The scenario
The store sells wireless earbuds, smartwatches and phone accessories in Kuwait, direct to consumer, on a Shopify site with KNET, Tabby and cash on delivery at checkout. Four years in, two staff running ads and fulfilment together, and a catalogue where a large share of traffic arrives already knowing the exact model name it wants, because that is how people shop for electronics — they search the model, not the category.
Demand was never the issue. What the ads never did was answer the question that search traffic actually carries once it lands: whether this particular seller, and not the big electronics retailer one tab over, is the safe place to hand over a card number for a product that could just as easily be a grey import with no local service behind it.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Own stock, in-house courier, same-day across Kuwait City for orders before mid-afternoon
- Team
- Two founders, one running ads and one answering WhatsApp and fulfilment
Ads that answered the spec sheet, not the doubt keeping people from paying
When we tagged the previous quarter of spend, every ad turned out to be a variation on the same format: a white-background product shot, a bullet list of specs and a price. It is the ad any electronics account runs, and it is the ad every visitor has already seen ten times on ten other listings before landing on this one.
The specs themselves were not wrong. The problem was that nobody clicking through from a Google search for a model name needed to be told the battery life again — they already knew it, because that is what they had just searched. What they had not decided was whether to trust this particular seller with a card number, especially with the big electronics retailers one tab away and a story about a fake charger or a dead grey-import earbud in every Kuwaiti group chat.
The WhatsApp inbox made the real objection obvious: two years of the same three questions, asked before almost every sale, and none of them ever answered in an ad. The account did not have a demand problem or a targeting problem. It had a trust problem wearing a specs problem's clothes.
What we did — the creative angles
The losers are here on purpose. A test with only winners was never a test.
Check the serial number before you pay
Winner"Before you pay — check this number matches the box."
- Format:
- Vertical, 15 seconds, phone camera on the serial sticker and the box seal, no music
This was the single most frequent WhatsApp question, and no ad had ever addressed it directly. Leading with the exact fear — receiving a swapped or fake unit — instead of a feature reframes the ad as a buyer's checklist rather than a sales pitch, and it disqualifies the shopper who was never going to trust an online seller anyway. It was the clear winner on hook rate and it held its cost per purchase as spend increased, which none of the spec-led creatives managed.
The same warranty you'd get at the mall
Winner"Yes — one-year warranty, same as the mall. Here's the card."
- Format:
- Vertical, 12 seconds, close-up on the warranty card and the local service-centre sticker
The second most common message, and a pure trust objection from shoppers who had been burned before by an imported unit with no local service. A twelve-second ad that shows the card and the service-centre sticker did more for the retargeting audience's cost per purchase than any discount the store had run, because it removes the exact reason a saved item never becomes an order.
Ordered before 3, on your desk tonight
Winner"Order before 3pm — it's on your desk tonight, not next week."
- Format:
- Vertical, 10 seconds, courier handoff then the same box being opened at a desk that evening
Same-day delivery is the one thing a Kuwait-based store can promise that a cross-border listing on a marketplace cannot, and this ad said so directly instead of burying it in a shipping-policy link. It converted especially well on retargeting for shoppers who had already compared prices elsewhere and come back, because speed was the tie-breaker once price had stopped being the question.
Yes, it works with the new console
Winner"Yes, it works with the new controller. We tested it ourselves."
- Format:
- Vertical, 16 seconds, side-by-side plug-in test with the newly launched console
Timed to a gaming console launch, when search volume for accessory compatibility spikes before most sellers have even confirmed it works. Answering the compatibility question first, on video, beat every ad that simply showed the accessory looking good, because the audience arriving in that window had one specific question and nothing else mattered until it was answered.
The live price-comparison screen recording
Lost"Watch us compare five stores live."
- Format:
- Vertical, 20 seconds, screen recording flipping between browser tabs and prices
A reasonable idea that lost for an instructive reason: it handed the exact habit this category is fighting — the ten open tabs — back to the shopper as an ad, and reminded her to go do it herself. It pulled cheap clicks and almost no orders, and every click it did generate went straight into an active comparison session it had just modelled, rather than a checkout.
The polished influencer unboxing
Lost"Unboxing the one everyone's talking about."
- Format:
- Vertical, 25 seconds, studio lighting, professional voiceover, no phone-camera texture
It underperformed on both hook and hold. The polish that was meant to signal quality instead read as a paid promotion, and the big electronics retailers already run a bigger-budget version of exactly this format, so the ad invited a comparison it could not win. A rougher, phone-shot version of the same unboxing tested weeks later outperformed it on every metric.
The instalment maths, said out loud
Neutral"That's about 12 a month with Tabby, not 150 at once."
- Format:
- Vertical, 14 seconds, text-on-screen with the instalment split over the product
Neutral in cold prospecting and genuinely useful further down. Framing the Tabby instalment as a monthly figure did not stop the scroll on a first view, but it earned its place as a retargeting creative for anyone who had reached a gaming laptop or high-end product page and left, which is exactly where a price objection on a big-ticket item actually lives.
What we did — the optimizations, in order
Tag the back catalogue before making anything new
We exported the previous quarter of ads, tagged every one by angle, hook type, format and language, and 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 been tried on a catalogue where half the ads look identical with different photos. The share-of-spend figure also becomes the one number the engagement is judged on, which is more honest than an account-level return that moves for a dozen unrelated reasons.
Build the angle sheet from the inbox and the search terms, not a brainstorm
We read two years of WhatsApp messages, product reviews and Google Shopping search terms landing on the store, and pulled out the eight questions and doubts that came up most often across the catalogue.
Why: In electronics the buyer arrives already informed about the product and unsure about the seller, which is a different sheet from most categories. Building it from real search terms and real messages, rather than a guess at what a shopper might worry about, put trust and delivery ahead of any feature on the list.
One testing campaign per platform, offer and audience frozen
One campaign per platform including Google Shopping, one ad set per angle, equal budgets, five to seven days, with the audience, the price and the landing page held identical across every variant.
Why: Google Shopping behaves differently from social — the buyer has already searched — so testing it alongside the social platforms rather than treating it as an afterthought is what showed us the authenticity hooks worked everywhere, while the console-launch hook only earned its budget on the platforms carrying discovery traffic.
Read hook, then hold, then click, then cost
We judged each creative in that order and recorded the specific step it failed at, rather than a single verdict for the ad as a whole.
Why: The influencer unboxing and the price-comparison ad both showed a poor cost per purchase, but one failed at the hook and the other failed at the click, and the fix for each is opposite. Diagnosing the step is what made the next round of production cheaper instead of just different.
Promote the winners into a separate scaling campaign
Winners moved out of the test and into a scaling campaign that stays untouched structurally, while the testing campaign kept its own ring-fenced share of the budget for the next round.
Why: Testing inside the campaign carrying the store's revenue resets its learning every time a new ad is added, which is exactly what had been happening for four years. Two campaigns with two separate jobs read slightly harder but scale far more predictably.
Write the hook library down, with the season already planned
Every hook tested, its result and the reason went into one shared document, along with the White Friday, back-to-school and next console-launch variants to shoot four weeks ahead of each date.
Why: This category has three predictable spikes a year, and the accounts that win them are the ones with creative ready before the spike, not shooting it during the spike. The written library is what stops the store paying to relearn the same lesson at every White Friday.
What changed
The headline number is in the table above, and the shape of it matters more than the size: cost per purchase came down while the account kept testing on Google Shopping alongside the social platforms, which is the only version of that result worth trusting. It moved because the share of spend sitting on creatives under the target cost went from a minority to a majority, not because a single video went viral.
The more useful finding was about the kind of ad that works in this category. Three of the four winners led with a trust objection — authenticity, warranty, delivery speed — rather than a feature, and the two clearest losers were the two that leaned hardest on either a comparison the store could not win or a polish it did not need. The instinct going in had been to make the product look more impressive; what actually sold was answering the question the shopper already had open in another tab.
There was a second-order effect the founders noticed before we pointed it out. Because the winning ads pre-answered the authenticity and warranty questions, the volume of pre-purchase WhatsApp messages dropped, and the ones that still arrived were closer to a decision to buy rather than a decision to trust.
What we would do next
Take the authenticity and warranty pattern onto the product pages themselves, because the same two worries are still answered by an ad and then left unaddressed on the page the ad sends people to — that is a conversion job, not a creative one, and it is where the next block of margin sits.
Second, build the White Friday and back-to-school hook variants a full month early rather than the week before, and pre-load the testing budget so the winning trust angles are already proven before the spend increases for the season.