# Executive creative audit

3 November 2026. Review and mockups only; no deployment, production edits, database access or commits.

## Overall judgement
The established identity works. The pricing argument, recognisable logo, strong product cut-outs and clear green actions distinguish this from a generic affiliate feed. Six KEEP, four REFINE, zero RETHINK. The strongest intervention is precise language and mobile hierarchy, not a new identity or campaign platform.

## Inventory and source of truth
The live /promo-ads/ page contains G2, G5, G6, G6b and BF1–BF6: ten concepts. Its introduction describes only eight, while the supplied brief describes five. All ten are included to avoid selecting an arbitrary five. Live full-resolution images were compared by SHA-256 with the local originals; see evidence/original-manifest.json. Original source is prototypes/proper-deals-20261030/render_promo.py, supported by render_ig.py, grades.py and data/daily-2026-11-01.json and bf-check-2026-11-01.json. The older README is context, not current vocabulary authority.

## Strengths to preserve
Keep the actual category logo “TOMMO’S PROPER LEGO DEALS”, its hierarchy and proportions. Keep product-led layouts, informal British copy, the RRP → Typical price → Best offer explanation, and Tommo’s verdict. Keep the distinction between critical educational ads and positive deal recommendations. The six-product grids are useful overview creatives; forcing all ads into a single-product template would reduce range and campaign variety. Keep BF3’s “over the odds”: it describes a quantified comparison, rather than alleging deception.

## Ranked opportunities
1. G2: make the RRP denominator explicit and replace an approximate “just typical” statement with the actual 62p difference. Accuracy supports trust without changing the concept.
2. G5: retain the comparison, expose the actual 26%/32% figures and enlarge the price evidence. The original compact side-by-side concept is stronger than a multi-slide reveal that requires another swipe to understand the point.
3. BF5: bring the £3.01 comparison forward and avoid calling a dated offer a retailer’s Black Friday price without proof.
4. BF1: retain the text-led teaser, remove the unsupported prediction of a specific 50% retailer headline.

## Brand consistency
Current public terms: RRP is LEGO’s recommended retail price; Typical price is what UK shops typically charge; Best offer is today’s lowest UK price in the stated coverage; real discount is measured against Typical price; Tommo’s verdict is the editorial grade. Proper Deal is a large real discount plus the cheapest price in four months. Internal numeric grade thresholds are not public copy. “Proper price” appears in older V4/M materials and is superseded; it is not reintroduced.

Current creative tokens: navy #14213D, cream #F6F3EE, mustard #F2B134, pale blue #E3ECF9, action green #2BB673. The logo uses its existing yellow, preserved as a supplied image. Pink/red negative verdicts are useful deliberate variation, not a reason to recolour the whole system. The website and card CSS declare Inter but depend on system fallback; mockups explicitly embed the repository’s Inter font files to make the intended typography portable. No production font change is proposed.

## Reference review
Inspected original logo-system.png; desktop/mobile website exports; chart-2026-11-01-top.png; Telegram chart.png; email/email-preview.jpg; current ad-B-adult_fans-proper.png; historical review/a-v4.png and review/card-m.png; Tommo’s original desk image and Sunday Sort poster. Current B/C layouts supersede the older vocabulary. A standalone current M social advertisement was not conclusively located: card-m.png is a historical card reference, not proof of a current M ad. Long website/email exports were available but downscaled in the inspection viewer; source HTML and visible card text supplemented them. Sunday Sort treatment was checked from its poster and portfolio source, not a full audio/video editorial review. None of these references is redesigned.

## Pricing and claim checks
Verified arithmetic using displayed portfolio prices, not new retailer requests: G2 (£69.99−£52)/£69.99 = 25.70%, rounded 26%; £52−£51.38 = £0.62 above typical (1.21%). G5 roses (£54.99−£37.50)/£54.99 = 31.81%, rounded 32%; (£50−£37.50)/£50 = 25%. BF3 £75−£64.79 = £10.21; 15.76% above typical rounds to 16%; RRP discount rounds to 17%. BF4 £47.50 against £79.99 is 40.62% off; against £72 is 34.03% off. BF5 £75−£71.99 = £3.01; 4.18% above typical; RRP discount 16.66%, rounded 17%.

These are saved 1 November examples, not live prices. Every product mockup says so. Prices, retailer attribution, stock, lowest-price history and shipping eligibility need campaign-time confirmation. The saved BF data documents an older grading rule than the current public brand definition: do not use that metadata alone to substantiate every Proper Deal claim. No ranking or grading algorithm has been changed. “Shops say” is stronger than a calculated RRP comparison unless an actual retailer claim is retained. Product imagery is reused from project assets for private review; redistribution/licensing permission was not independently established. Do not treat this pack as clearance to publish.

## Territories considered, not automatically produced
Discount Reality Check is already strong in G2/G5; refine those. A sequential reveal could lose the comparison when users skip frames, so no extra carousel or Story export is justified here. A Tommo portrait is established and recognisable but would displace product/evidence in these feed ads; retain the existing verdict identity rather than inventing an endorsement. BF3 already delivers an effective Price Check / Deal to Avoid treatment and remains unchanged.

## Validation
Run five-second mobile comprehension tests first, with fresh viewers asked which figure is RRP, which is typical and whether the offer genuinely saves money. Then compare originals and approved refinements with matched audience, spend, placement and dates. Predefine the required sample and primary conversion metric from baseline traffic; no performance evidence is available to justify claiming an uplift now. Grid fine print is supporting information rather than the scroll-stopping hook. A lower CTR can still be acceptable if comprehension and qualified subscriptions improve.

## Approval
All four alternatives are PENDING APPROVAL. Six KEEP decisions are NO CHANGE. No alternative is permission to alter existing creatives. Sources and QA are recorded separately; any remaining inspection limitations are explicit.