mattgierhart/PRD-driven-context-engineering / .claude/skills/prd-v08-drift-baseline-compare
prd-v08-drift-baseline-compare
VerifiedEstablish baseline → snapshot → compare → history monitoring for any KPI, config, or metric that can drift during PRD v0.8 Deployment & Ops. Triggers on requests to monitor drift, baseline a value for later comparison, or when user asks "how do we track if X changes?", "baseline this", "config drift", "performance regression", "metric drift", "compare to last week", "is this getting worse?". Outputs MON-DRIFT-* entries with baseline + comparison rules.
Drift: Baseline / Compare / History
Position in workflow: v0.8 Monitoring Setup → v0.8 Drift: Baseline / Compare / History → v0.8 Runbook Creation
Execution Mode
Default is standard. See .claude/rules/08-skill-execution-modes.md for selection logic.
| Mode | What this skill produces |
|---|---|
| quick | One metric / config / dataset baselined; weekly compare schedule; simple threshold alert |
| standard | 3–5 things monitored; baseline + tiered thresholds (warn / critical); compare cadence; history retention |
| deep | Full portfolio; multi-dimensional comparison (per segment, per environment); regression-cause matrix; auto-baselining after intentional change |
What This Does
Generalizes a pattern AgriciDaniel's claude-seo skill encodes for SEO drift — baseline → snapshot → compare → history — into a reusable monitoring shape that works for any metric, config, or dataset that can change over time and needs to be watched.
This is drift monitoring, distinct from alerting on absolute thresholds. Alerting answers "is X over the line right now?" Drift monitoring answers "is X different from last week's normal?" — which catches slow regressions that absolute thresholds miss.
Examples of things worth drift-monitoring:
- KPI: activation rate week-over-week
- AI search position: ChatGPT/Perplexity ranking for target queries
- Config: feature-flag rollout percentages
- Performance: p95 latency by endpoint
- Cost: per-user infra cost
- Marketing: per-channel CAC trend
- Content: changelog post engagement
- Third-party: vendor pricing pages (price hikes), competitor feature pages (parity loss)
How It Works
- Pick what to monitor — One thing per MON-DRIFT- entry. Must be:
- Quantifiable (number, percentage, list, configuration value)
- Snapshotable (captureable at a point in time, ideally automatically)
- Causally interpretable (when it changes, you know enough to investigate)
- Capture the baseline — Take a snapshot. Date it. Store in version control or a known location (
status/baselines/,monitoring/snapshots/, etc.). - Define drift thresholds:
- Warn: meaningful change (e.g., 10% drift in a KPI; any change in a config value)
- Critical: serious change (e.g., 25% KPI drop; breaking config change)
- Recalibrate: intentional change that should refresh the baseline (e.g., after a feature rollout, the baseline is wrong; refresh it)
- Set compare cadence — How often does this get re-snapshotted?
- Hot (hourly/daily): production KPIs, AI search positions during a launch
- Warm (weekly): standard product KPIs, content engagement
- Cool (monthly): vendor pricing, competitor feature parity, infra cost
- Build the compare procedure — A script or runbook that:
- Takes a new snapshot
- Diffs against baseline
- Computes drift % per dimension
- Emits warn/critical signals at thresholds
- Appends to history log
- Plan auto-baselining after intentional change [standard+] — When the team makes a deliberate change (ships a feature that should improve activation), the old baseline becomes wrong. Define what triggers a baseline refresh and who approves it.
Example
Monitoring AI search position for the target query "best CRO tool for SaaS founders" across 3 surfaces. (Generalized from AEO Audit re-test cadence.)
Baseline (2026-04-01):
ChatGPT: not mentioned (rank: --)
Perplexity: rank 4 of 5, miscategorized as "analytics"
AI Overviews: not mentioned
Sources cited: [competitor1.com, competitor2.com, reddit.com/r/SaaS]
Thresholds:
- Warn: rank changes by ≥1, sources list changes, miscategorization persists
- Critical: dropped from any surface where previously mentioned
Compare cadence: weekly during launch month, monthly after.
Re-snapshot (2026-04-15):
ChatGPT: rank 3 of 5 (NEW) ← improved
Perplexity: rank 3 of 5, category corrected (NEW) ← improved
AI Overviews: rank 5 of 7 (NEW) ← improved
Sources cited: [+ ourdomain.com (NEW), ...]
Compare output: 3 surfaces improved. Likely cause: alternatives-pages campaign + new CRO-anchored blog post indexed.
History append: 2026-04-01 → 2026-04-15 row added to history log.
Action: No alert; positive drift. Baseline can stay (next compare in 2 weeks). If the trend reverses, the baseline catches it.
What You Get Back
- MON-DRIFT-* entries (one per monitored thing) — Baseline snapshot, thresholds, cadence, compare procedure, history pointer
- History logs in
status/drift-history/<id>.jsonl(or similar) — Append-only record of compare results - Compare scripts (
scripts/drift/<id>.shor equivalent) when automatable
When to Use It
| Trigger | Mode |
|---|---|
| Post-launch — need ongoing watch on KPIs | standard |
| AEO Audit results to maintain | standard |
| Competitor feature parity tracking | standard |
| Vendor / third-party pricing watch | quick |
| Performance regression hunting | deep |
| Infrastructure cost trend | quick |
Do not use for things that are already covered by threshold-based alerting (RED/USE metrics, MON-* alerts). Drift is for trends; alerting is for absolutes.
Consumes
- MON-* monitoring infrastructure (from v0.8 Monitoring Setup) — Data sources for KPI/performance drift; existing dashboards
- KPI-* metric definitions (from v0.3 + v0.9) — What's worth watching
- CFD-* competitor research (from v0.2 + ongoing) — For competitor-parity drift
- GTM-AEO-* Coverage Matrix (from v0.9 AEO Audit) — For AI search position drift
- BR-PRICING-* (from v0.3 + v0.9 Offer) — For pricing drift on our side; vendor-pricing drift watches their side
- DEP-* release entries — Intentional changes that may trigger baseline refresh
Produces
- MON-DRIFT-* entries with
Type=Drift-Baseline— One per monitored thing - History logs in
status/drift-history/— Append-only diff history - Compare scripts in
scripts/drift/(when automatable) - Baseline-refresh approvals — Recorded when an intentional change resets a baseline
Output Template
MON-DRIFT-XXX: Baseline — [What is being monitored]
Type: Drift-Baseline
Category: [KPI | AI-Search | Config | Performance | Cost | Vendor | Competitor]
Owner: [Person / role]
Status: Active
Subject: [Specific thing — e.g., "ChatGPT rank for query 'best CRO tool for SaaS'"]
Baseline (YYYY-MM-DD):
[Snapshot — values, list, config dump, etc.]
Thresholds:
Warn: [Condition — e.g., "rank change ≥1 OR sources list changes"]
Critical: [Condition — e.g., "dropped from any surface previously mentioned"]
Recalibrate-trigger: [What intentional change resets the baseline]
Compare cadence: [Hourly | Daily | Weekly | Monthly | Triggered]
Compare procedure: [Script path OR manual runbook reference]
History: status/drift-history/<id>.jsonl
Re-snapshot history:
- YYYY-MM-DD: [summary of last compare result]
- YYYY-MM-DD: ...
Linked IDs: MON-XXX (data source), KPI-YYY or CFD-ZZZ or GTM-AEO-AAA (subject)
Anti-Patterns
| Pattern | Signal | Fix |
|---|---|---|
| Stale baseline | Comparing today's value against a 12-month-old baseline that no longer reflects "normal" | Add recalibrate-trigger conditions; refresh after intentional changes |
| Drift alerting on noise | Critical alert fires every week for 2% movement | Threshold is too tight; widen warn, tighten critical |
| No history log | Snapshot, compare, alert, forget | Append-only history is the whole point; without it, you can't see trends |
| Monitoring everything | 50 MON-DRIFT-* entries, nobody reads any | Pick the 3–5 things that matter; archive the rest |
| No causal interpretation | "Drift detected" with no investigation path | Each drift signal should map to candidate causes (recent release, competitor move, AI surface change) |
| Drift instead of alerting | Using drift for things that need absolute thresholds | Drift = trends; absolutes (latency > 500ms) = alerting |
Quality Gates
Before activating drift monitoring:
- Subject is quantifiable and snapshotable
- Baseline captured with explicit date
- Warn and critical thresholds defined
- Compare cadence committed
- Compare procedure (script or runbook) exists
- History log path defined and writable
- Recalibrate-trigger documented (when does baseline get refreshed?)
- Owner assigned
Downstream Connections
| Consumer | What it uses | Example |
|---|---|---|
| Runbook Creation (RUN-) | Drift criticals trigger investigation runbooks | MON-DRIFT-AEO critical → RUN-investigate-aeo-loss |
| Launch Metrics | Drift on launch KPIs informs go/no-go decisions | KPI-activation drift ≥10% → re-look at GTM |
| Feedback Loop Setup | Drift in customer-facing metrics drives CFD- investigation | Activation drift down → CFD- interview cohort |
| v0.9 AEO Audit (re-run) | Drift in AI search positions triggers re-audit | AEO drift → re-run aeo-audit |
| v1.0 Continuous Discovery | Drift patterns inform discovery questions | "Why is X drifting?" → Teresa Torres discovery |
Detailed References
- AgriciDaniel's
seo-driftskill — original baseline/compare/history pattern for SEO - Site Reliability Engineering (Google) — SLO-burn-rate alerting (related but different)
- (No bundled
references/— pattern is reusable; the content lives in each MON-DRIFT- entry)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 321k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |