Loyalty Programs That Actually Work for Small African Retailers

For two years, Mama Wanjiru bought milk, bread and airtime from the same duka in Umoja every morning on her way to the matatu stage. Then a new shop opened one street closer to the stage. She did not complain and she did not say goodbye. She simply stopped coming, and the owner only noticed three weeks later, when the morning takings looked thin and he could not say why.

That is how small shops lose money. Not in one dramatic theft, but one quiet regular at a time. If you run a duka in Nairobi, a provisions shop in Lagos or a cold store in Kumasi, your survival does not depend on the crowd that passes your door. It depends on maybe forty people who come back every week. A loyalty program is nothing more than a deliberate plan to keep those forty and slowly grow the number. This guide is about doing that without plastic cards, without an agency, and without giving away your margin.

Take two customers in a Lagos provisions shop. The first walks in because you ran a promo, spends 4,000 naira and never returns. The second spends 2,500 naira every week because she trusts your prices and you stock the brands she likes. After six months, the promo customer gave you 4,000 naira. The quiet regular gave you around 60,000, and she is still coming.

Now add what she did not cost you. No flyer, no boosted post, no discount to pull her through the door. Winning a stranger costs money every single time; keeping a regular costs you consistency and a little attention. The figures change with the country, a 300 shilling basket in Nairobi, a 50 cedi basket in Accra, but the ratio holds everywhere: a weekly customer is worth ten to twenty one-off buyers over a year. So the sharp question for a thin-margin shop is not how to get more people through the door. It is how to make the people already buying come back one more time each month.

The stamp card is a lovely idea and a terrible object. The idea works: once a card is half full, a customer feels invested and comes back to finish it rather than start from zero somewhere else. That pull is real, and you should keep it.

The object fails. Cards dissolve in a rainy-season pocket, hide in the second handbag, or go through the wash. A customer swears she had eight stamps when she had five, and you have no record either way. Worse, the card teaches you nothing: you cannot see who your best buyers are, what they take, or when they stopped coming. My honest advice is to not print stamp cards in 2026 at all. Keep the buy-ten-get-one logic, it is excellent. Just keep the count somewhere that cannot fall out of a pocket.

Nobody on this continent wants another plastic card. Everybody has a phone number, remembers it, and carries it everywhere by definition. That makes the number the natural loyalty account. At the counter you ask, which number should I put your points on, and that is the whole enrolment. Nothing to print, nothing to reorder, nothing to forget at home.

The number also happens to be the most useful identifier in African retail. It is the M-Pesa line in Kenya, the MoMo wallet in Ghana, the name on a Moniepoint transfer in Nigeria, and the WhatsApp contact, all in one. A customer who gives you her number has not just joined a points scheme. She has opened the one channel that can bring her back next month. A plastic card can do none of that.

Only three loyalty models are worth a small shop’s time. Points on every purchase, redeemed later against goods. A free item after a set number of visits: buy nine, the tenth is on the house. Or a plain regulars-only perk, like first call on stock that runs out fast.

The test for choosing between them is brutal and simple: can the customer repeat the rule back to you in one sentence? Buy nine, get the tenth free passes the test. Earn one point per hundred naira, reach silver at five hundred points, double points on Tuesdays does not. If you sell small items at steady prices, count visits. If baskets swing a lot, a bottle of soda one day and a full month of provisions the next, count spend with points instead, so the big basket earns more than the small one.

Whatever you do, resist copying the scheme of a supermarket chain. A tiered program with expiry rules needs a marketing department to run it. You need a sentence.

Your customers already pay you through an M-Pesa till, an Opay or PalmPay transfer, MTN MoMo, or cash. A loyalty program that ignores this becomes an extra ritual at the counter, and anything that slows a Saturday queue dies within a month.

So attach loyalty to the payment itself. When someone pays to your till number, her phone number is right there in the confirmation. A bank transfer in Lagos carries a name you can match. Earning credit should take zero extra seconds for the customer and at most one tap for whoever is serving. And record cash sales against the same customer when she gives her number, because a regular is a regular whether she pays from her phone or with notes. If your records split her into a cash person and a MoMo person, half her history disappears and your best customer looks like two average ones.

Never push a form across the counter. Attach the question to the benefit, in the moment: let me put this on your rewards so today counts, which number should I use? Said that way, the customer is getting something, not filling something. Then state plainly what you will do with the number, points plus the occasional message when her regular items land, and what you will never do, which is sell it or flood it. In markets where everyone gets scam calls, that one sentence of honesty is what separates you from the noise.

In Kenya, Nigeria or Ghana, WhatsApp is where a message from a shop actually gets read. Not SMS, which people ignore, and not email, which they never open. That reach is exactly why you must handle it with care: a customer who mutes or blocks you is gone from that channel for good, and no reward wins her back.

Two messages a month is plenty. Send only what the reader herself would call useful: the Peak milk you buy every week is back, or a small end-of-month rice offer reserved for regulars. Send it to the people who buy that item, not to the whole list. Use a broadcast list rather than a group, so nobody sees anyone else’s number or replies.

Run one test before every send: if this message arrived from a shop you visit, would you be glad or annoyed? If the honest answer is annoyed, delete it. The channel is worth more than any single promotion you will ever push through it.

The lazy version of loyalty is a discount, and on thin margins it is slow poison. Knock ten percent off a bag of rice every week to reward loyalty and two things happen: you train customers to wait for the cut, and you hand away most of what you earned on the item. That is not loyalty. That is a price war you started against yourself.

Reward with things that cost less than they feel. A free low-cost item after ten visits: the soda costs you 40 shillings and lands like a gift. First call on stock that sells out, which costs you nothing at all. The small extra dropped in a top regular’s bag, the dash in Lagos and Accra, the nyongeza in Nairobi. Points that take a dozen visits to redeem, so eleven purchases fund the reward before it costs you once. Loyalty means being the shop people choose even when you are not the cheapest. If every reward is a price cut, you are teaching them the opposite lesson.

Before you spend one shilling on rewards, remember the one that is free: greet the customer by name and remember what she takes. The supermarket at the mall cannot do that. When you hand Mama Nkechi her usual two cartons before she asks, you have given her a reason to walk past three cheaper shops on her way to yours.

The limit is your memory. You hold your top ten regulars in your head, not a hundred, and the day your cousin covers the counter, everything you know stays home with you. That is the real argument for writing loyalty down. Not to replace the warmth, but so the warmth survives your absence.

A satisfied regular telling her neighbour to buy from you converts better than any signboard, because the trust arrives before the customer does. Put a small reward behind it: when a regular sends someone who becomes a customer, both get extra points or a small item on the next visit. On one street in Surulere or one estate in Ruaka, word of mouth compounds fast, and it costs a fraction of what any advert would.

A loyalty program you never measure is a mood, not a program. Three numbers settle the question. What share of this week’s sales came from known customers rather than strangers. How many regulars came back this month compared with the month before you started. And which rewards get claimed: if the referral bonus has moved nobody in eight weeks, drop it and put that value into the reward people do chase.

Expect one uncomfortable discovery. In most small shops, about twenty customers quietly produce close to half the sales. Once you can see their names, you know exactly whose milk to reserve, and whose absence to notice within days instead of weeks.

Everything above can start in a notebook, and for a very small shop a notebook is fine for a while. It stops being fine when the queue is long, a points count is disputed, or you are away for a week.

This is the job a point of sale does quietly. With digabloPos, the phone number is the customer account: points build on each sale by themselves, whoever is behind the counter can see the balance, and cash and mobile money land in the same customer history instead of splitting one person in two. It keeps working when the network drops, which on a Lagos street or during a Nairobi outage is not a detail, and it syncs once the connection returns. It also shows you the best-customers report, so the twenty names carrying your shop stop being a guess. The loyalty stays yours. The tool just makes sure it survives a busy Saturday.

Do not plan a launch. Do this instead. Today, choose one rule and one reward: buy ten, get one free, with a reward that costs you little. Write the rule on a card by the till so your staff say it the same way every time. From tomorrow, ask every paying customer for a number, attached to the benefit. Within two weeks you will have forty or fifty numbers, which is enough. Send one WhatsApp broadcast, restock news only, to the people who buy that item, and watch who walks in. After a month, look at your three numbers, keep what pulled people back, and cut what did not.

The cost of starting is close to zero: digabloPos is free to start, and the loyalty follows the numbers you are already collecting at the counter. The cost of waiting is another Mama Wanjiru, gone three weeks before you notice.