spillover-estimator
# Spillover Estimator
Estimate cross-channel spillover without pretending to prove perfect attribution.
## Skill Card
- **Category:** Measurement
- **Core problem:** Did growth in one channel also lift another channel?
- **Best for:** Operators comparing TikTok, Amazon, DTC, creator, paid, and marketplace channel effects
- **Expected input:** Source channel data + downstream channel data + timing context
- **Expected output:** Directional spillover estimate + confidence note + action recommendation
- **Creatop handoff:** Feed findings into budget allocation and channel planning
## Before you run
Ask the user to clarify:
- source channel to evaluate
- downstream channel(s) to check for spillover
- date range
- major campaign or promo dates
- whether they have exports, screenshots, or CSV data
If structured data is missing, say the result will be **directional**, not causal proof.
## Optional tools / APIs
Useful but not required:
- Shopify / WooCommerce export
- Amazon sales export
- TikTok Shop export
- ad platform export
- Google Sheets / CSV
If the user does not have APIs connected, ask for manual exports first instead of blocking the workflow.
## Workflow
1. Confirm channel scope and time window.
2. Collect source-channel change signals.
3. Collect downstream-channel change signals.
4. Align timing around campaigns, creator drops, content bursts, or promo windows.
5. Judge whether the downstream lift looks:
- likely related
- weak / mixed
- insufficient evidence
6. Explain the estimate with honest caveats.
## Output format
Return in this order:
1. Executive summary
2. Spillover estimate
3. Evidence blocks
4. Confidence and caveats
5. Recommended next step
## Fallback mode
If the user only has weekly snapshots, rough screenshots, or partial exports:
- use simple directional comparison
- do not claim causal attribution
- clearly label missing data and confidence limits
## Quality rules
- Never overclaim causality from timing alone.
- Prefer directional clarity over fake precision.
- Separate channel correlation from verified lift.
- Make the user’s next measurement step obvious.
## License
Copyright (c) 2026 **Razestar**.
This skill is provided under **CC BY-NC-SA 4.0** for non-commercial use.
You may reuse and adapt it with attribution to Razestar, and share derivatives
under the same license.
Commercial use requires a separate paid commercial license from **Razestar**.
No trademark rights are granted.
标签
skill
ai