Looking for our creative agent?See it
All playbooks

Get more customers · Play

Map which ad angles you've never tried

Most brands run the same three ad ideas in different colours.

Many stores run the same few ad ideas in different colors. Connect Meta Ads, open the Meta Ad Library, and ask Claude to tag your ads by angle, audience and format, show the grid with empty cells, and do the same for your competitors. Fill the most promising empty cells with new ads and compare on cost per sale.

  1. 1
  2. 2

    Ask

    Tag every active and recent ad I've run by angle (problem, benefit, social proof, comparison, offer, founder story), by audience it speaks to, and by format (static, video, UGC, carousel). Show it as a grid and highlight the empty cells. Do the same for my 3 closest competitors.

    Paste this into Claude, ChatGPT, Codex or Cursor.

  3. 3

    Act

    Fill the most promising empty cells with new ads. Let Meta sort them.

What it tends to findExample

5 ads in 6 months, all the same angle. Competitors had launched 25 that month.

Watch out

Empty doesn't mean good. Start with cells where competitors have long-running ads.

Step by step

  1. 01Connect Meta Ads read-only through a Claude connector or MCP server, or export your ad list, and give Claude the Meta Ad Library through browsing.
  2. 02Ask the playbook prompt to tag ads by angle, audience and format and show empty cells for you and your competitors.
  3. 03Check which empty cells competitors fill with long-running ads. Empty does not mean good.
  4. 04Launch new ads in the most promising empty cells on a capped test budget.
  5. 05Compare new ads on cost per sale against current ones, and re-run the grid monthly.

Questions

Run these next

Every number on these pages is an illustrative example, not a result from a named store. Read-only access is enough for every check; make any change yourself.

See where your funnel leaks

Book a call. We'll show you where you're losing sales and what to fix first.

Takes a minute. Then you pick a time.