding or changing tracking, calculate the Measurement Readiness & Signal Quality Index.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 25,234 | 43,626 | +73% | 1 | 1 | 0% | 4,626 | 6,897 | +49% | 0 | 0 | — |
case-02 | fail→pass | 30,073 | 37,584 | +25% | 1 | 1 | 0% | 6,734 | 6,870 | +2% | 0 | 0 | — |
case-03 | fail→pass | 46,400 | 38,632 | -17% | 1 | 1 | 0% | 8,246 | 8,158 | -1% | 0 | 0 | — |
case-04 | pass→pass | 16,300 | 16,267 | -0% | 1 | 1 | 0% | 2,727 | 3,717 | +36% | 0 | 0 | — |
case-05 | pass→pass | 13,417 | 20,489 | +53% | 1 | 1 | 0% | 2,307 | 4,224 | +83% | 0 | 0 | — |
case-06 | pass→pass | 16,952 | 14,631 | -14% | 1 | 1 | 0% | 2,592 | 3,293 | +27% | 0 | 0 | — |
case-07 | pass→pass | 15,580 | 17,341 | +11% | 1 | 1 | 0% | 2,736 | 3,690 | +35% | 0 | 0 | — |
case-08 | pass→pass | 12,826 | 16,516 | +29% | 1 | 1 | 0% | 2,243 | 3,751 | +67% | 0 | 0 | — |
case-09 | pass→pass | 15,049 | 15,417 | +2% | 1 | 1 | 0% | 2,385 | 3,229 | +35% | 0 | 0 | — |
case-10 | pass→pass | 13,736 | 17,657 | +29% | 1 | 1 | 0% | 2,321 | 3,464 | +49% | 0 | 0 | — |
case-11 | pass→pass | 11,673 | 16,395 | +40% | 1 | 1 | 0% | 1,982 | 3,503 | +77% | 0 | 0 | — |
case-12 | pass→pass | 9,366 | 12,942 | +38% | 1 | 1 | 0% | 1,670 | 2,610 | +56% | 0 | 0 | — |
case-13 | pass→pass | 10,263 | 9,848 | -4% | 1 | 1 | 0% | 1,887 | 2,472 | +31% | 0 | 0 | — |
case-14 | pass→pass | 9,829 | 14,865 | +51% | 1 | 1 | 0% | 1,633 | 2,989 | +83% | 0 | 0 | — |
case-15 | fail→pass | 17,312 | 17,881 | +3% | 1 | 1 | 0% | 2,808 | 3,559 | +27% | 0 | 0 | — |
case-16 | pass→pass | 18,543 | 23,870 | +29% | 1 | 1 | 0% | 2,575 | 4,541 | +76% | 0 | 0 | — |
case-17 | pass→pass | 21,195 | 23,676 | +12% | 1 | 1 | 0% | 3,334 | 4,342 | +30% | 0 | 0 | — |
case-18 | pass→pass | 18,234 | 24,274 | +33% | 1 | 1 | 0% | 3,051 | 5,054 | +66% | 0 | 0 | — |
case-19 | pass→pass | 15,917 | 19,340 | +22% | 1 | 1 | 0% | 2,569 | 3,768 | +47% | 0 | 0 | — |
case-20 | pass→pass | 15,583 | 15,540 | -0% | 1 | 1 | 0% | 2,678 | 3,441 | +28% | 0 | 0 | — |
case-21 | pass→pass | 16,626 | 24,342 | +46% | 1 | 1 | 0% | 2,655 | 4,646 | +75% | 0 | 0 | — |
case-22 | pass→pass | 16,380 | 27,825 | +70% | 1 | 1 | 0% | 3,318 | 5,192 | +56% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.