A Claude toolkit for e-commerce teams testing short form UGC and product ads: forty resources (ten prompts, ten templates, ten workflows, five audits and five trackers), three complete worked cases and a plan for when there is too little data to choose a winner.
<aside> 🚀
🚀 FREE STRATEGY CALL Want to grow your product through viral reach? We help teams across iGaming, mobile apps, AI products, and SaaS build creator-led organic distribution generating tens of millions of views at CPMs starting from $0.03.
<aside> 🚀
</aside>
On the call you'll receive:
| In the post | Count | Where it lives on this page |
|---|---|---|
| Review what the current ads are telling you | 4 | The H2 section with that name: P01, T01, W01, TR01 |
| Check the promise in the ad against the product page | 4 | The H2 section with that name: P02, T02, W02, A01 |
| Choose the question for the next test | 6 | The H2 section with that name: P03, P04, T03, W03, A02, TR02 |
| Decide what changes between versions | 6 | The H2 section with that name: P05, P06, T04, T05, W04, W05 |
| Write a brief the creator and editor can use | 7 | The H2 section with that name: P09, P10, T09, T10, W09, W10, A05 |
| Track what ran and what changed | 5 | The H2 section with that name: T06, W06, A03, TR03, TR05 |
| Read the results and plan the next batch | 8 | The H2 section with that name: P07, P08, T07, T08, W07, W08, A04, TR04 |
| 40 resources: 10 prompts, 10 templates, 10 workflows, 5 audits and 5 trackers | 40 | The seven sections above. The manifest lists them all, and the zip holds every file |
| Complete example: testing two openings | 1 | The three worked cases, case 1 |
| Complete example: investigating clicks without purchases | 1 | The three worked cases, case 2 |
| Complete example: refreshing an existing creative | 1 | The three worked cases, case 3 |
| What to do when there is too little data to choose a winner | 1 | Case 1, plus P08 Thin Data Check and W08 Too Little Data Workflow |
| Instructions and a completed example for every resource | 40 | Inside each toggle, and in the matching file in the zip |
| Free Access | 1 zip and 28 files | Downloads, in Start here |
Jump to: Start here. The shared files. The seven sections of the library. The three worked cases. How to use it every week. How this was made.
README.md, then docs/FIRST-TEST-WALKTHROUGH.md.context/campaign-product-brief.md and add your approved claims to registers/claims-ledger.csv.trackers/TR01-evidence-log-tracker.csv. That is your first real result. Plan for about 20 minutes. The timings are estimates, not measured.<aside>
How to run it. Copy a prompt from a toggle on this page or from its file. Paste it into a Claude chat, then paste your inputs below the last line. Templates, audits and trackers also work by hand with no model. Claude prepares analysis, briefs and recommendations from what you paste in. It has no access to your ad account and it does not make videos. Manual exports and file handoffs are the default. A human media owner decides budgets and publication.
Before you paste anything. Before using these prompts you should omit any PII (personal identifiable information) and check with your compliance team and/or your AI policy.
</aside>
<aside>
Downloads. Start with the zip. It holds all 60 files: the 40 assets, the five tracker registers, the manifest, the shared files, the source dataset, the three cases, the walkthrough and a reconcile script.
carrywell-claude-ad-testing-toolkit.zip
Files one by one: 28 files you can open on their own </aside>
What is inside the zip
<aside>
30 minutes on Zoom. Your product, your niche and the creator formats already working in it.
</aside>
Every prompt, template and workflow reads from the same five shared files. They are for the fictional Carrywell Modular Desk Organiser. Replace the content with your own and keep the headings.