The scenario
The brand sells home blood-pressure monitors and digital thermometers in Kuwait, direct to consumer, on a Shopify store with KNET, Tabby and cash on delivery at checkout. One founder, no in-house media buyer, and a customer base that skews heavily toward adult children ordering a device for a parent who has never owned one.
Demand was never the issue. Every winter flu season and every stretch after Ramadan, when households book their check-ups, orders climbed on their own. What the brand had instead was an ad account leaning on language that sounded borrowed from a product insert — accuracy claims, precision claims, the kind of phrasing a platform's health policy exists to catch.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Own stock in a Kuwait warehouse, next-day delivery across all six governorates
- Team
- One founder, two part-time WhatsApp support agents
Ads that read like a warning label, not an answer
When we pulled the previous quarter of ads, the pattern was hard to miss: every creative opened with a line about accuracy and readings, the kind of phrasing that sounds like it belongs on a product insert rather than an Instagram feed. Two of the four ads used the word "accurate" in the first three seconds, and one implied the device could catch a problem before it became serious — a line that turns into a health claim the moment you say it out loud.
None of this was working. Cost per purchase had been climbing for two straight months, and the brand's answer had been to write one more line about precision, which is the one thing nobody scrolling Instagram at night is actually thinking about. The real conversation was happening somewhere else entirely: in the WhatsApp inbox, where almost every message asked some version of the same two questions before anyone asked about price.
Those two questions were about the box, not the reading. Is this the real device, registered and sold through a proper channel, and is it simple enough to hand to a parent who has never used one in his life. The account did not have a product problem or a targeting problem — it had spent months answering a question nobody was asking, in language a platform could flag.
What we did — the creative angles
The losers are here on purpose. A test with only winners was never a test.
The registration question, answered before it's asked
Winner"Before you buy any monitor — ask for the registration number."
- Format:
- Vertical, 15 seconds, phone camera, the box and the sticker shown close up
This was the exact phrase sitting in the WhatsApp history, word for word, and no ad had ever used it. Leading with the customer's own worry instead of a benefit turns the ad into advice rather than a pitch, and it does the disqualifying work for you: a shopper who does not care about authenticity is not really the audience anyway. It was the clear winner on hook rate and it held its cost per purchase as spend increased.
Buying it for your father, not for yourself
Winner"My dad checks his numbers every morning. This is what I got him."
- Format:
- Vertical, 20 seconds, a daughter setting up the monitor at a kitchen table
Almost every order in this category is bought by an adult child for a parent, and the brand's ads had never once acknowledged that. Shifting the ad from "you" to "your father" matched who was actually scrolling and paying at checkout, and it outperformed every version written in the second person on both hook rate and completed purchases.
Two minutes at home instead of a clinic queue
Winner"No appointment. No queue. Two minutes at your own table."
- Format:
- Vertical, 18 seconds, unedited, real kitchen, morning light
Convenience turned out to matter more than precision language, especially for a device meant to be used daily rather than once. This creative performed best on hold rate specifically, and it worked well as a second ad shown after the registration hook had already earned the click, because it answers the next objection: that this replaces a habit, not a hospital visit.
Same box, compare the code yourself
Neutral"Check the code on the box against the one online."
- Format:
- Vertical, 12 seconds, close-up on a serial code and a phone checking it
Flat cold, useful warm. It did not win attention on its own, but it earned a place as a retargeting creative for people who had reached the product page and left, because a lingering authenticity doubt is exactly what sits between a saved item and an order in this category.
Know your numbers before it's too late
Lost"Do you actually know your numbers? Find out before it's too late."
- Format:
- Vertical, 14 seconds, stock-style graphic with a pulse animation
This was the closest the brand came to a real problem: the line implies a health outcome, reads as a warning rather than a product ad, and one platform flagged the account for review over it. It also underperformed on every metric before the flag arrived — fear did not convert here, it just made people scroll faster past a claim they did not trust.
The white-coat demonstration
Lost"Doctors recommend checking your blood pressure daily."
- Format:
- Vertical, 16 seconds, a model in a white coat demonstrating the cuff in a clinical set
Strong on paper, cold in the auction. A studio-clinical setting made the product look like hospital equipment rather than something a household keeps in a kitchen drawer, and citing what doctors recommend edges toward a claim the brand cannot make on its own behalf. The same message worked once it was rewritten as a customer's own morning routine instead.
Before the season, not during it
Winner"Stock up before the season starts. Delivered tomorrow."
- Format:
- Vertical, 13 seconds, text-on-screen with a delivery-day promise
A straightforward seasonal push tied to the point every household actually thinks about a thermometer, timed a few weeks ahead of the winter flu season rather than during the rush when couriers slow down. It read as practical rather than promotional, and it converted well as a top-of-funnel ad once the trust-led hooks had already done the harder work.
What we did — the optimizations, in order
Tag the back catalogue before writing a single new line
We exported the previous quarter of ads and tagged every one by angle, hook type, format and claim risk, then worked out what share of spend had sat on creatives below the target cost per purchase.
Why: Without this you are guessing at what has already cost you money, and the claim-risk flag matters as much as the cost read: a cheap ad that borders on a health claim is not a win, it is a liability waiting to surface on the account.
Build the angle sheet from the inbox, with a compliance pass
We mined two years of WhatsApp messages, reviews and return reasons for real questions, wrote each one first in Kuwaiti dialect, then checked every line against platform health-ad rules before it went into production.
Why: The brand's actual worries were about authenticity and ease of use, never a health outcome, so every real angle already sat inside what is allowed to say. Doing the compliance pass at the angle stage is far cheaper than reshooting after a disapproval.
Production sprint in real homes, not a studio
We shot ten pieces over two weeks in a real kitchen and living room, with three hook variants cut on top of the two strongest bodies rather than one hook per shoot day.
Why: This category's buyer trusts a phone camera in a real home more than a lit set, and shooting light meant we could cut a fresh variant the same week a new WhatsApp question started repeating, instead of waiting for the next booked shoot day.
One test campaign per platform, everything else frozen
One campaign on Meta and one on Instagram placements, one ad set per angle, equal budgets, five to seven days, with audience, offer and landing page held identical across all of them.
Why: With a small, need-based audience like this one, a moved audience or a changed offer makes the read meaningless within days, and this account could not afford to burn a week on a test that told it nothing it could act on.
Read hook, then hold, then click, then cost — and flag claims separately
We judged each creative in that order and separately logged which lines were borderline on health-claim language, regardless of how well the ad was performing on cost.
Why: A creative can win on cost and still be a liability if a line reads as medical advice. Treating compliance as its own column rather than an afterthought is what let the brand keep its winners live without a policy strike hanging over the account.
Promote the winners, retire the claim-heavy set entirely
Winning creatives moved into a scaling campaign, and we pulled every ad using clinical or fear-based language at once, even the one still holding an acceptable cost, rather than letting it fatigue on its own.
Why: The risk of a platform-level flag on the whole ad account is worse than one ad's mediocre cost, and removing the pattern rather than the single ad stopped it from reappearing in the next production round without anyone noticing.
Write the angle library, seasonal variants included
Every hook tested, its result and its compliance read went into one shared document, along with three seasonal variants scheduled ahead of the winter flu season and the post-Ramadan check-up period.
Why: The asset that lasts past this engagement is the written record of which real objections sell in this category without a claim, which is what stops the brand relearning the same lesson — or risking the same warning — every winter.
What changed
The headline result is in the table above, and the shape of it matters more than the size: the cost per purchase came down while every clinical-sounding line left the account entirely, not because one ad went viral. It moved because the share of spend sitting on creatives below the target cost went from a minority to a clear majority.
The more useful finding was about the kind of ad that works in this category. Three of the four winning creatives led with a trust worry or a caregiving moment rather than a claim about the product, and the two clearest losers were the two that leaned closest to a health outcome. The brand's instinct had been to sound more precise; what actually sold was answering the authenticity question the customer was already asking, in her own words.
There was a second effect the founder noticed before we pointed it out. Once the winning ads answered the registration and authenticity questions up front, the volume of pre-purchase WhatsApp messages dropped, and the account's one platform review closed cleanly, because nothing left running still read as a claim.
What we would do next
First, take the authenticity and registration answer onto the product page itself, because the same question the ad now answers in three seconds is still unanswered on the page it sends people to, with no batch photo or registration proof visible without scrolling.
Second, build the winter and post-Ramadan variants a full month early rather than the week before, and pre-load the delivery messaging for the week couriers are busiest, since the peak in this category is predictable to the fortnight.