gooseworks-ai/goose-skills / skills/marketing/analyze-performance
analyze-performance
VerifiedWeekly performance read across every channel we run: what worked, why, and what to make next, judged on the business's real number (store revenue, qualified pipeline, retained users), not platform claims. Use when: write my Monday report; how are my ads doing — which ad won?; are my ads actually making money (Meta says 4x ROAS but revenue is flat); why did sales drop this week; did the test win; is my tracking working; monthly numbers for the replan; anything on fire today. Not for acting on the recommendations (pausing, scaling, launching) — use run-paid-campaigns; not for replanning the budget — use plan-campaign.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 15.0 KB |
| references/creative-and-channels.md | markdown | 7.0 KB |
| references/diagnosis.md | markdown | 5.5 KB |
| references/local-mode.md | markdown | 2.1 KB |
| references/measurement-contract.md | markdown | 10.3 KB |
| references/monday-report.md | markdown | 8.9 KB |
| references/monthly-and-pulse.md | markdown | 4.4 KB |
| references/revenue-join.md | markdown | 4.6 KB |
| references/run-memory.md | markdown | 14.7 KB |
| references/tests-and-learnings.md | markdown | 5.5 KB |
| skill.meta.json | config | 5.9 KB |
| tests/test-by-hand.md | markdown | 6.9 KB |
references/monthly-and-pulse.md
View on GitHub# Monthly unit-economics read and the daily pulse
## Monthly unit-economics read (first Monday of the month — its own file, summarised in that week's report)
Feeds plan-campaign's monthly replan. Saved as `reports/performance/monthly/<YYYY-MM>.md` and registered as a brand document (the `save-to-brain` skill (`../../save-to-brain/SKILL.md`), keep-history — a period artifact; local mode: its local branch, under `gooseworks-local/<brand-slug>/`). Delivered through the `write-report` skill (`../../write-report/SKILL.md`). Needs a customer-level store export (orders by customer for the last ~120 days) for the cohort lines — ask for it in that week's one input message.
- **MER and new-customer MER trend** over the last 3 months, monthly, vs break-even.
- **nCAC trend** vs allowable; **payback**: months of contribution to recover nCAC (B2B: CAC payback in months).
- **Cohort repeat** (DTC): of last month's and the month before's new customers, the share that bought again within 30/60/90 days, from the store export — judge customers by what they do over time, not day one.
- **Channel scorecard**: each channel vs its target from the brief, over the month, on blended/store-joined numbers; channels missing target two months running are named for plan-campaign.
- **Tests ended this month** and their calls; **learnings due for review** (REVIEW BY passed) — proposed in the decisions block to re-confirm, downgrade or retire (they need the marketer's words).
- **Brand signals**, only if a sibling has written them: research-market's sentiment counts from "What people are saying" (share of voice only if the section has it), improve-search-visibility's last check and citation movement. This playbook does not measure them.
## Daily pulse (interim, exception-only)
Interim until a watch-and-alert layer exists. Runs only when the marketer asks ("anything on fire?") or on a schedule they explicitly approved. **Says nothing unless a threshold breaks** — a pulse that reports every wobble trains over-reaction.
**It opens with what it can see** ("I can see Goose ads on Meta (31% of Meta spend); nothing else"), and never gives an all-clear on spend it can't see — instead: "Paste today's numbers from <platform> if you want those checked."
**Inputs, by mode** (2 minutes, plus anything the marketer pastes):
- **GooseWorks connected:** `meta_read` performance over the last 1–3 days, and `push_status` for the push ids in run-paid-campaigns' `state.md`.
- **Local mode** (`local-mode.md`): the marketer's ads connector over the last 1–3 days (account-wide, read-only), else today's export or screenshot; the paid run's local `state.md` for what was launched. No `gooseworks:` call. With neither a connector nor numbers pasted, there is nothing to check: say so in one line.
Checks (on what the inputs show; "Goose ad" below means any ad the inputs cover):
- Pacing: month-to-date spend more than ~15% off plan.
- Spend anomaly: a Goose ad set with zero delivery, or a day's spend > ~1.5× its normal.
- Rejections / failed creatives on a recent push.
- A Goose ad's CPA past ~2× allowable on mature data with real spend.
- A known stock-out or ended offer still being advertised.
- Zero reported conversions on a day with normal spend (cheapest catch of a tracking break).
- Fatigue: a Goose ad's 3-day CTR well below its 28-day CTR at steady spend.
After a launch, the first pulse (T+2) is the 48-hour read: pacing, rejections, early CTR vs the brand's norm.
Output when something breaks: one line per issue — what, evidence, the proposed action and who takes it (almost always run-paid-campaigns, as a pause/swap proposal behind its gate). Otherwise: "Nothing needs you today in what I can see" — plus the coverage line.
Scheduling either the Monday report or the pulse is a standing configuration change: offer once, and create it only on an explicit yes, by marketing-guide §3 Schedules:
- **GooseWorks connected:** `gooseworks:automation_upsert`; say scheduled runs use credits; the method copied into the workspace and the self-contained prompt in `run-memory.md` § Setting up the schedule.
- **Local mode:** the host's own scheduler, created off; its prompt names this playbook, the brand, local mode, the pulse or weekly job, and the scheduled rules in `run-memory.md` § Scheduled runs (read and report only; ask nothing). It reads the connector, or the exports the marketer left in the folder. No host scheduler → say it isn't available here. Never `gooseworks:automation_upsert`.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |