The blueprint run end to end on a real case: Wildhorn (wild American bison organ capsules) against Primal Edge Bison, a competitor selling the same product under another brand. Every step shows the prompt sent, what came back, and the decision taken.
Skill ad-blueprint-gpt-image-2, images generated with OpenAI gpt-image-2 (quality high, 1:1). Image 1 of every generation is the real Wildhorn jar photo.
Role: you are my head of creative strategy for an e-commerce brand selling on Meta. Context - My brand: WILDHORN — Wild American Bison organ capsules — trywildhorn.com - The competitor: PRIMAL EDGE BISON — buy-primaledge.com — they sell the same product under a different brand (same 4 bison organs, same promise). - I'm giving you a folder of 35 of their ads: iCloud / ADS ECOM / Primal (static images + videos). - I'm also giving you a folder of winning templates, sorted by concept. Mission, in this order 1. Full market research on Primal Edge (statics, video transcripts, TrendTrack). 2. 1:1 adaptation of the winning templates across the winning angles. 3. Generation with GPT Image 2 + QC. 4. Comparison site + demo link. Get my sign-off on the market research before step 2.
Here is the folder of 35 ads from PRIMAL EDGE BISON. 1. Transcribe every video. 2. For EACH ad: format · exact headline · subheadline · hook · angle · mechanism · visual format · proof · offer · CTA. 3. Group the ads by angle, count the duplicates. 4. Rank the angles by weight.
| What we measured | Finding |
|---|---|
| Volume analyzed | 35 statics + 15 video transcripts + the product page |
| Static vs video | 45 of 46 active ads are images — statics carry the account |
| Most duplicated ad | "I Tested 3 Organ Supplements For My Beer Belly" — 13 copies, run through a persona page |
| The thesis behind everything | "A beer belly isn't a fat problem. It's an estrogen problem." The enzyme in belly fat converts testosterone into estrogen, which stores more fat: a loop. |
| Product story | 4 organs, 4 jobs: Kidney blocks the conversion · Liver regulates · Heart burns the store · Testicle is the source |
Using the TrendTrack MCP, pull everything on PRIMAL EDGE BISON: Meta pages, active ads and 6-month trend, scaling and duplicated ads, video transcripts, traffic and 30-day trend, offers, landing pages. Cross-check with the 35-ad analysis and cite every number.
| # | Angle | The idea |
|---|---|---|
| A1 | Beer belly = estrogen | The king angle. The belly isn't beer or discipline: an enzyme in belly fat turns testosterone into estrogen, which stores even more fat. A loop. |
| A2 | Man boobs / puffy nipples | The second door in, and the strongest scroll-stopper. "It's not fat. It's a hormone signal." |
| A3 | Wild vs farmed | Attacks the category: farmed reindeer, feedlot beef. "2,000 lb wild vs 350 lb farmed." |
| A4 | "I tested 3 organ supplements" | Listicle advertorial. Their most duplicated ad (13 copies). |
| A5 | Anti-waste | "I spent $1,750", "12 supplements every morning", "It's not a supplement problem, it's a FOOD problem." |
Day 3 / 7 / 14 / 30 timeline · ancestral and patriotic with holiday promos · scarcity and authenticity ("BEWARE OF FAKES", apology letter) · cortisol belly with a GLP-1 twist · the doctor who said "it's just age".
Pain & mechanism — MAN BOOBS GONE IN 30 DAYS (NOT SURGERY. NOT PUSH-UPS. NOT A COMPRESSION SHIRT.) · PUFFY NIPPLES? THAT'S NOT CHEST FAT. IT'S A HORMONE SIGNAL. · FLUSH THE CORTISOL — LOSE THE BELLY · A stubborn gut isn't a discipline problem.
Comparison — 3 ORGAN SUPPLEMENTS. 1 SOFT CHEST. WHICH ONE ACTUALLY WORKS? · WHICH ANIMAL'S ORGANS WOULD YOU TRUST? · SCAM? — WE TESTED THE MOST POPULAR BEEF-ORGAN SUPPLEMENTS · THIS IS WHAT'S INSIDE EACH CAPSULE.
Product — 4 WILD BISON ORGANS — One Capsule. The Full System.
Story — 26 WINTERS ON A RANCH. ONE BOTTLE PASSED. · FOR THE MEN WHO BUILT THIS COUNTRY
Urgency — ⚠ OFFICIAL APOLOGY STATEMENT · ⚠ BEWARE OF FAKES · LAST UNITS. JUNE STOCK.
Video hooks — "If your gut feels like a bag of concrete you carry around all day…" · "Your wife says snack less. Your doctor says it's just your age…" · "If you're swallowing 12 supplements every morning and your gut still hasn't moved…"
Avatar — American man, 45-65, married, beer belly or man boobs. Tried gym, keto, tongkat, shilajit, zinc; spent $1,000-2,000. Sucks his belly in at the mirror, keeps his t-shirt on at the pool.
What proves it to him — the belt moving one notch, seeing his feet, tucking his shirt in.
Winning formats — us-vs-them table, anatomical diagram, product with the 4 organs, raw UGC, cinematic ranch, text-only letter, patriotic promo.
Save the brand kit for WILDHORN in a single folder: /root/workspace/wh-remix/kit/. It must contain: - the final market research report (angles, headline bank, video hooks, avatar); - the product sheet (exact price, offer, guarantee, allowed claims, banned words); - the product reference photo; - the list of every creative already produced: concept · angle · headline · JSON · image. This kit is the only source for every future batch.
New batch for WILDHORN. 1. Reload the brand kit: /root/workspace/wh-remix/kit/. Do NOT research anything new, do NOT use TrendTrack. 2. Here are the new concept templates I picked: concepts.zip (79 templates). 3. Accept every template, sort them into concept families, and adapt each one 1:1 to WILDHORN on EVERY winning angle of the kit, with the headlines from that angle's bank (rewritten, aggressive, conversion-driven). 4. Double every creative (2 headline variations). 5. Skip any combination already in the list of creatives made. 6. One JSON per creative, generation with GPT Image 2 (quality high, 1:1), QC. 7. Add the new creatives to the comparison site and the demo link. Start with the first concept and show it to me before running the rest.
| # | Concept family | Templates | Status |
|---|---|---|---|
| 1 | US vs THEM / comparison | 10 | 10 / 10 done |
| 2 | Before-after / timeline | 8 | 8 / 8 done |
| 3 | Problem reframe (X is not the problem) | 8 | 8 / 8 done |
| 4 | Mechanism / anatomy / ingredients | 6 | 6 / 6 done |
| 5 | Origin / ranch / real food | 8 | 8 / 8 done |
| 6 | Offer / bundle / price | 11 | 11 / 11 done |
| 7 | Bold claim + hero product | 9 | 9 / 9 done |
| 8 | Meme / humor / UGC | 10 | 10 / 10 done |
| 9 | Off-avatar, re-skin last | 9 | 9 / 9 done |
One ad per template (R4): 79 templates in, 79 ads out. Inside each family, the 5 winning angles from the research are spread across the templates.
| # | Concept family | A1 Belly = estrogen | A2 Man boobs | A3 Wild vs farmed | A4 Tested 3 | A5 Anti-waste | Ads |
|---|---|---|---|---|---|---|---|
| 1 | US vs THEM / comparison | 2 | 2 | 2 | 2 | 2 | 10 |
| 2 | Before-after / timeline | 4 | 1 | 1 | 1 | 1 | 8 |
| 3 | Problem reframe (X is not the problem) | 3 | 1 | 1 | 1 | 2 | 8 |
| 4 | Mechanism / anatomy / ingredients | 1 | 1 | 1 | 1 | 2 | 6 |
| 5 | Origin / ranch / real food | 2 | 1 | 2 | 1 | 2 | 8 |
| 6 | Offer / bundle / price | 3 | 3 | 2 | 1 | 2 | 11 |
| 7 | Bold claim + hero product | 2 | 3 | 1 | 1 | 2 | 9 |
| 8 | Meme / humor / UGC | 3 | 3 | 2 | 1 | 1 | 10 |
| 9 | Off-avatar, re-skin last | 1 | 2 | 2 | 2 | 2 | 9 |
Grouped by concept family. Each pair: the template on the left, our ad on the right. Generated with gpt-image-2 (high), Image 1 = the real jar, Image 2 = the template. Every image checked at full resolution.





















































