The scenario
The brand's bestseller by a wide margin is a couple's fragrance set, two bottles boxed together, sold almost entirely as an anniversary or wedding gift on a Shopify store with KNET, Tabby and cash on delivery at checkout. A smaller discovery range of single-note samples sits alongside it as the low-commitment first order for anyone unsure which of the brand's six core scents suits them. Four years in, the founder still runs every WhatsApp conversation from her own personal number, which had already been temporarily restricted once for sending too many broadcasts too quickly.
Every message the brand had ever sent went to the same person: whoever's card was charged at checkout. Nobody had asked the obvious question that the product itself raised every day — a couple's set is bought by one partner and worn by two people, and the wearer who never appears in the order data might be the easier sale on her own account entirely.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Own stock, gift-wrapped in house, courier delivery across Kuwait
- Team
- Founder plus one part-time assistant handling WhatsApp and packing
The brand kept messaging the buyer and ignoring the person actually wearing the scent
The cohort read was blunt: split by first product, repeat purchase within ninety days on the couple's set was close to nothing, while the small discovery range that should have been the lower-value product was quietly outperforming it. That is backwards for a business trying to build repeat revenue, and the reason became obvious once we mapped who actually receives each order.
A discovery-set buyer is shopping for herself, so every message sent to her afterward reaches the person who will actually decide whether to buy again. A couple's-set buyer is usually shopping for someone else's anniversary or wedding, so every message sent to her afterward reaches someone who already has the gift and has no obvious reason to buy the same set twice — while the partner who wore it, formed an opinion about it, and might want her own bottle of one of the two scents had never received a single message from the brand in four years.
The broadcast cadence made this worse rather than better. With no segmentation between the two very different buyers, the whole list received the same monthly discount blast, which trained the anniversary gift-buyer to expect a lower price on a purchase she was only ever going to make once a year anyway. The brand had built its entire retention effort around the wrong half of its own customer base.
What we did — the messages and flows
The losers are here on purpose. A test with only winners was never a test.
Capturing the recipient's own WhatsApp number at delivery, with consent
Winner"We're delivering your gift today — should we send the care instructions to her number too?"
- Format:
- Delivery-day WhatsApp message to the buyer, with an opt-in link for the recipient's number
This was the single missing piece in four years of messaging. Framing it as a practical delivery detail rather than a marketing ask made buyers comfortable sharing it, and it gave the brand its first legitimate, opt-in way to reach the person who actually wears the fragrance rather than the person who only ever bought it once.
A check-in with the recipient a week after delivery, not the buyer
Winner"How is it wearing on you? If you want your own bottle, here's a discovery set to start with."
- Format:
- WhatsApp Business API template message, sent only to opted-in recipients
This message reached a person the brand had never spoken to before, at the exact moment she had formed an opinion about the scent she was wearing. Offering a discovery set rather than a full bottle matched the actual size of the decision she was being asked to make, and it converted at a noticeably higher rate than any offer previously sent to the original buyer.
A replenishment nudge for discovery-set buyers who never upgraded
Winner"You tried three scents a while back. Ready for the full bottle of your favourite?"
- Format:
- WhatsApp message timed to roughly ten weeks after the discovery-set order
Discovery-set buyers were the brand's best-converting cohort and the one it had never followed up with at all. Timing the nudge to when a sample typically runs low, rather than on a generic monthly schedule, caught the customer at the moment of an actual need instead of an arbitrary calendar date.
An anniversary win-back sent to the original couple's-set buyer
Winner"Your anniversary is coming up. Same set, or something new this year?"
- Format:
- WhatsApp message sent three weeks before the customer's own purchase-anniversary date
A couple's set is not a repeat purchase on a random cycle; it is tied to a specific yearly date the brand already had in its own order history and had never used. Reaching the buyer three weeks ahead of her own anniversary, rather than waiting for her to think of the brand unprompted, turned a one-time purchaser into the account's first meaningful year-over-year repeat cohort.
A blanket monthly discount sent to the entire list regardless of segment
Lost"15% off everything this month" sent to every contact on the list
- Format:
- Monthly WhatsApp broadcast, no segmentation
This was the brand's only retention activity before the engagement, and it was quietly training its most valuable segment — anniversary gift-buyers — to expect a discount on a purchase that only happens once a year regardless, while contributing nothing to the recipients who had never been messaged at all. It also carried the highest opt-out rate of anything the brand sent, which is what finally justified retiring it.
What we did — the optimizations, in order
Read repeat rate by first product, not as one blended number
We split ninety-day repeat purchase by whether the customer's first order was the couple's gift set or the discovery range, rather than reading one repeat rate for the whole store.
Why: A blended repeat rate was hiding the fact that the bestselling product was structurally unlikely to repeat, while the smaller discovery range — the one everyone treated as an afterthought — was the brand's real gateway product.
Move off the founder's personal number and capture the recipient too
The brand moved to the official WhatsApp Business API with template messages, and the delivery-confirmation flow now asks the buyer, with clear consent, whether the recipient's number can also receive care instructions and future messages.
Why: A personal number that has already been restricted once cannot be the channel a growing retention programme depends on, and without a legitimate, consented way to reach the recipient, half of every couple's-set order was permanently invisible to the brand.
Build separate flows for the buyer and the recipient
Order confirmation and delivery update go to the buyer as before; a scent check-in and a discovery-set offer go to the recipient a week later; a replenishment nudge goes to discovery-set buyers at roughly ten weeks; an anniversary reminder goes to couple's-set buyers three weeks ahead of their own purchase date.
Why: Two people with two entirely different relationships to the product need two different messages, and merging them into one flow is what had made the brand's retention effort invisible to the person most likely to actually buy again.
Retire the blanket discount and set a segmented cadence
The monthly blast to the whole list was retired, replaced with two to four messages a month, each segmented by whether the contact is a buyer or a recipient and by what they last bought.
Why: A generic discount to everyone had been training the wrong behaviour in the brand's most valuable segment while doing nothing for the segment it had never spoken to. Segmenting the cadence let each message actually fit the person receiving it.
Match the offer to the repurchase cycle each product actually has
A discovery-set-to-full-bottle offer timed to when a sample runs low, and a couple's-set reminder timed to the customer's own anniversary date rather than a store-wide sale calendar.
Why: A gift set bought once a year and a sample used up in ten weeks do not belong on the same promotional calendar, and forcing them onto one is why the brand's only offer before this project fit neither product well.
Track opt-out by message type and keep a weekly log
Every flow and broadcast is tracked separately for opt-out rate and revenue per recipient, with the two-person team reviewing the log weekly and killing anything whose opt-out rate rises, however well it appears to sell in the moment.
Why: The blanket discount had been carrying the account's highest opt-out rate for months before anyone noticed, because nobody was tracking opt-out by message. A weekly log is what stops a small team repeating the same mistake with the next new flow.
What changed
The table above carries the numbers once the founder signs them off, and the shape worth noting is that the repeat-rate gain came almost entirely from a customer the brand had never messaged before: the gift recipient. Reaching her a week after delivery, at the moment she had an opinion about the scent, produced a conversion path that simply had not existed.
The anniversary reminder mattered in a quieter way. Because it is tied to a fixed yearly date already sitting in the order history, it is the first piece of retention infrastructure the brand has that keeps working without anyone remembering to build a new campaign, which was never true of the old monthly discount.
Retiring the blanket discount also mattered on its own. Opt-out fell once messages stopped training the brand's most loyal segment to wait for a lower price, and revenue per recipient rose because every message now had a specific reason to exist rather than a calendar slot to fill.
What we would do next
Extend the recipient-capture idea to corporate and bulk gifting orders around National Day, where a single buyer's purchase can introduce the brand to a dozen new recipients at once if each one is invited to opt in.
Second, build a light referral step into the recipient's first check-in message, since she has just been introduced to the brand by someone she trusts, which is a warmer opening than any cold audience the paid side of the account will ever reach.