RFM ANALYSIS
See Who Is About to Leave, Before They Do
Three numbers about every customer — how recently they came, how often, and how much they spend — sorted into ten groups that build themselves. No spreadsheets, no analyst, nothing to maintain.
Ten Groups, Sorted Before You Open the Screen
This is the RFM filter in your customer base. Every count is live, so you know how many people are in a group before you click it — and what to do about them once you have.
Customer base
Bought recently, buy often, and spend the most.
Recognition, not discounts. They were already coming.
Sample counts from a coffee shop after six months. Tap any segment chip to filter the table — the panel below is our guidance, not part of the app.
Three Questions, Asked of Every Customer
RFM has been used by retailers for forty years because it works with almost no data. Three numbers, and every customer sorts themselves into a group.
Recency — when were they last in?
The single most predictive number about a customer. Someone who came last week is worth far more attention than someone who spent twice as much a year ago.
Frequency — how often do they come?
Habit is what loyalty is actually made of. Two visits a month for a year beats one enormous purchase, for almost every business on this list.
Monetary — how much do they spend?
The one people reach for first, and the least useful on its own. It only means something next to the other two.
Then the groups appear
Combine the three and everyone lands somewhere — Champions, Sleeping, At risk. Nobody sorts anything by hand, and the counts update as people visit.
Who Lands Where, and What to Say
Two of these need money spent on them. Two need nothing at all. Knowing the difference is the whole value of this page.
Six Decisions This Actually Changes
Segments are only worth having if something happens differently because of them.
Where the discount goes
Not to Champions
The instinct is to reward your best customers with the biggest offer. They were already coming. Send the discount where it changes an outcome instead.
Who gets the campaign
148 people, not 1,284
A win-back sent to everyone reaches people who came in yesterday. Sending to the right group is better marketing and, on paid channels, nine tenths cheaper.
When to intervene
Need attention, before At risk
There is a window where a lapsing customer is still winnable. This is the screen that shows you the window while it is still open.
What to test, and on whom
Medium is your sandbox
Your largest and most ordinary group. Try a new offer here first — if it moves them, it will move most of your base.
When to stop spending
Inactive costs you nothing to ignore
Not every customer is worth chasing. Recognising the group that has genuinely gone saves you money on every paid send afterwards.
Who to know by name
Ninety-six people
Your Champions are a small enough group that your team could learn their faces. That is worth more than any automated message you could send them.
Settings Most People Never Find
RFM is included on every plan. These are the parts that decide whether the groups tell you the truth.
Set your own thresholds
What counts as “recent” for a coffee shop is not what counts for a garage. Until you set the boundaries to your trade, every group on the screen is slightly wrong.
Counts before you click
Each segment shows its size on the chip itself. You know whether a campaign is worth writing before you have written it.
Message a segment directly
Filter to a group, then send push, email, or SMS to exactly those people. The segment is the audience — nothing to copy out or re-upload.
Rewards for one segment only
On reward cards you can restrict a reward to a single group. A Champions-only perk costs nothing to offer and means a great deal to receive.
Spend has to be entered
The monetary third of RFM comes from the purchase amount your team types at the scanner. Skip it and the groups are built on two numbers, not three.
It re-sorts itself
Nobody maintains these lists. A Sleeping customer who walks in tomorrow moves group on their own, and the counts move with them.
Four Ways to Read This Wrong
RFM is simple enough to be dangerous. These four mistakes are the common ones.
Leaving the thresholds on default
“Sleeping after thirty days” is right for a café and absurd for a garage nobody visits twice a year. This is the first thing to change and the thing almost nobody changes. Until you do, every count on the screen is quietly misleading.
Judging it in the first month
RFM needs history to mean anything. In week two everybody is a Beginner, and that is correct rather than broken. Give it a full customer cycle — two or three months for most businesses — before you build a campaign on it.
Discounting your Champions
The most expensive habit on this page. They were already coming, so every discount you send them is margin you gave away for visits you had. Give them access, recognition, or a name at the counter instead.
Chasing Inactive because the number is big
A large Inactive group looks like an opportunity and is usually just history. On paid channels it is the fastest way to spend money on people who left long ago. One last offer if you must, then leave them out.
Find Out Who Is About to Leave
RFM segments are included on every plan, including the entry tier, and they build themselves from ordinary scans at the counter. The trial costs nothing and needs no card.