RFM segmentation: stop blasting the same email to your entire list

There's one thing that drives me nuts when I look at the CRM of most brands we meet. A list of 40,000 people, and one single email going out to everyone. The customer who bought 8 times this year gets the exact same message as the guy who downloaded a promo code 18 months ago and never bought a thing. There it is. Same content, same offer, sent at the exact same time.

No wonder the numbers plateau. Obviously, if you write to everyone, your message lands right for nobody.

Done well, RFM segmentation fixes this without building some monstrous machine. Let me walk you through how it works, then the trap everyone falls into, over-segmentation, which eats a ton of your time for zero ROI.

What RFM actually means

RFM stands for Recency, Frequency, Monetary value. Nothing more.

• Recency is the date of someone's last purchase. Yesterday or two years ago changes everything.
• Frequency counts how many orders they've placed with you.
• Monetary adds up everything they've spent in total value.

On each criterion, you assign a score. Usually 1 to 3 to keep it simple. That gives you a ratio like 3-3-3, or 1-2-3, or 1-1-1.

A 3-3-3 customer bought recently, buys often and spends a lot. That's your VIP. That's the person you want to spoil, the one you push private sales to first, the one you reserve new collections for before anyone else.

At the other end, you've got the 1-1-1. One purchase two years ago, just once, a $10 basket. You're clearly not going to talk to them the same way. Either you wake them up with a win-back sequence, or you stop targeting them so you don't wreck your deliverability.

That's the whole point of RFM: it's instantly readable. You don't need a data scientist to understand that a 3-3-3 beats a 1-1-2.

The trap of over-complicated segmentation

Let me be clear on this because it's the number one mistake I see. Segmentation that's too complex is useless. You end up with 27 segments, you lose track, your teams build hyper-personalised messages for micro-populations, and in the end the ROI just isn't there.

You spend three days building a flow for 40 people. That's not worth it, time-wise.

What I recommend, especially at the start, is splitting your base into three big blocks. It's basic but it works well.

• Leads, the ones who've never bought. They're in your list, they handed over their email through a pop-up or a form, but they've never pulled out their card.
• One-time buyers, the ones who bought once. Your goal with them is simple, you want the second order.
• Repeat buyers, the ones who've bought several times. These are becoming loyal, you need to nurture them.

When you look at the breakdown, you often land on numbers that sting. Like people who bought once making up 35 to 40% of your subscriber base, and those who bought several times barely 10%. The rest is leads who never converted.

Just seeing those three rates, you already know where to put your energy. Each block calls for its own actions. For leads you're chasing the first order. For one-time buyers, the repurchase. For repeat buyers, the job is not letting them slip away.

Once those three are running smoothly and you've got them down, that's when you can refine with real RFM scoring, 3-3-3, 2-3-1 and the rest. But not before.

Applying RFM in your campaigns and flows

Segmentation is your foundation. You need a clean list and decent deliverability, otherwise everything else collapses. But segmentation is worthless if you don't actually use it in your sends.

In the welcome series

Take the welcome series, it's obvious here. Someone enters the flow and has never bought, you push them the long journey, you educate them on the brand, on the product, you build the relationship before selling.

But a customer who's already ordered twice and comes back through this flow, you're not going to hit them with the same intro emails. You already have plenty of data on them. So you can do smart cross-selling, suggest products that complement what they've already bought. A genuinely specific message.

Splitting the welcome into Level 0 for the lead, Level 1 for the one-time buyer and Level 2 for the loyal customer changes everything in terms of message relevance.

In newsletter campaigns

As a standard we run around 8 campaigns a month, two newsletters a week. On premium we go up to 12, three a week. Beyond that it gets complicated, you're forcing the pace, you send for the sake of sending and you've got nothing left to say. You just need to have something to talk about and not force it.

During the big moments, a private sale launch, a new collection, Christmas, that's where segmentation really makes the difference. You can target people who engaged with your latest emails, or on the flip side re-engage those who didn't open. You can reserve early access for your 3-3-3s and push a more aggressive offer to one-time buyers to trigger the repurchase.

In Always On flows

Beyond the welcome, all your automated flows benefit from being plugged into segmentation. Cart abandonment, checkout abandonment, post-purchase, win-back.

Win-back depends directly on RFM. Recovering a customer who hasn't been active in a while. And to trigger it properly, you take the average time between two orders across your base. If your customers rebuy every 90 days on average and someone blows well past that, that's the moment to activate. It's been a year since you ordered, come back, here's what's new.

And one important thing on deliverability. Truly inactive people, the ones who haven't opened anything in months, you don't target them with your regular campaigns. Otherwise you tank your open rates and Gmail starts sending you to spam. You isolate them in a dedicated reactivation sequence, and if it doesn't take, you cut them.

Don't segment on RFM alone

RFM gives you the base structure, but you can and should cross it with other logics depending on your business.

If you have retail on top of digital, separate your store customers from your online customers. Omnichannel shakes up quite a few flows. A customer who buys in-store doesn't have the same expectations or journey as your 100% web buyer.

On international, my view is clear-cut. When your base is small, like 50 or 100 people in a market, there's no point wasting time creating segments by language. You address everyone in French, or in your main language, and that's it. Once the base is large enough in a market, that's when you start building a dedicated segment and translating. Before that, it's wasted time.

One last point on data. Watch out for statistical anomalies in your exports. Sometimes you get some wild orders, a big B2B order, an outlier purchase, that skew your average basket or LTV. Clean that up before basing your segments on it, otherwise you're making decisions on garbage numbers.

Where to start this week

You don't need a six-month project to get going. Here's where I'd start.

Export your base and look at the split between leads, one-time buyers and repeat buyers. That simple ratio already shows you where you're losing people.

Then split your welcome into three levels so you stop talking to a loyal customer like a stranger. Also calculate your average time between two orders to plug your win-back into it, and set your inactives aside so you don't wreck your deliverability.

And above all, resist the urge to over-segment everything. Better three clean blocks that actually run than twenty segments nobody looks after.