Case studies · spec work on real brands

Five projects. One way of working.

Each one starts with what customers say and ends with a decision. Short version here. Open any "See the numbers" box for the detail.

01Customer research 02Competitor map 03Hooks 04Creator briefs 05Test report
01PetLab Co. · dog probiotic

Most owners don't know why their dog itches

QuestionWhere are itchy-dog owners in their thinking?
DidRead 10 Reddit threads, pulled 16 real quotes, checked Google Trends.
DecisionStart ads with the problem, not the product.
See the numbers

Search interest (Google Trends, US, 12-month avg): "dog yeast infection" 62 · "dog licking paws" 53 · "probiotics for dogs" 51 · PetLab brand searches near 0.

Awareness: mostly problem aware. Few owners search the brand, so brand-first ads would talk to almost nobody.

The angle: owners treat the skin, but the brand argues the itch is often linked to the gut. That's why creams and food swaps only help for a while.

Owners who described the problem

10 Reddit threads
12 of 19

couldn't name a cause for the itching.

No idea whyNamed a cause
02PetLab Co. vs PetJoy, BetterWild

Everyone sells the science. Nobody sells the calm dog.

QuestionWhich angles are crowded, and which are open?
DidMapped 11 live competitor ads, scored 10 angles with ICE.
DecisionTest the open lane first, with two safe bets beside it.
See the numbers

Proof the mechanism angle works: PetJoy's mechanism ads have run for 1,300 to 1,600 days. Long-running ads are usually profitable ones.

Top 3 angles by ICE score: "My dog won't take pills" (9.0) · owner story of the calm dog they got back (8.8) · yeast and gut link (8.7).

Test plan: 1 desire angle (open lane) + 2 angles close to what already works. If the open lane wins, the brand owns it.

What 11 competitor ads are selling

Meta Ad Library

Striped rows are open lanes. No competitor is running them.

03Lanera · foundation for mature skin

Hooks for women 50+ that pass Meta's rules

QuestionHow do you say "this is for you" without naming her age?
DidBuilt hooks from real 4.8★ reviews. The creator says her own age.
DecisionEvery claim traces to the product page or a real review.
See the rules I used
  • Avoid: "Are you over 50?" Calling out the viewer's age breaks Meta's personal attributes policy.
  • Avoid: "Look 10 years younger." It can't be proven.
  • Use: "I'm 61, and…" The creator talks about herself.
  • Use: claims from the product page, like "doesn't settle into wrinkles."

First 3 seconds, three ways

Same body, 3 hooks
Reels0:03Main"Someone asked if this settles into lines. I'm 61. Watch it at hour ten."Comment reply
Reels0:03Test"I stopped wearing foundation. Every one cracked by lunch."Failed solution
Reels0:03Test"My husband asked what I changed. It was one bottle."Social proof
04Lanera · UGC production

Briefs a stranger can film on the first try

QuestionCan a creator who's never heard of the brand get it right first time?
DidWrote two full briefs: script, shot list, rules, specs, and fees.
DecisionPlan the proof shot, so the claim is shown, not said.
Open brief: Comment reply Open brief: Before and after

Inside a brief

Shot list excerpt
LAN_S2 · Shot list8 shots · 1 creator · 1 day
01Reads the comment, eyebrow raisedCar · Must
03Smile lines at 8am, just appliedVanity · Must
07Same smile lines at 4pm, same lightVanity · Must
08Scrolls reviews, then shade quizCar · Must
9 sectionsReels safe zonesDo's and don'tsUsage rightsFile naming
05Lanera · creative test report

The worst ad wasn't the problem. The checkout was.

QuestionAfter $100 per ad, what do we kill, fix, or scale?
DidSet break-even first, then followed each ad down the funnel.
DecisionFix checkout first. It changes the result of every ad.
See the numbers

Break-even: $29.97 price − $12 costs = $17.97 CPA, or 1.67 ROAS.

AdHookHoldCTRCPAROAS
Social proof19%9%1.5%$16.671.80
Comment reply36%3%0.7%$50.000.60
Failed solution31%10%2.1%$33.330.90

Practice dataset built for this case study, not the brand's real results.

One call per ad

Break-even CPA $17.97

Social proof

Scale carefullyCPA $16.67

Only ad under break-even. Weak hook, so test 3 new ones.

Comment reply

Rewrite bodyHook 36% → Hold 3%

Best hook, but people leave fast. The body breaks the promise.

Failed solution

Fix checkout40 carts → 3 buyers

Best hold and CTR. The ad works. The store loses them.

Next step

Want this for your brand?

Send me one ad you're running. I'll reply within 24 hours with one thing I'd change.

Get started Back to home

All five are spec projects built from public data: Meta Ad Library, Google Trends, Reddit, and brand product pages. None of these brands are clients.