The scenario
The brand designs and sells sport hijabs, full-coverage leggings and long-sleeve training sets in Kuwait, direct to consumer, on a Shopify store with KNET, Tabby and cash on delivery at checkout. One founder, a small in-house sewing and quality team, and a product range built specifically for women who train, run or attend a women-only gym class without changing what they wear to get there.
The following was real. Women-only gym culture in Kuwait had grown steadily, and the founder had built a loyal Instagram audience from consistent, well-lit product photography: flat lays, mannequin shots and studio portraits with the model standing still. What the catalogue never had was a single second of the product actually moving on a real body.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Own stock, local courier, next-day across Kuwait
- Team
- One founder, a small sewing and quality team, one part-time content creator
A catalogue that looked perfect standing still and said nothing about moving
When we tagged the previous quarter of spend, every creative in the account was a variation on the same shot: the product on a mannequin, or a model standing square to the camera in good light, holding still. The photography was genuinely good. It just answered a question nobody in the audience was asking, which was whether the set looked neat in a photo. Nobody buys modest activewear to look neat in a photo.
The real question sat in two years of WhatsApp messages and Instagram comments, asked a dozen different ways that all meant the same thing: does it stay covered when I bend, when I squat, when I stretch to shelve a weight. That is the actual purchase decision, and a still photograph cannot answer it, however well it is lit.
The founder's instinct, when performance softened, had been to book another studio day and refresh the same kind of shot. That refreshed the wardrobe on screen without refreshing the argument the ad was making. The account did not have a photography problem. It had a movement problem, and no amount of better lighting on a still pose was ever going to fix it.
What we did — the creative angles
The losers are here on purpose. A test with only winners was never a test.
The squat test, filmed once, unedited
Winner"Squatting in it right now — watch the hem."
- Format:
- Vertical, 15 seconds, one continuous take, gym floor, phone camera
This directly answers the single most common question in the brand's inbox, and it does the only thing that can actually answer it: shows the fabric under load, in real time, with no cut. It disqualifies nothing and reassures everything, and it was the clear winner on hold rate because women were watching the whole fifteen seconds waiting to see the hem move.
A prayer-motion coverage check
Winner"Rukoo, sujood, standing back up — full coverage the whole time."
- Format:
- Vertical, 18 seconds, real prayer motion in the sports set, filmed at home
A second objection nobody had ever put in an ad: the set also has to survive prayer without a wardrobe change. It is a claim no generic imported activewear brand can credibly make, and it became the best-performing creative in retargeting because it settled a doubt no other ad acknowledged.
A real class, not a shoot
Winner"This is my actual Thursday class, not a photoshoot."
- Format:
- Vertical, 20 seconds, filmed inside a women-only studio with the instructor's permission
Puts four real customers in one frame, moving through a normal circuit rather than posing, and it consistently outperformed anything shot with a hired model, because the brand's own customer recognised herself in it.
The dance-studio backdrop
Lost"Move with confidence, in coverage that keeps up."
- Format:
- Vertical, 15 seconds, professional model against a polished dance-studio backdrop
A reasonable hypothesis that lost for a specific, repeatable reason. The polish read as a showroom rather than real life, and viewers could not tell from a choreographed sequence whether the fabric held under an ordinary squat, which is exactly the doubt this audience arrives with.
The price comparison against a global sportswear brand
Lost"Same coverage. Look at the difference in price."
- Format:
- Vertical, 14 seconds, split screen with the price on each side
Lost for the same reason price comparisons usually lose here: it pulled in cheap clicks from discount hunters, not buyers with a coverage objection, and it invited comparison with a brand whose fit reputation this store had not yet earned.
Fabric stretch, close up
Neutral"Watch this fabric stretch and bounce back."
- Format:
- Vertical, 10 seconds, macro shot of the fabric being pulled by hand over the knee
Neutral cold, but useful further down. A close-up on fabric recovery did not stop the scroll alone, but it earned a place in retargeting for shoppers who had already watched a movement video and wanted proof the stretch held.
Sizing without trying it on
Winner"I'm a medium everywhere else. Here is what I ordered."
- Format:
- Vertical, 16 seconds, a customer comparing her usual size against what fit in this brand
Built to test whether the pattern worked on the second-biggest objection: sizing without a fitting room, since most orders are never tried on before delivery. It performed almost as well as the squat test, confirming the pattern was about answering a real doubt, not the squat itself being magic.
What we did — the optimizations, in order
Tag the back catalogue before making anything new
We exported the previous quarter of ads and tagged every one by angle, hook type, format and language, then worked out what share of spend had sat on creatives below the target cost per purchase.
Why: Every creative in the catalogue was a photograph, so the audit doubled as a diagnosis: the account had never once tested a video showing the product in motion. The gap was not a weak angle among many; it was an entire format missing.
Build the angle sheet from the inbox, not a brainstorm
We read two years of WhatsApp messages, Instagram comments and return reasons, and pulled out the questions that came up most. Each became an angle, written first as the customer's own sentence in Kuwaiti dialect.
Why: The founder already knew the coverage question by heart because she answered it by hand every day. What she had never done was turn it into advertising material instead of customer service, and an angle sheet built from real messages made that translation obvious rather than debatable.
Shoot movement, not poses, and make it look unproduced
The production sprint replaced the studio day with unedited, single-take clips filmed on a phone in a real gym, a bedroom and a women-only studio, each built around a specific movement rather than an outfit.
Why: A polished shoot signals effort, but this audience reads polish as distance from real life, which is exactly what they doubt. Rough footage of a real squat carries more evidence than a beautifully lit photograph, because it shows the moment the product holds or does not.
One testing campaign per platform, frozen everywhere else
One campaign per platform, one ad set per angle, equal budgets, run for a fixed window, with the audience, the offer and the landing page held identical across every one of them.
Why: If the offer or the audience moves at the same time as the creative, a strong result cannot be traced back to the movement footage specifically, and the founder would have been left guessing whether it was the squat or a coincidence.
Read hook, then hold, then click, then cost
We judged each creative in that order and wrote the reason it failed at whichever step it failed, rather than recording a single verdict on the ad as a whole.
Why: The dance-studio ad and the price-comparison ad looked similar in a cost-per-purchase column but failed for opposite reasons — one never earned trust, the other earned the wrong audience — and only reading the funnel in order surfaces which fix each one actually needs.
Promote the winners into a separate scaling campaign
Winners moved out of the test and into a scaling campaign that never gets restructured, while the testing campaign kept its own ring-fenced share of the budget for the next round of angles.
Why: Testing inside the campaign carrying the store's revenue resets its learning every time a new ad enters it. Two campaigns with two jobs made the winning movement footage stable enough to carry a real budget increase into the peak season.
Write the hook library down, seasoned four weeks ahead
Every hook tested, its result and the reason went into one shared document, alongside a January, post-Ramadan and National Day variant of the winning pattern, shot ahead of each season.
Why: January is the loudest peak here but not the most profitable one, and the weeks after Ramadan bring a buyer with less competition for her attention. A written library means those variants exist before the demand does, not scrambled together the week traffic arrives.
What changed
The headline number is in the table above, and its shape matters more than its size: cost per purchase moved down while the account shipped a genuinely new format for the first time, which is the harder version of that result to earn. It moved because the footage answered an objection still photography could never touch, not because one clip went viral.
The more durable finding was which movement mattered. The squat and the prayer-motion test both outperformed every polished shoot, and the price comparison lost for the same reason the dance-studio ad lost: neither gave the viewer evidence, only a claim. The ad has to show the fabric surviving a real movement, and nothing substitutes for that.
The founder also noticed something we did not have to point out: the questions arriving before purchase shifted from coverage toward sizing, because the footage had already answered what used to take her three messages to talk a customer through.
What we would do next
Take the movement pattern onto the product page itself, because the same coverage question is answered in the ad and then left to a still photograph the moment a shopper clicks through. That is a conversion job rather than a creative one, and it is margin sitting unclaimed.
Build the indoor-summer variant early, showing the same sets worn for a home workout, since that is where this category's demand shifts once outdoor heat sets in and women-only studios get busier.