scar niche: $1,700 spent, stuck on creative

bigcoffee

Been running ads for a scar treatment product and I'm bleeding roughly $700 after $1,700 in spend. I came in knowing creative would be a grind, but I've hit a wall and need to hear how others actually think through this.

What I've done so far:

Pulled creatives from a spy tool, rewrote hooks, added captions, trimmed them, then ran on Meta. Not best practice but I had to start somewhere.
One video actually worked: ten sales on roughly $270 spent. Then Meta flagged it on a weird policy violation that's likely tied to using an implied before/after.
Posted a cleaned-up version and it dropped to three sales on the same budget. The risk-free version lost whatever juice the original had.
Also tested an advertorial angle. Straight-up article ad linking to a problem-led page with a CTA to my product. Click-through sat around 7%, but zero conversion. Both the ad and the page sold the same problem, but somewhere a link in the chain broke.

Numbers so far: $1,700 in, net down about $700, one promising ad dead, and the replacement is weaker. I have a gut feeling this audience is skeptical enough that static images won't do it, they need video trust. But that's an assumption I haven't tested, and right now my money's gone.

Two questions I keep circling:

How do you actually learn to read these signals so you're not just guessing and burning cash? I've been asking AI for every decision and every explanation. It sounds smart but I'm not sure it's teaching me anything.
If you were in my seat, what's your next move to get this from a costly experiment to a repeatable funnel?

Appreciate any brutal honesty on what I'm missing.

dailygrind

Burning roughly two thousand dollars on recycled creatives is a fast track to broken unit economics. Slapping new labels on assets built for a different audience will never yield a sustainable CAC to LTV ratio.

To fix this, implement a structured testing process:

Analyze search intent. Run a quick gap analysis in Ahrefs or audit Google Search Console to see the exact queries users type when seeking scar solutions. Build angles around those real pain points.
Check landing page compliance. Run a Screaming Frog crawl to ensure site health, then ensure the copy aligns with the ad. Ad networks flag before-and-after imagery aggressively, so compliant hooks are essential.
Own the creative pipeline. Stop letting AI handle media buying. Study structured resources, like Dara Denney's frameworks, to design original assets tailored to the target market.

metricsmuse

That 7% CTR ad that didn't convert is your most useful data point. It's not a creative problem - people got interested, then bailed on the page or checkout. Before you touch another creative, figure out where they dropped.

On the winning ad, 10 sales on $273 is solid. The v2 dropping to 3 usually means you changed more than just the flagged element. Worth going back and comparing them frame by frame - hook, first three seconds, claims. Whatever you softened to pass review was probably the thing making it work.

Scar treatment is a health claim category, so this'll keep happening. Before-and-afters, cure language, anything implying results gets flagged. Testimonial framing where the person describes their own experience survives better than you making the claim.

Learning this stuff mostly comes from running things and reading the breakdowns - placement, age, country, funnel steps. Most diagnosis is just finding the step where the number falls off a cliff.

I'd stop testing new creative until you know if the problem's the page.

wordweaver

$1,700 is not a dataset, honey. That's a single test flight with maybe one real signal-the $273 ad that did 10 sales. Everything else is just expensive noise you're trying to read like tea leaves.

On the ban: been there. Meta's bot doesn't care about your careful copy. It's reading the visual promise. Before/after, close-up skin, anything screaming "this removes your defect" gets flagged regardless of wording. The ads that survive move the transformation off the body-show the routine, the ingredient, the "two weeks later" feeling. Let the viewer mentally supply the before. We had a near-identical ad die the same way. Rebuilt it with the same hook and script, just dropped the proof shot, and it landed within spitting distance of the original CPA.

That 7% CTR on your advertorial is the loudest number in the post, but it's not telling you about creative. It's telling you about the hook-people got interested, read the thing, then didn't buy. Trace that path: LPV → ATC → purchase. If ATC is fine but checkout leaks, it's price, shipping, or trust. If nobody adds to cart, your advertorial is selling a different product than your PDP is.

Stop guessing if an ad is "good." Start asking "which step lost the money." Write down what you expect a number to be before you look at it. Change one layer at a time. AI is decent for hypotheses, useless for judgment-it'll confidently explain any number you feed it, including the wrong ones.

Static vs video? Gut's probably right for scar treatment-video does the trust work. Keep video on cold, try statics on retargeting where that trust is already earned.

chloeest

Testing methodology over creative volume - that's the pivot point you're at. Your v1 already proved demand before the ban hammer hit, so the signal is there. A 7% CTR on the advertorial says the hook works, zero conversions says the landing page isn't delivering on the promise. Classic message-match gap.

Pause the spend. Spin up one campaign: the winning v1 angle, re-edited to clear policy, $50/day cap. Run it for a week and track only three things - CTR, time on page, add-to-cart. The creative that holds attention longest is the one building trust. Everything else is noise.

Happy to eyeball the ad and landing page if you want a second set of eyes - no strings, just a quick audit. DM me the links.

ranktracker

Two classic bumps in the road. For the banned creative, platforms get super twitchy with scar products. Your safest bet is to steer the copy and visuals toward general confidence and wellness, avoiding any direct before/after promises. That's how you stay compliant.

That second issue is the real money pit. A solid CTR on your advertorial with zero sales almost always means a disconnect in your funnel. The advertorial has to pre-sell the exact product you're sending them to, and the landing page needs to pick up that thread seamlessly. What kind of scars is this product targeting?

pixelperfect

Okay, let's slow down for a second, because I can feel the frustration in this post. That banned ad? That's the best lead you've got right now, and you're moving past it too fast.

You had 10 sales for about $273 spend. That's a $27 cost per sale. Whatever that creative was saying, it was working. Then version 2 gets 3 sales on a similar spend, and your brain is screaming "version 2 is worse!" Maybe. But it could also be random noise, the account taking a hit after the ban, or you accidentally cutting the exact line that was doing the heavy lifting when you cleaned it up. Three sales versus ten at that spend level... that's not enough data to know for sure what happened.

On the ban itself: scar treatment lives in the health and personal attributes category. The flag is almost certainly before-and-after imagery or copy that implies a physical change in the person reading it. this isn't a vague policy, it's a very specific one. It's worth reading those two sections properly instead of guessing which piece to cut. Most people in this niche get banned over and over for the exact same core reason while thinking each strike is a new problem.

that advertorial with a 7% click-through rate that didn't convert? that's not a bad page, that's a mismatch. a 7% CTR is high. people clicked because the ad promised something the page didn't deliver. Read the ad and the first paragraph of that advertorial back to back, out loud. Usually, it's two completely different conversations.

Now, on your actual question. you can't diagnose anything at $1700 spread across that many changes. You've switched creative source, format, angle, and funnel type. There's no control variable anywhere, so none of your learning so far is reliable. That's the real bottleneck, not the specific creative.

here's what I'd do first. Take the angle from that banned winner, build one compliant version, and run it against one variable at a time. Same audience, same page, same daily budget. Change only the hook. Four ads that differ in just one way will teach you more than forty ads that differ in ten.

Static versus video is a fair question, but it's not your bottleneck. You'll be able to answer it in a week once you have a control to test against.

bigcoffee

Good catch on the pixel gap. that 15 sessions vs 50 outbound clicks is a red flag - either the pixel wasn't firing or the click tracking is inflated. Worth checking your UTM parameters and whether the outbound click event is set to "link click" vs "landing page view."

On the ad ban: "recommended by my surgeon" is a classic trigger for medical claims. Glad you removed it. Running v2 in the same ad set as v1 means it likely inherited the learning phase from v1, but if v1 was already flagged, that could carry negative signals. i'd spin v2 into a fresh ad set with no history.

What's your current cost per session on the advertorial? and are you seeing any add-to-cart or checkout data from those 15 sessions?

metricsmuse

50 outbound clicks turning into just 15 sessions? that's where your money's bleeding out. you're losing roughly 70% of paid traffic before the page even loads. no amount of creative tweaking will fix that.

Usually it's page speed - run the landing page through PageSpeed Insights on mobile. Anything over three seconds and you're hemorrhaging visitors. Redirects can add to it too, if your ad URL bounces through a redirect before landing, that's another chunk gone. also check how it loads inside Instagram's in-app browser specifically - plenty of pages break there but work fine everywhere else.

Fix that before touching anything else. same ads, same spend, you'd get three times the traffic actually arriving.

on the surgeon line - yeah, that's a third-party endorsement claim and it's a hard no in health. that was probably the thing doing the heavy lifting in your creative. The closest legal version is a customer describing their own experience in their own words, not citing a professional.

The v1 vs v2 in the same ad set thing isn't your main issue. Ads don't inherit performance data from each other, they compete for delivery within the ad set. if v1 had history, the algorithm favors it, so v2 gets the worse slice of the audience. Different problem, but way smaller than the traffic loss.