RFM Analysis: How to Segment Restaurant Customers
RFM analysis groups customers by last order, order count and amount spent. A simple example showing whom to win back and whom to reward with a bonus.
RFM analysis rates a customer on three questions: when they last ordered (Recency), how often they order (Frequency) and how much they have spent (Monetary). As a result your best customers, new customers and customers drifting away are grouped separately, and you have an offer for each.
Sending the same promo to every customer wastes money. A loyal customer doesn't need a discount, while a customer drifting away needs more than a plain message. Below we explain what RFM is and how to use it.
Key points:
- R is how long it has been since the last order.
- F is how many times they have ordered.
- M is how much they have spent in total.
- The offer depends on the group: who to win back, who to reward.
What is RFM and why do you need it?
RFM is a way to rate a customer by last order time, number of orders and money spent. It shows what your customer base is made of: who is loyal, who is new, and whom you are about to lose.
It needs customer data: who ordered what, and when. At an aggregator you don't have this data, while in your own channel every order goes into the CRM. So RFM is for restaurants that have their own customer base.
What are the main groups?
Customers are usually divided into five groups: best, loyal, new, at risk and lost. You set the exact names and boundaries yourself.
| Group | What the customer looks like | What to do |
|---|---|---|
| Best | ordered recently, often, spent a lot | say thank you, offer new dishes first, no discount needed |
| Loyal | orders regularly | keep them with loyalty (points, cashback) |
| New | just made a first order | give a small bonus on the second order |
| At risk | used to order often, now silent a long time | send a reminder and a personal offer |
| Lost | haven't ordered for a long time | one last try, or stop |
How is it calculated?
Three numbers are taken for each customer: days since the last order, number of orders and total amount. You score each from 1 to 5 (5 is best) and combine the scores. For example, 5-5-5 is the best customer, 1-4-4 is a customer who used to be good and is now drifting away.
This can be done in Excel, but when customers grow in number, it is hard by hand. In Delever CRM, RFM segmentation is ready: the customer profile shows order history, average check and activity.
How do you message a segment?
Each group gets its own text and offer. To a customer at risk you can say "we miss you, a bonus on your next order". To your best customer, "be the first to try a new dish" works well: you show respect without giving a discount.
The message channel matters too: SMS, push or Telegram. In Delever you can send to the base split by segment (RFM, last order time, average check). SMS providers (Eskiz, PlayMobile, SMS-center, SMS-fly) connect under your own contract. We covered win-back messages separately.
What are the most common mistakes?
The first mistake is sending the same discount to everyone. The second is writing endlessly to lost customers. The third is building segments once and never recalculating them.
Customer status changes: today's loyal customer may be at risk in three months. Update segments once a month. When giving a bonus, don't forget your margin: how to calculate a loyalty program.
Conclusion
- Check whether you have a customer base: phone numbers collected in your own channel.
- Group customers by the three indicators.
- Send each group its own offer and measure the result.
We show how to see RFM segments in Delever: delever.io or @delever_bot on Telegram.
Frequently asked questions
What is RFM analysis?
A way to group customers by last order time (R), number of orders (F) and amount spent (M).
What data does RFM analysis need?
For each customer: order date, count and amount. This data is collected in the CRM through your own channels (website, bot, app).
How often should RFM be updated?
Once a month is enough. Customer status changes, so update segments regularly.
Can I do RFM for aggregator customers?
It is hard: the aggregator doesn't give you customer data. That is why you need to move customers to your own channel.
Want to see Delever on your own venue? Book a free 30-minute demo at delever.io or message @delever_bot on Telegram.