The scenario
The brand sells premium straighteners, dryers and stylers direct to consumer, built a genuinely loyal Instagram following on demo content, and had never sent a single WhatsApp message to a customer despite selling a product with an obvious replenishment story: brush heads wear out, cords fray, and a serious user of a hot tool eventually wants a second one for a bag or a second bathroom.
The founder had built the business almost entirely on new-customer acquisition through Snapchat and Instagram ads, which worked, but meant every dinar of revenue was being earned fresh each month rather than compounding through anyone who had already bought. A Shopify order-confirmation email was the only message a customer ever received after checkout, and open rates on it were close to nothing.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Own stock, local courier, next-day across Kuwait
- Team
- Founder plus one part-time customer-service hire
Happy customers, no channel back to them, and a list nobody had built
The reviews were genuinely good and the Instagram comments read like a brand with real loyalty, which made the repeat-purchase number more surprising than it should have been. When we pulled the cohort data by first product bought, a clear pattern showed up: buyers of the mid-range straightener came back for a second tool or an accessory at a noticeably higher rate than buyers of the flagship dryer, who tended to approach the purchase as a one-time, considered decision and never hear from the brand again.
Nobody had ever tried to change that, because nobody had a channel to try it in. Email existed on paper but was read by almost no one, and the brand had no WhatsApp presence beyond the founder's personal number, which she was not going to hand out to every customer. Every marketing dinar went to finding a new buyer, which is the most expensive way to grow a brand that already has thousands of past customers sitting quietly, half of whom asked a pre-purchase question on WhatsApp and never heard from the brand again once the order shipped.
The founder's assumption was that a hair tool is bought once and there is nothing more to sell. The problem was never the product's repurchase cycle. It was that the brand had never opened a channel its own customers were already waiting on.
What we did — the messages and flows
The losers are here on purpose. A test with only winners was never a test.
WhatsApp opt-in captured at checkout and again on delivery
Winner"Get your warranty confirmation and care tips on WhatsApp" as the opt-in line
- Format:
- Checkout checkbox plus a second opt-in prompt on the delivery-confirmation message
Framing the opt-in around the warranty and care tips rather than "join our list" gave the customer an immediate, concrete reason to say yes, on a product where the warranty question is one of the most common pre-purchase worries. Asking a second time on delivery caught the customers who skimmed past the checkout checkbox, on the official WhatsApp Business API rather than a personal number.
A replenishment check-in timed to when brush heads and cords actually wear
Winner"How's the straightener holding up? Brush heads usually need a refresh around now."
- Format:
- One-to-one WhatsApp message, sent at the product's own replenishment interval
This was written to arrive at the point of genuine need rather than on a generic monthly schedule, which is why it read as a check-in rather than a sales pitch. It produced the highest reply rate of any flow, most of it in plain Kuwaiti dialect, and a meaningful share of those replies turned into an accessory order without a discount attached.
Sixty-day lapsed segment given one specific reason to return
Winner"We restocked the shade everyone asked about" sent only to people who had asked about it
- Format:
- Segmented one-to-many broadcast, tied to a specific past enquiry
A specific reason tied to something the customer had actually asked about outperformed a blanket discount to the same segment sent two weeks earlier, and it did so without training the list to wait for the next markdown. The lapsed customers who came back from this message bought at full price.
A generic monthly broadcast to the entire list
Lost"Check out what's new this month" sent to every opted-in number regardless of what they had bought
- Format:
- Unsegmented one-to-many broadcast, sent on a fixed monthly date
It was the fastest way to see an opt-out spike in the whole engagement. Sending the same message to someone who had bought a dryer last week and someone who bought a straightener eight months ago reads as noise to both of them, and the opt-out rate on this message was several times higher than on any segmented send, which is exactly the signal that kills a young WhatsApp channel before it earns trust.
A formal Arabic thank-you template after purchase
LostA standard-Arabic formal thank-you message, translated word for word from the English template
- Format:
- Automated post-purchase WhatsApp template, formal register
It read like a bank notification rather than a message from a brand the customer had just spent real money with, and reply rates on it were near zero compared with the dialect version tested against it. Rewriting the same message in Kuwaiti dialect, in the customer's own register, is what turned a dead template into a channel people actually replied to.
What we did — the optimizations, in order
Read the cohorts by first product bought
We pulled repeat purchase rate at thirty, sixty and ninety days, split by the customer's first product rather than as one blended figure for the store.
Why: The blended repeat rate was hiding a real gateway product — the mid-range straightener — whose buyers behaved completely differently from the flagship dryer's buyers. That single split is what told us which cohort to build the flows around first.
Capture WhatsApp opt-in at checkout and again on delivery
We set up opt-in on the official WhatsApp Business API, added at checkout and repeated on the delivery-confirmation message, framed around the warranty and care tips rather than a generic newsletter ask.
Why: A channel a customer actually reads is worth more than a bigger email list nobody opens, and building it on the official API from day one avoids the number ban that hits brands sending from a personal WhatsApp or an unofficial tool.
Write the flows in Kuwaiti dialect, not translated formal Arabic
Order confirmation, delivery update, the replenishment check-in, the review ask and the win-back message were all written first in Kuwaiti dialect, tested against a formal-Arabic version translated line for line.
Why: The dialect version consistently earned replies where the formal one earned silence, because it read like the brand's own voice rather than a bank circular. Reply rate is the single clearest signal that a WhatsApp channel is alive rather than just delivered.
Set a cadence of two to three broadcasts a month, segmented
We capped one-to-many sends at two to three a month, each segmented by last product bought and recency, replacing the ad-hoc unsegmented sends the founder had tried once.
Why: The one unsegmented broadcast that was tested proved the point on its own: it produced the highest opt-out rate of anything sent in the engagement. A capped, segmented cadence is what keeps a young list from learning to mute the number within its first month.
Match the offer to the repurchase cycle, not a flat discount
Accessory and brush-head reminders went out as care-and-refresh messages timed to actual wear, while a specific restock alert went only to the customers who had asked about that item, rather than a store-wide markdown.
Why: A flat discount trains a list to wait for the next one; an offer tied to a real reason to return, at the point the customer would need it anyway, sells at full price and does not erode the brand's margin on every future launch.
Give lapsed segments a specific reason, not a blanket discount
Sixty and ninety-day lapsed customers were segmented by their original product and given a message tied to something specific — a restock, a new colour, a seasonal bundle — rather than one discount code sent to everyone equally.
Why: A specific reason respects that the customer already knows the brand, which is a different job from convincing a stranger. It also keeps the win-back cost lower than an acquisition ad, because the person already trusts the product.
Measure reply rate, revenue per recipient and opt-out weekly
Every flow and broadcast was scored weekly on reply rate, revenue per recipient and opt-out rate, and the one unsegmented broadcast that spiked opt-outs was retired rather than repeated.
Why: A message that sells but spikes opt-outs is borrowing against the list's future value, and the only way to catch that trade before it costs real revenue is to watch opt-out rate with the same seriousness as any sales number.
What changed
The table above will carry the confirmed figures, and the pattern worth stating now is that the improvement did not come from discounting harder — it came from finally having a channel and using it with restraint. Repeat purchase rate moved because the replenishment and restock messages gave people a specific, timely reason to come back, not because the brand competed on price against itself.
The opt-in work mattered as much as the flows themselves. Framing it around the warranty and care tips, rather than a generic list signup, produced an opt-in rate the founder had not expected from a one-time-purchase product, and it is now the largest owned channel the brand has, ahead of email by a wide margin.
The two losing sends were instructive rather than merely disappointing. The unsegmented monthly broadcast proved that a young WhatsApp list punishes noise fast, and the formal-Arabic template proved that register matters as much as message: the same information, in the brand's own dialect, is a different result entirely.
What we would do next
Build a referral message into the flow for the flagship dryer's buyers specifically, since they repurchase least but gift and recommend most, according to the review comments already sitting on the product page unread by the marketing team.
Second, extend the replenishment cadence into a light subscription offer for brush heads and accessories timed to the wear cycle, so the second order becomes something the customer expects rather than something a message has to convince her of each time.