The scenario
The brand has run roughly a dozen drops over three years, on a Shopify storefront with KNET, Tabby and cash on delivery, and every one of those orders left a phone number behind. By the time we started, the founder's phone held a WhatsApp broadcast list of several hundred past buyers — a genuinely valuable asset most stores this size never build at all, because it came free with every fulfilment message rather than a paid opt-in campaign.
The list had never been treated as more than a megaphone. Every new drop, the founder wrote one announcement and sent it to the entire list at once, whether someone had bought a wedding-season abaya once two years ago or a Ramadan jalabiya every single year since launch. The list was an asset. The way it was used was closer to a habit.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Made-to-order runs, in-house alterations, courier across Kuwait
- Team
- Founder plus one part-time stylist managing WhatsApp
Hundreds of past buyers, and one message sent to all of them every time
When we split the list by what people had actually bought before, a gateway pattern showed up immediately. Customers whose first purchase was a Ramadan jalabiya came back at a noticeably higher rate than customers whose first purchase was a one-off wedding-season abaya, which makes sense once you say it out loud: a jalabiya gets worn every night of Ramadan and replaced yearly, while a wedding abaya is bought for an occasion that might not repeat for years. The blast treated both customers identically.
The broadcast habit was also quietly training people to leave. A discount went out after nearly every drop to try to move remaining stock, sent to the same full list regardless of who had just paid full price for the same collection days earlier. Those full-price buyers, who should have been the brand's most loyal segment, were opting out at a higher rate than anyone else on the list — they were being taught that waiting a week after any drop got them a lower price.
None of the individual messages were badly written. The founder writes warmly and in real Kuwaiti dialect, which is more than most stores this size manage. The problem was never the copy. It was sending the same message, on the same day, to a jalabiya buyer, a wedding-abaya buyer and someone who had bought once and never opened a message since.
What we did — the messages and flows
The losers are here on purpose. A test with only winners was never a test.
Early access for past buyers, forty-eight hours before the public announcement
Winner"You're seeing this before Instagram does. The new collection, two days early."
- Format:
- WhatsApp message with product previews, sent only to the past-buyer list
This gave the WhatsApp list a reason to exist beyond being told the same news as everyone else at the same time. Past buyers who had first crack at sizes before a drop's stock went public converted at a meaningfully higher rate than the cold Instagram audience seeing the same collection two days later, and the early-access group's orders also settled faster because sizing had already been discussed in a previous purchase.
The broadcast segmented by occasion instead of sent to everyone
WinnerJalabiya buyers get the Ramadan collection message first; wedding-abaya buyers get the occasion-wear message first
- Format:
- Three list segments by last-purchased occasion, each receiving a different lead message for the same drop
The same drop often includes pieces for more than one occasion, and a jalabiya buyer opening a message led with wedding-season styling had every reason to scroll past it as irrelevant. Leading with the piece closest to what she had actually bought before lifted open rates and, more importantly, meant the message that reached her felt like it was written for her rather than blasted at her.
An alteration check-in a week after delivery, not a day-of thank-you
Winner"How's the length? If it needs a hem adjustment, it's free within two weeks."
- Format:
- WhatsApp message timed one week after delivery, replacing the same-day order confirmation as the only follow-up
This did two things at once: it caught fit problems while an alteration was still easy to arrange, and it gave the brand a legitimate second touchpoint that was not a sales pitch, which past buyers responded to noticeably better than they responded to any promotional message. A customer whose length problem is fixed for free is also a much more likely repeat buyer than one who quietly decided the piece never fit.
One identical announcement to the whole list, every drop
Lost"New drop is live!" sent to every saved number at the same time, unsegmented
- Format:
- The founder's long-standing habit, kept for one more cycle as the control
Kept deliberately to confirm what a fully unsegmented broadcast was actually producing: a flat repeat rate and an opt-out rate that ticked up specifically among the brand's best customers, the ones receiving a message that had nothing to do with what they had bought before. It is the pattern every other message on this list existed to break.
A discount blast to the full list after every drop, regardless of who bought
Lost"Missed it? 15% off the remaining pieces" sent to full-price buyers days after their own order
- Format:
- A blanket discount broadcast to every saved number, days after each drop closed
This was the single largest driver of opt-outs on the entire list, and it was aimed squarely at the customers the brand could least afford to lose: people who had just paid full price. It trained the list to wait for a discount after every drop rather than to buy on the announcement, which is the opposite of what a retention message should do, and it was retired within one cycle once the pattern in the opt-out data was clear.
What we did — the optimizations, in order
Read repeat rate by what someone bought first, not as one number
We split ninety-day repeat purchase rate by the occasion of each customer's first order — jalabiya, wedding abaya, everyday abaya — rather than reading one blended rate for the whole list.
Why: A blended repeat rate was hiding a genuine gateway product — the jalabiya — inside an average pulled down by wedding-abaya buyers who were never likely to repurchase quickly by the nature of the product. Splitting the read is what let the brand see which customer to build a flow around first.
Segment the existing list before writing a single new message
The full WhatsApp list was tagged by last-purchased occasion and recency, splitting it into segments the brand could message differently rather than treating it as one audience.
Why: The channel already existed and was already free; the only thing missing was treating it as a database instead of a broadcast list. Segmenting before writing anything new is what made every message that followed feel personal rather than blasted.
Build the flows a broadcast habit had been skipping
We wrote an early-access flow for past buyers, an alteration check-in a week after delivery, and a review request timed after the occasion rather than the delivery date, all in Kuwaiti dialect.
Why: A single founder writing one announcement per drop has no time left to also write a fit check-in or a review ask, so those touchpoints simply never existed. Building them as flows meant they went out every time without depending on the founder remembering to write them under drop-week pressure.
Set a cadence instead of one blast per drop
The list now receives two to three messages a month rather than one blast per drop plus an occasional discount: the early-access message, the segmented drop announcement, and the alteration or review touch, spaced by segment.
Why: The old habit was quiet for weeks and then loud once with an unsegmented blast, which is a rhythm that trains a list to ignore the brand until a discount shows up. A steady, segmented cadence is what actually keeps a list reading rather than muting.
Match the offer to the occasion, not a blanket discount
The blanket post-drop discount was retired and replaced with early access as the reward for past buyers, and any discount that does run is aimed only at the lapsed segment rather than everyone including full-price buyers from days earlier.
Why: A discount that reaches someone who just paid full price teaches her to wait next time, which is the opposite of what this brand needs from a segment that is already its most loyal. Early access rewards the same loyalty without discounting a made-to-order product whose margin cannot easily absorb it.
Win back the sixty and ninety-day lapsed with a specific reason
Customers who had not opened a message in sixty or ninety days were pulled into their own segment and messaged with something specific to their last purchase — a new fabric weight for summer if they had bought a heavier piece, or a length range now offered if theirs had been an issue before.
Why: A blanket "we miss you" message competes with every other brand sending the same line. A specific reason tied to that customer's own last order reads as attention rather than automation, and it is the version of win-back this list had never tried.
Measure opt-outs per segment and prune fast
Opt-out rate is now tracked per message and per segment rather than as one account-wide number, and the blanket discount blast was retired the moment its opt-out rate was shown to be highest among the brand's own recent full-price buyers.
Why: An account-wide opt-out number would have kept the discount blast alive for months, because it looked acceptable averaged across the whole list. Segmenting the measurement is what caught that it was specifically alienating the customers most worth keeping.
What changed
The table above carries the numbers once the brand signs them off, and the shape worth flagging now is that the improvement came almost entirely from using an asset the brand already had rather than building a new one: the same phone numbers, segmented and messaged differently, produced a materially different result.
The clearest single change was in who was opting out. Before segmentation, the brand's most loyal customers — recent full-price buyers — were leaving the list at the highest rate of any segment, which is close to the worst possible outcome for a retention program. Once the blanket discount stopped reaching them and early access replaced it as their reward, that reversed.
The alteration check-in produced a result nobody had predicted going in: it surfaced fit problems the brand had never heard about because customers who felt a piece did not quite work simply stopped buying rather than complaining. A handful of free adjustments turned into repeat customers who, on the old system, would have quietly become churn with no visible reason attached.
What we would do next
Build a referral message for the jalabiya segment specifically, since a Ramadan piece is bought inside friend groups and gatherings far more than a one-off wedding abaya is, and that segment already has the highest repeat rate to build on.
Second, connect the alteration check-in directly to the tailor's calendar so a fit adjustment can be booked from inside the WhatsApp thread itself, rather than starting a second conversation once a customer replies that her length needs work.