RFM analysis in plain words: who your best customers really are

RFM analysis in plain words: who your best customers really are

You have a thousand people in your database. Who should you write to? Who should you reward? Who should you win back? If you send the same message to everyone, some customers will unsubscribe — and some of the ones worth keeping, you simply won't notice losing.

RFM solves this with three questions. No formulas — questions any business owner can understand.

The three questions

  • R — how recently were they here? Yesterday or six months ago.
  • F — how often do they come? Once a week or once in a lifetime.
  • M — how much do they spend? Grab a coffee or close a big bill.

The answers to these three questions sort your database into clear groups. And here's the most valuable part: each group needs to be treated differently.

What groups emerge

🟢 Core regulars — come often, were here recently, spend well

There are few of them, but they bring in the most money. The biggest mistake is flooding them with promotions: they'd come anyway, so you're just giving away a discount for free. The right move is recognition: exclusive offers, early access, personal attention. Their job is not to leave. Yours is not to give them a reason to.

Four groups of customers

Coreoften, recently, high spend
Do not disturb. Recognition, not discounts.
Newcomersonce, recently
Give them a reason to return this week.
Dormantused to come often, gone
A certificate with a short deadline.
Occasionalonce, long ago, low spend
Do not spend budget. General newsletter.
Each group is handled differently — that is the whole point of segmentation.

🟡 Newcomers — were here recently, but only once

The most underrated group. This person tried you and is still deciding whether to come back. This is exactly where a second visit is worth the most — and where a short-deadline bonus works best. Miss the window this week, and they'll drift away.

🟠 Dormant — used to come often, haven't shown up in a while

Your former best customers. Winning them back is the most profitable task in your database: they already know the way and already trusted you. This is where an automatic voucher for people who haven't visited in a while works well.

⚪ One-timers — visited once, long ago, for a small amount

No need to spend budget on them. Let a general newsletter remind them you exist — nothing more.

What to do with this in practice

You can't build these lists by hand — by the time you've counted them, they're already outdated. In UDS, the groups update themselves — there's RFM analysis and ready-made segments, and on top of them you can add your own tags: "dairy allergy," "takeout only," "comes with kids."

From there, campaigns go out not to "everyone," but to a segment — turning a blast into a relevant message. Vouchers, by the way, can also be distributed by segment.

Where to start today

Even without fancy tools, ask your database the three questions and find one group — the dormant ones. That's the cheapest money sitting in your database right now.

See how your database breaks down into groups: 7 days free or a 20-minute walkthrough.