The scenario
The brand sells baby health essentials — thermometers, nasal aspirators, feeding hygiene kits — alongside a household vitamin range, in Kuwait, direct to consumer, on a Shopify store with KNET, Tabby and cash on delivery at checkout. Three years in, a founder-run team, and a customer base of mostly new and expecting mothers who found the brand through a WhatsApp group referral rather than an ad.
The first order was rarely the problem. A newborn's first thermometer, a first bottle of children's vitamins after a paediatrician visit — the brand converted those well. What happened after delivery was nothing: no message, no reminder when the vitamin bottle was due to run out, and no channel the brand actually owned beyond an order-confirmation email most customers never opened.
- Monthly revenue band
- pending client sign-off
- Average order value
- pending client sign-off
- Fulfilment
- Own stock, local courier, next-day delivery across Kuwait
- Team
- Founder plus one part-time customer service agent on WhatsApp
A refill business with no idea when anything needed refilling
Nearly everything the brand sells runs out on a schedule: a bottle of children's vitamins lasts about a month, thermometer batteries and probe covers wear down, and nasal aspirator refills are needed on their own rhythm. None of that schedule existed anywhere in the business. Every repeat order was the customer remembering on her own and searching for the brand by name again.
When we read the cohorts by first product bought, the picture was not what the founder expected. The bestselling item, a starter thermometer kit, had one of the weakest repeat rates in the catalogue, while a lower-priced vitamin sample size that almost nobody bought first quietly had the highest return rate of any product in the store — a signal the brand had never once looked at.
There was no WhatsApp channel doing any of this work, despite WhatsApp being where nearly every customer question and referral already happened. The brand had built a genuine gateway product without knowing it, and it had no way to reach the person who bought it again once the box left the warehouse.
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 twice: checkout and delivery confirmation
Winner"Get your delivery updates on WhatsApp — one tap, no forms."
- Format:
- Opt-in checkbox at checkout, repeated as a one-tap link on the delivery-confirmation page
Framing the opt-in around delivery updates rather than marketing removed the hesitation new mothers showed toward a generic newsletter sign-up, and asking a second time on the delivery-confirmation page caught people who had skipped it at checkout while they were focused on the address. This became the channel the flows below were actually built on.
A replenishment check-in timed to the actual product cycle
Winner"Your little one's vitamins are probably running low about now."
- Format:
- WhatsApp template message sent close to the average time a bottle runs out, per product
This single message, timed to the product's own cycle rather than a fixed calendar date, carried more of the repeat orders than any broadcast the brand had tried before. It reads as a helpful reminder from someone who knows the product, not a sale, which matched exactly how the brand's existing WhatsApp-group reputation already worked.
The gateway-product upsell on the order-confirmation flow
Winner"Most families who tried this also keep this size on hand."
- Format:
- One-line suggestion inside the order-confirmation WhatsApp message, not a separate broadcast
Once the cohort read showed which product actually predicted a second order, surfacing it inside a message the customer was already reading — rather than a new broadcast she might mute — quietly lifted the number of customers who bought the true gateway item, without adding to the message volume anyone had to manage.
A weekly general wellness broadcast to the full list
Lost"This week's health tip for your family."
- Format:
- Unsegmented weekly WhatsApp broadcast to every opted-in number
This was the first thing tried, based on the idea that more contact means more recall, and it produced the highest opt-out rate of anything sent during the engagement. A weekly message with no connection to what the recipient had actually bought reads as noise, and it taught a visible share of the list to mute the number within the first month.
What we did — the optimizations, in order
Read the cohorts by first product, not by month
Repeat purchase rate at thirty, sixty and ninety days, split by the very first item each customer bought, rather than a single blended repeat-rate number for the whole store.
Why: A blended number would have kept pointing the brand back at its bestseller, which was actually one of the weaker repeat drivers. Splitting by first product is what surfaced the real gateway item and told us where to put the first flow.
Capture WhatsApp opt-in twice, framed around delivery not marketing
An opt-in on the official WhatsApp Business API at checkout, framed as delivery updates, repeated once more on the delivery-confirmation page for anyone who had skipped it.
Why: A checkbox that says "marketing" gets skipped by new mothers who already feel oversubscribed to baby-brand newsletters. Framing it as the channel their delivery updates actually arrive on made the opt-in itself the first thing that worked.
Build the flows around each product's own cycle
Order confirmation, delivery update, a replenishment check-in timed to the average time each specific product runs out, the review ask, then a gateway-product suggestion — written in Kuwaiti dialect throughout.
Why: A single flow for the whole catalogue would have sent a vitamin reminder a month too early or a thermometer message that never fires, since the two products run out on entirely different schedules. Building around each product's own cycle is what made the replenishment message read as helpful rather than random.
Segment the cadence instead of broadcasting to everyone
The unsegmented weekly broadcast was replaced with two to four messages a month, each sent only to the segment whose last product and timing actually matched the message.
Why: The weekly general broadcast had produced the highest opt-out rate of anything sent, and it was the clearest evidence in the whole engagement that volume without relevance trains a list to mute the number rather than buy from it.
Match the offer to the repurchase cycle, not a flat discount
Vitamin refills were offered as a small subscription rather than a one-off discount, while thermometer and aspirator accessories were offered as a care-and-refresh reminder tied to age milestones rather than a calendar promotion.
Why: A consumable and a durable accessory do not repurchase the same way, and offering both a flat percentage off trains the customer to wait for the next discount instead of buying on the schedule the product actually needs.
Build the win-back for lapsed cohorts by product, not a blanket discount
Sixty and ninety-day lapsed segments, each given a specific reason tied to their own first product — a running-low reminder for vitamin buyers, a seasonal readiness message for thermometer buyers.
Why: A blanket discount to everyone who has not ordered in ninety days teaches the whole list to wait it out. A specific, product-linked reason to come back reaches people at the moment their own need actually resurfaces.
Measure repeat rate, revenue per recipient and opt-out together
Every message's revenue per recipient was read alongside its opt-out rate, not on its own, and the general weekly broadcast was killed the moment its opt-out rate crossed the agreed threshold.
Why: A message can look like it is selling while quietly burning the list it is selling to. Reading the two numbers together is what caught the weekly broadcast early enough to still have a healthy list left for the flows that actually worked.
What changed
The headline number is in the table above, and it moved because the brand finally had a channel timed to when a customer actually needed to buy again, not because of a single clever message. Revenue per recipient rose alongside the repeat rate, which matters: a channel that only grows engagement without growing revenue is a vanity metric in disguise.
The clearest lesson was about the gateway product. Once the cohort read pointed at the real driver of a second order rather than the bestseller, the upsell built around it quietly outperformed every discount the brand had tried before, because it was recommending something the data said this specific customer was likely to need rather than something the brand wanted to move.
Opt-out rate told the other half of the story. It rose sharply on the one unsegmented broadcast and fell back to a low, stable level once the cadence was segmented by product and timing, which is the clearest evidence that relevance, not frequency, is what a Kuwait WhatsApp list actually rewards.
What we would do next
First, extend the replenishment cycle logic to every product in the catalogue rather than the two it launched on, and start testing subscription pricing on the vitamin range specifically, since its repurchase pattern is the most predictable in the store.
Second, build a referral flow off the back of the WhatsApp-group behaviour that already brings new mothers to the brand, since the channel doing that work for free deserves a message of its own rather than staying invisible in the results.