How to spot at-risk customers before they stop buying

Most customers do not leave a shop with a complaint. They simply come in less often, then stop. By the time anyone notices, they have been buying elsewhere for months. The earlier you see the change, the better your chance of finding out why and winning them back.

This guide shows how to recognise customers who are drifting away, how to group customers by the way they buy, and how to plan retention contact that respects their choices. You can apply it with a customer notebook, a spreadsheet or a retail system with customer records.

Start with the customers you can recognise

You can only follow customers you can recognise when they buy. That usually means a customer record linked to their purchases, through a loyalty card, a phone number given at the till or an account for credit customers. Ask for details only when the customer is happy to give them, and record whether, and how, they agree to be contacted.

Three signals worth watching

Customer analysts often use three simple measures, sometimes called recency, frequency and value:

• Recency: how long since the customer last bought something.
• Frequency: how often they usually buy.
• Value: how much they usually spend.

None of these alone tells you a customer is at risk. The signal is a change from that customer's own normal pattern:

• a weekly customer who has not been in for three or four weeks;
• a customer who still comes, but now buys only one or two small items;
• a customer who always bought from one category, such as baby care or phone accessories, and has stopped buying it;
• a loyalty member who has stopped collecting or using points.

Comparing each customer with their own history matters. A customer who buys once a month is not at risk after three weeks; a daily customer might be.

Group customers by buying pattern

Grouping customers helps you decide where to spend your time. A practical set of groups:

• High-value repeat customers: buy often and spend well. Look after them first.
• New customers: one or two purchases so far. The second and third visits decide whether they become regulars.
• Loyalty-active customers: collecting and using rewards.
• At-risk customers: regulars whose pattern has recently changed.
• Lapsed customers: have not bought for well beyond their usual gap.

Review the groups regularly. Customers move between them, and the movement is often the most useful information.

Understand why before you act

Before contacting an at-risk customer, look at their history for clues:

• Did they stop buying a product that was out of stock several times?
• Did a price change or the end of a promotion coincide with the drop?
• Do they live near a branch that changed opening hours or staff?
• Did they have a complaint, a return or a refund recently?

Often the reason is something you can fix, such as a product you no longer carry or a recurring stockout. Fixing the cause helps every customer affected by it, not only the ones you contact.

Plan retention contact that respects consent

When you decide to contact customers, keep it relevant and respectful.

1. Contact only customers who agreed. Check each customer's marketing consent and preferred channel, whether SMS, WhatsApp or email. Customers who did not agree should not receive marketing messages.
2. Make the message fit the reason. A customer who stopped buying baby care because of stockouts wants to know the product is back, not a general discount.
3. Keep the audience small and specific. A message to twenty lapsed customers of one category is more relevant than a message to everyone.
4. Have someone approve before sending. Review the audience, the message and the channel before anything goes out.
5. Record the result. Note who came back, so you learn which approaches work.

Messages that are too frequent or irrelevant do more harm than good. They teach customers to ignore you, or to withdraw their consent.

Do not forget the customers who stay

Retention is not only about the customers who are leaving. Thank your high-value regulars, make sure your best-selling products are reliably in stock, and give new customers a reason to come back a second time. These habits prevent many customers from drifting away in the first place.

Frequently asked questions

How long without a visit makes a customer "at risk"?

It depends on the customer's usual pattern. A useful rule of thumb is to look closer when a customer's gap since their last purchase is clearly longer than their usual gap between purchases, for example twice as long.

Should I offer a discount to every at-risk customer?

No. A discount can help, but it is often not the reason the customer left. Find the likely cause first; restocking a product or resolving a complaint may matter more than a discount.

Can a system contact at-risk customers automatically?

We recommend that every campaign is reviewed and sent deliberately by a person. A system can suggest who to contact and why, but your team should decide what is sent and to whom.

Conclusion

Spotting at-risk customers is about noticing changes in each customer's own pattern, understanding the reason and responding in a way that respects their consent. A small, regular review of your customer groups catches problems while they can still be fixed.

In NexaPOS, ML Insights groups customers into high-value, new, loyalty-active, at-risk and lapsed customers, explains why each at-risk customer was flagged, and previews which customers can be reached with their consent. Recommendations never send a campaign automatically; your team creates and approves every campaign.