B2B AI DirectoryB2B AI Directory
Image GenintermediatePro

Turn paid ad winners into a reusable visual pattern library

Analyze your best-performing paid ads, extract the visual patterns behind them, and turn those patterns into reusable creative templates for future campaigns.

What you will have

A visual pattern library with winning hooks, layouts, formats, proof styles, and reusable design templates for future paid campaigns.

Setup time
3-4 hours
Time saved
6-10 hours per creative planning cycle
Estimated cost
$100 to $500 per month
Tools used
6 tools

Why this works

Most creative reviews stop at surface-level conclusions like 'this ad worked' without capturing the underlying visual pattern. This workflow turns performance winners into reusable creative logic: layout, proof type, visual hierarchy, message density, and CTA treatment. The result is not just more ad variants, but a system for producing variants based on evidence.

Step-by-step workflow

Preview the workflow

The first 2 steps are open. Pro unlocks the remaining steps, copy-paste prompts, pro tips, tool-by-tool setup guidance, and implementation details.

1

Export recent ad creative performance

30-45 min

Export your last 30-90 days of creative performance from Meta Ads or your primary paid channel. Include creative name, image or video URL, campaign, audience, spend, impressions, clicks, CTR, CPC, conversions, CPA, and any notes about the offer. Put the data into Google Sheets with one row per creative.

Output

A structured creative performance sheet with metrics and creative references.

Meta AdsGoogle Sheets
Pro tip

Do not only include winners. Include losers too. Claude needs contrast to understand which visual traits may be helping or hurting performance.

2

Tag each creative by visible traits

45-60 min

Add columns for visual format, primary image type, amount of text, proof style, headline placement, CTA treatment, color intensity, product visibility, and whether a human face appears. Fill what you can manually, then use Claude to normalize messy tags into consistent labels.

Output

Creative library tagged by visual and messaging traits.

Google SheetsClaude
Pro tip

Keep labels simple. 'Screenshot with callouts' is useful. 'Modern SaaS vibes' is too vague to brief or replicate.

Prompt template
Normalize these ad creative tags into a clean taxonomy.

Raw creative data:
{{creative_data}}

Current tags and notes:
{{raw_tags}}

Create a controlled taxonomy for:
1. Visual format
2. Main visual subject
3. Proof type
4. Message density
5. CTA treatment
6. Layout pattern
7. Audience cue
8. Product visibility

Return the cleaned tags for each creative and a short definition for every tag.

Pro workflow preview

Previewing 2 of 7 steps

Pro membership

Unlock the full workflow

Get the remaining 5 steps, copy-paste prompts, pro tips, tool-by-tool setup guidance, and weekly new workflows.

$9/month

Identify winning and losing visual patterns
Turn patterns into reusable creative cards
Create design templates for the top patterns
Generate new variants from the pattern library
Review results and update the library
See Pro plan
3Identify winning and losing visual patterns
Locked
4Turn patterns into reusable creative cards
Locked
5Create design templates for the top patterns
Locked
6Generate new variants from the pattern library
Locked
7Review results and update the library
Locked

Expected results

Patterns documented

5-12 reusable patterns

A 30-90 day creative export usually contains enough variation to identify a small number of repeatable visual patterns.

Creative briefs generated

10-25 new variants

Each approved pattern can produce multiple controlled variants without requiring a net-new creative strategy session.

Planning time saved

6-10 hours per sprint

The library reduces repeated manual review of past ad screenshots and performance spreadsheets.

Learning quality

Pattern-based testing

Creative tests become easier to interpret because each new variant is tied to a named hypothesis and visual pattern.

Related workflows

Continue with workflows that share a similar GTM motion, category, or tool stack.