Customer Segmentation with RFM: A Simple Way to Know Who to Contact and Why

A shop sends the same promotion to every number it has. Loyal customers who would have bought anyway receive a discount they did not need. People who bought once two years ago receive a message they do not remember asking for. And the customers who are about to drift away receive nothing that speaks to them. Customer segmentation with RFM is a simple method for treating different customers differently, using data you already have.

Three questions

RFM sorts customers by three measures:

• Recency. How recently did they last buy?
• Frequency. How often do they buy?
• Monetary value. How much do they spend?

A customer who bought last week, buys every month and spends well is very different from one who bought once, long ago, for a small amount. Each deserves a different message, and some deserve none.

What you need

A table of sales with three columns:

• Customer, identified by phone number or customer ID.
• Date.
• Amount.

Twelve months of history is a good window for most businesses. Use a longer one if customers buy rarely, as with furniture or annual services.

Calculating RFM in a spreadsheet

Step 1: summarise by customer

For each customer, calculate:

• Days since last purchase. Today's date minus their most recent purchase date.
• Number of purchases in the period.
• Total spent in the period.

A pivot table does this in a minute.

Step 2: score each measure

Give each customer a score from 1 to 3, or 1 to 5 for more detail, on each measure.

A simple three-level version:

Recency

• 3: bought in the last 30 days.
• 2: bought 31 to 90 days ago.
• 1: more than 90 days ago.

Frequency

• 3: five or more purchases.
• 2: two to four.
• 1: one.

Monetary

• 3: top third by total spend.
• 2: middle third.
• 1: bottom third.

Set the thresholds from your own data and buying cycle. A coffee shop and a furniture store need very different numbers.

Step 3: combine

Each customer now has three scores, such as 3-3-3 or 1-1-2. Group them into a handful of named segments.

Five practical segments

Champions

High on all three. They bought recently, buy often and spend well.

What to do: thank them personally. Give early access to new products. Ask for reviews and referrals. Do not discount to them by reflex, since they are already buying.

Loyal

Good frequency and recency, moderate spend.

What to do: suggest related products or larger packs. A simple loyalty reward. Keep the relationship warm.

New

Recent first purchase, low frequency so far.

What to do: the second purchase is where most customers are lost. Check that the first order went well, share how to use or care for the product, and send a timely reminder or a small incentive for the next order.

At risk

They used to buy often or spend well, and their last purchase is getting old.

What to do: the most valuable segment to act on. Contact them personally. Ask whether something went wrong. Remind them of what they liked. An offer can be justified here.

Lapsed

Low on recency and frequency. One or two purchases long ago.

What to do: one win-back message, clearly relevant. If there is no response, stop. Repeatedly messaging people who have gone harms your sender reputation.

Some businesses add a sixth group, big spenders with few purchases, who bought a large amount once or twice. They merit a personal approach.

Turning segments into action

Segmentation is useful only if it changes what you send. Build a simple plan:

• Monthly: recalculate the segments.
• Champions: one personal thank-you each quarter.
• New: an automated check-in after delivery and a reminder at the usual reorder interval.
• At risk: a weekly list of ten to contact personally.
• Lapsed: one win-back campaign each quarter.

Send only to customers who agreed to receive messages, and always offer a way to stop.

Watch the movement

The most useful view is who moved since last month:

• New to Loyal: onboarding is working.
• Loyal to At risk: find out why before they go.
• At risk to Lapsed: win-back did not work, or was not tried.
• Lapsed to active: learn what brought them back.

A good customer starting to slip is the earliest warning a business gets.

What RFM tells you about the business

• Concentration. If a small group provides most of your revenue, protect it, and recognise the risk.
• Leaky first purchase. Many one-time buyers point to a problem with the product, the delivery or the follow-up.
• Size of the dormant list. How much business a good win-back effort could recover.
• Where acquisition works. Compare the segments of customers from different sources. A channel that brings many one-time buyers is worth less than it looks.

Limits

• RFM looks at the past. It does not know why someone stopped buying.
• It treats all products alike. A customer who buys a durable item every three years is not lapsed.
• It ignores margin. A big spender on discounted goods may be less profitable than a smaller regular.
• It needs clean customer identification. Duplicates split one customer's history across several records.

Use it as a starting point, and add what you know about each customer.

Common mistakes

• Too many segments. Five that you act on beat twenty that you only chart.
• Thresholds copied from another business.
• Calculating once and never updating.
• Sending discounts to champions.
• Ignoring the new customers, where the largest losses occur.
• No consent. Segmentation does not create permission to message.

Tools

A spreadsheet handles a few hundred customers comfortably. As the list grows, a CRM that records each purchase against the customer and recalculates the groups saves the monthly effort and lets you act on a segment directly. Polynusa keeps customers, their purchases and the WhatsApp conversation in one workspace, with tags you can use for segments and follow-ups. See the pricing page.

Summary

Score each customer on how recently, how often and how much they buy. Group them into champions, loyal, new, at risk and lapsed. Thank the best, help the new towards a second purchase, contact those at risk personally, and try once to win back the lapsed. Update monthly and watch who moves.

FAQ

What does RFM stand for?

Recency, frequency and monetary value. Recency is how long ago a customer last bought. Frequency is how many times they have bought in a period. Monetary value is how much they have spent. Together they describe how valuable and how engaged a customer is.

How much data is needed for RFM segmentation?

A list of sales with customer, date and amount, covering at least six to twelve months. A few hundred transactions are enough for a small business to see useful groups. The method works in a spreadsheet and needs no special software.

How often should RFM segments be updated?

Monthly for most small businesses, or before each campaign. Customers move between segments as they buy or stop buying, and those movements are useful signals, especially a good customer starting to slip.