What the founders believed walking in
Two clinicians sold their practice and bet everything on a nerve-support supplement. Paid ads, a marketplace listing, a storefront. Over a year in, no profit, savings draining, and a fear they said out loud: that they would die on the way up the mountain before the business turned.
They had done the responsible thing. The category is heavily policed, so they sanded the copy down to stay compliant. Then sales collapsed, and the belief that formed was reasonable: the message had gone soft, so the fix was to be bolder again. Let us say more in the ads. Turn the claim back up.
It was a fair read. The old copy used to convert, the new copy did not, and the only visible difference was how much it dared to say. When your numbers are falling, that is the obvious lever to pull.
What we saw
We did not touch the copy. We laid their own funnel out in front of them, one stage at a time.
The click was already solved. Cost per click was under a dollar and the ads were pulling roughly seventeen hundred clicks a week. For a supplement, that is not a weak ad. That is a strong one. Nobody argued once the number was on the screen.
So the leak was not the ad. It was everything after the click, and it was the compliance posture itself. The landing page was ambiguous about what the product was for, so people clicked and left. And the platform reads more than the ad. It crawls the landing page, and it watches a tracking tag that is shared across accounts. That is why, after the ban, even their unrelated accounts got shut off. The problem was never one weak ad. It was a poisoned setup that new creative alone could not clean.
Then we reconciled the dashboard against the source. It showed zero purchases over thirty days. The tracking tag, over twenty-eight days, showed 56 purchases. The scoreboard was lying to them, and their fear was reading the lie as failure.
What we decided, and what we chose not to do
The tempting path in this category is to game it. Rent cheap accounts, burn three or four a day at twenty dollars each, keep relaunching under new company names. We named that path plainly, then refused it. A real business with real customers treats account-burning as a liability, not a trick, because you cannot sell a company built on disposable accounts.
We also did not go bolder on the claim. When every competitor is making the same forbidden promise, the promise stops being leverage anyway. The concession we made to the founders was honest: we could not give them assurances, the odds were roughly even, and pretending otherwise would have been disrespectful. Then we showed them the move that did not need the risk to get the reward.
What we built
We rebuilt the paid surface around claims that could stand on their own, and gave the fuller story a channel built for it: a newsletter the customer asks to receive.
A compliant funnel bridge. The ad points to an educational page led by a clinician video, not a buy page. The page’s job is to teach and to earn an email, not to close, and education honestly done needs room a sales page never gives it. Buy-now lives only in the sticky navigation, in a bright color, never as the page’s main button. The moment it becomes the main button, the page becomes a sales page and gets read as one. Some of what the brand needed to communicate to its customers, the fuller education, the ongoing guidance, the stories customers told in their own words, needed a different mechanism than an ad ever allows, and the newsletter became that mechanism: a channel the customer chooses, with the space to say things properly and the same standard of honesty everywhere. Customer stories stayed on the customers’ own accounts and in their own words, because a customer can say what they want about their experience, while the brand republishing it as its own promise is the violation.
A claim swap, one rung down. We downgraded every line we could not defend to the rung beneath it. A pain signal is a bodily function, so “supports nerve communication” holds where “relieves pain” does not. A count of satisfied users holds where a success percentage does not. And we sold the life past the problem, so we rarely had to make the forbidden claim at all.
A clean rebuild, in one week. Three locked steps. Rebuild the landing page to full compliance, with the legal footer, privacy policy, and disclaimers missing the first time. Build the new compliant ads in a shared sheet for group review before anything went live. Then launch on a fresh ad account with fresh tracking tags, because a contaminated setup cannot be cleaned with new creative. The relaunch was aimed at the first week of November.
The failure became a checklist. The page that got them banned should never have gone live without its legal footer and terms. So that check is now a required step in a pre-launch checklist. The incident is not a memory anymore. It is a gate.
We also kept two versions of the message. The strict platform gets the most careful variant. The looser channels, like the marketplace, keep the bolder one. We did not let the strictest platform tax every other channel too.
What changed
The rebuild held, and the numbers moved in the direction that matters.
The unit economics flipped. Orders went from losing money on every ~$30–40 sale to a profitable ~$82 average order. The compliant ad that came back held about a 1.36 return on what went into it. The marketplace held roughly 2.57 to 3 times. These are one brand’s figures in one engagement, not a rule. Spend stayed capped at the founders’ own comfort number, around twenty to twenty-five thousand a month, judged month to month.
The thing they thought they had wasted turned out to be an asset. Around 500 customers added in the rebuild, roughly 1,800 across the brand’s life, a high reorder rate, a product people came back for. That is real equity. It just does not show up in a bank balance, which is why the fear felt so total when the dashboard read zero.
The loop now runs on numbers instead of feelings. Find the pocket where the ads return two times, then scale slowly, no more than about twenty percent a day, so the tracking keeps learning. Hold the return at or above 1.5. Grow the owned email list as a hedge. Steer or stop if the numbers stop moving.
The lesson
When a regulator takes your claim away, the answer is almost never to whisper the same claim more quietly. It is to rebuild. Swap every claim you cannot defend for the one you can, on every surface. Give the education and the ongoing conversation a mechanism built for them, a newsletter the customer opts into. Sell the life past the problem, so the forbidden claim stops being the thing the business depends on.
The founders arrived sure they had a copy problem. They had a placement problem, and placement was cheaper to fix and safer to keep. The higher bar in a locked-down category is not only a cost. It is a wall competitors have to climb too. Learn to advertise cleanly inside it, and the thing keeping you out starts keeping them out.