Customer segmentation for shops: build audiences from what people actually buy
Sending the same message to every customer wastes the good ones’ attention and annoys the rest. A handful of simple segments built from purchase history — recent, loyal, lapsed, big spenders — turn one list into five useful conversations.
5 min read
In short: segmentation means grouping customers by what they do, so each group gets a message that fits. Start withrecency (when they last bought),frequency (how often) andmonetary value (how much), often calledRFM, and addbirthdays, loyalty tier, product interest andonline behaviour (abandoned carts). In Wameed, an audience combines conditions withAND (for example last visit older than 60 days AND total spent above SAR 500), shows how many customersmatch and how many arereachable with consent on the chosen channel, and feedsWhatsApp template ande-mail campaigns. Marketing is on the Professional plan and above. Only customers who opted in to a channel are messaged.
You do not need a data team. You need five segments and the discipline to use them.
The data you already have
Every phone-number sale at the till or online builds a profile:
| Data | Source |
|---|---|
| Last purchase date | Till and online store |
| Number of purchases | Till and online store |
| Total spend | Till and online store |
| Average basket | Calculated |
| Products and categories bought | Till and online store |
| Birthday | If captured with consent |
| Loyalty tier and points | Loyalty programme |
| Channel | Shop, website, delivery |
| Abandoned carts and browsing | Online store |
| Consent per channel | Recorded at the till or checkout |
The key habit is capturing aphone number for every customer, with consent. Seecoffee shop loyalty andcustomer data and PDPL.
RFM in plain words
| Letter | Question | Example rule |
|---|---|---|
| R: Recency | When did they last buy? | Within 30 days = recent |
| F: Frequency | How often? | 5+ purchases in a year = frequent |
| M: Monetary | How much? | Total above SAR 1,000 = big spender |
Combine them into practical segments.
Five segments that work for most shops
| Segment | Rule (example) | Goal | Message |
|---|---|---|---|
| New customers | First purchase in the last 30 days | Second visit | “Thanks for choosing us; here is something for next time” |
| Loyal regulars | 5+ purchases, last visit within 45 days | Retain and reward | Early access, a thank-you, a tier upgrade |
| Big spenders | Total spend above your top 10% | Look after | Personal attention, preview of new arrivals |
| At risk / lapsed | Last visit older than 60–90 days, previously 3+ purchases | Win back | A reason to return. Seewin-back campaign |
| One-time buyers | One purchase, over 90 days ago | Re-engage cheaply | A light offer or a new-arrival note |
Extras:
- Birthday this month. Seebirthday and milestone campaigns.
- Abandoned cart (online). Seeabandoned-cart recovery.
- Category fans (bought perfume twice, bought oud).
- Branch-based (customers of the Jeddah shop).
Building an audience in Wameed
- Pick a preset (such as lapsed customers or big spenders) or build your own.
- Add conditions, each combined with AND: last visit within or older than N days; total spent above or below; birthday month or the next N days; loyalty tier; online purchases, storefront behaviour or abandoned carts. Up to twenty conditions are allowed.
- Preview: see how many customersmatch and how many arereachable on the chosen channel (the rest lack consent). The preview also shows sample names.
- Save the audience to reuse.
- Attach it to a campaign or an automation.
Audiences are built over till and online purchases together, so a customer who buys in store and online is one person.
Consent and channels
- Customers must haveopted in to the channel you use. Those who have not are skipped and counted.
- Channels areWhatsApp (approved templates only) ande-mail. There is no marketing SMS in Wameed, and push campaigns are not offered.
- Quiet hours apply (21:00–09:00 by default); messages launched in the window are deferred, not dropped.
- Every consent act is recorded as evidence, and customers can opt out by themselves.
SeeWhatsApp opt-in consent andWhatsApp quiet hours and frequency.
Using segments well
- One goal per message. A win-back note is not a promotion for loyalists.
- Different offers for different segments. Don’t give a 20% discount to someone who would have paid full price.
- Control frequency. Don’t message the same person every week.
- Test small. Send to 10% first and compare.
- Measure by segment, not just overall.
- Refresh segments; people move between them.
Measuring
| Metric | Why |
|---|---|
| Delivered / read | Did it arrive? |
| Redemptions of the attached coupon | Did it work? |
| Attributed revenue | What it earned |
| Opt-out rate | Is it too much? |
| Segment size over time | Is your base healthy? |
Campaign reports show delivery through to read, redemptions and attributed revenue. Compare against a holdout group if you can.
Mistakes to avoid
- Messaging people who did not opt in.
- A single blast to everyone.
- Offers that train customers to wait for a discount.
- Stale segments.
- Ignoring online behaviour.
- No test before a big send.
Frequently asked questions
Do I need a data analyst?
No. A handful of simple segments from purchase history is enough.
What counts as “reachable”?
Customers in the audience who have consented to the campaign’s channel.
Can I combine till and online data?
Yes; the audience engine spans both.
Which plan has audiences?
Professional and above.
- #تقسيم العملاء
- #الجماهير المستهدفة
- #تسويق حسب سلوك العميل
- #قاعدة العملاء
- #customer segmentation
- #marketing audiences
- #RFM retail
Share this article
Ask about your own shop
Thirty minutes on your products, your tax setup and your hardware — not a slide deck.
Keep reading
7 min read
White Friday and National Day sales: plan stock, price, staff and systems so the peak makes money
The biggest sales days are also the days when tills crash, stock runs out and margins vanish. A six-week plan for stock, honest pricing, staffing, systems and customer messages turns the peak into profit.
6 min read
Measuring social ads for a shop: from clicks to real sales, without fooling yourself
An advert that gets clicks is not an advert that sells. Track UTM links, pixel events, coupon codes and your own analytics, count only paid orders, and compare cost per sale with margin. Here is a simple measurement system.
5 min read
Loyalty programmes: most fail, and the ones that work are simple
A programme nobody can explain at the counter in one sentence will not be used. Here is what separates the ones that change behaviour from the ones that just cost margin.

