Reading a creative system before the data arrives: a Jope case study
A day-one creative diagnostic for a vet-founded dog supplement brand, built from the public Meta ad library with no performance data. The point is the method, because a strategist who can find the leak in a system they have never touched will find it faster once the numbers are in front of them.
The short version
- What it was: I tagged 100 of Jope’s Meta creatives from the trailing 90 days across eight dimensions and read the system that produced them, with no spend or conversion data.
- The finding that mattered: roughly 80 of 100 creatives run the exact same headline, which is the fingerprint of a brand with no concept layer above its executions, so production keeps defaulting to the one proven block.
- What I built from it: five structured, testable hypotheses and seven net-new concepts across four untouched messaging territories, sequenced into a first test round designed to learn rather than just perform.
The setup
Jope sells a UC-II collagen joint supplement for dogs, positioned against the glucosamine category, founded by two veterinarians who built it after failing to help their own dog. The wedge is unusually strong: UC-II is clinically shown to be 59% more effective than glucosamine and chondroitin, and the product carries a 4.9 rating across more than 10,000 customers. The proof exists. My question was whether the creative was using any of it.
How I read it with no data
An ad library will not tell you what converts, but it tells you how a team thinks, because the choices repeat. So the read is structural rather than performance-based, built on three questions: where is this system likely to leak, what single point of failure is it carrying, and what has the brand never put into market at all.
What the library showed
- Volume and cadence. One hundred creatives in ninety days with a forty-plus burst across three days, but heavy churn, with many creatives dying in a day or two, which signals either premature kills before the algorithm exits learning or no clear winners scaling.
- Format mix. Roughly sixty-five percent video to thirty-five percent static, and the statics are repeated product shots, so the cheapest single-frame way to deliver the information half of the sell (why UC-II beats glucosamine) is going unused.
- Messaging concentration. About eighty of one hundred creatives share one headline (“I only wish I had tried sooner”) and one core copy block, so the account is running one ad many times against overlapping audiences rather than many ads.
- Landing pages. Seven or eight well-built pages exist, more sophisticated than most brands at this stage, but one copy block routes traffic to all of them, so earned clicks frequently land on a different promise than the one that pulled them in.
- Hook dependency. The two strongest video formats both require the founder on camera, so output volume is capped by one person’s calendar.
The finding that mattered most
Multiple creators do not independently converge on one headline by accident. That convergence happens when nothing in the system is asking them to build a second message, so everyone defaults to the proven one. The headline is not broken. The absence of a system that generates alternatives is the actual issue, and that is the thing a Creative Strategist is hired to fix.
From findings to hypotheses
A finding is an observation. A hypothesis is a structured bet with a reason and a checkable result, and it is the unit of work a strategist is accountable for.
| # | Observation | Diagnosis | Root cause hypothesis | Test | Expected result |
|---|---|---|---|---|---|
| H1 | ~80 of 100 creatives share one headline and copy block | Top-funnel fatigue risk; can’t isolate concept performance | No concept layer above the executions | Build three new territories (emotional transformation, prevention, vet authority) against the incumbent angle | One new territory holds at or above the incumbent on a non-overlapping audience, creating a second pillar |
| H2 | 65/35 video to static; statics are repeated product shots | Information half of the sell rides only in long video | The “why UC-II wins” message competes with emotion for runtime | A static system carrying the 59% comparison, the 90-day guarantee, the vet-made credential | Comparison and stat statics lift CTR on problem and solution-aware audiences at a fraction of video cost |
| H3 | Both strongest formats need the founder on camera | Velocity capped by founder availability | No validated non-founder hook format exists | Three founder-free formats: UGC compilation, screen-record social proof, before-and-after reel | One non-founder format hits viable hook and hold scores, decoupling volume from the founder’s calendar |
| H4 | 7-8 landing pages, one copy block routing to all | Message-to-destination mismatch suppressing conversion | Creative and pages managed as separate systems | Match each territory to its intent-aligned page | Conversion rises on matched pairs versus mismatched, isolating page intent as a lever |
| H5 | Many creatives died in 1-2 days; 40+ in a three-day burst | Possible premature kills before learning ends | Either judging on insufficient data or a completed promo misread as failure | Re-run a sample with adequate flight and budget | A share of “failed” concepts were never given enough data to judge, recovering usable concepts |
Every hypothesis is built from public creative, not spend, so the structural reads hold but the prioritization can shift the moment the account connects. The headline everyone runs might be the real winner rather than the lazy default, and knowing which is exactly why you connect the data before scaling.
The seven concepts
Seven concepts cover the four territories Jope is not touching, each chosen to pull a specific psychological lever. Hooks are verbatim and ready to test.
| Concept | The idea | Verbatim hook | Lever |
|---|---|---|---|
| The Comeback Reel | UGC compilation of dogs going from struggling to thriving, no founder, no lecture | “She couldn’t get up the stairs anymore. Watch what happened in 90 days.” | Transformation tension, founder-free |
| Don’t Wait for the Limp | Static aimed at younger owners, framed as proactive care before a problem exists | “80% of senior dogs get arthritis. Start protecting their joints now.” | Stakes tension, founder-free, new segment |
| The Pepsi Diaries | Mini-documentary from existing b-roll of the founders’ own dog, product in the final frame only | “Two vets couldn’t help their own dog. So they built what didn’t exist.” | Vulnerability tension, low dependency |
| The 10,000 Dog Wall | Screen-record scroll through real reviews with a climbing counter, all proof, no claims | “10,000 dogs can’t be wrong. Here’s what their owners are saying.” | Social proof tension, founder-free |
| The Shelf Test | A vet tears down a big-box label beside Jope’s, two cuts to test credibility vs. authenticity | “Your vet would never recommend this. Here’s what’s actually in your dog’s joint supplement.” | Conflict tension, founder-led |
| Built for the Trail | UGC from agility, hunting, and ranch communities showing dogs back in action | “She couldn’t keep up on the trail three months ago. Watch her now.” | Transformation tension, founder-free, niche segment |
| What Vets Give Their Own Dogs | Founders reframed from product authorities into insider sources, no pitch until the end | “I’m a veterinarian. This is what I actually give my own dog for her joints.” | Vulnerability and identity tension, founder-led |
Three of the seven need zero founder shoot time, which is the direct answer to H3 and the reason the pipeline keeps moving when the founders are unavailable.
The first test round
A slate of seven is a roadmap, not a first move. I would split the opening round three ways so it produces a winner and an insight at the same time:
- Exploit, two concepts that double down on what works: keep iterating the founder-responds-to-comment format, and add The Shelf Test, which finally shows the glucosamine argument Jope already owns in copy.
- Test, one concept that changes a single variable: What Vets Give Their Own Dogs uses the same founder asset but swaps the frame from pitch to insider confession, so anything that moves is attributable to the framing.
- Explore, one concept in untouched territory: Don’t Wait for the Limp or The Comeback Reel puts a genuinely new messaging pillar into market for the first time.
The remaining concepts sequence in behind based on what the first four teach.
The organizing idea
Every concept lives in one of three places: the science (The Shelf Test, Don’t Wait for the Limp), the joy (The Comeback Reel, Built for the Trail, The 10,000 Dog Wall), or the bridge between them (The Pepsi Diaries, What Vets Give Their Own Dogs). It works as an internal filter for the creative team rather than a tagline, and it gives every future concept one test before it gets briefed: does this show the science, the joy, or the connection. That is the difference between a pipeline and a pile of one-off ideas.
So here is the question I would put to the team in the first meeting: if the angle on eighty percent of your ads stopped working tomorrow, what would you run on Monday?
