New: Smart Segmentation — vertical-aware RFM, in Connected Pro. Take a look →

Product

Why we built customer segmentation that actually fits your restaurant

Two segmentation models — coffee chain vs. fine dining — side by side

A customer at a Madrid coffee chain comes in every weekday for a €4 cortado. A customer at a Girona fine-dining restaurant comes once a quarter for a €120 tasting menu. Both are Champions for their brand. Most loyalty platforms call only one of them that.

That’s the gap Smart Segmentation closes — and why we built it into Connected Pro.

The problem with one-size-fits-all RFM

RFM — Recency, Frequency, Monetary — is the workhorse of loyalty segmentation. Three scores per customer, combined into a code, mapped to a segment like Champion, Loyal, At Risk or Lost. It’s been the standard since direct mail invented it in the 1990s.

The catch: most loyalty platforms hardcode the model. The same monetary metric, the same lookback window, the same segment thresholds — applied uniformly across every brand on the platform. A coffee chain and a fine-dining restaurant get the same Champion definition. One of them is wrong.

We’ve seen what that costs in our own book. A coffee chain operator gets handed a Champion list dominated by people who bought a €40 birthday cake last quarter, not the daily-cortado crowd that actually drives revenue. A fine-dining operator gets a Champion list dominated by frequent low-ticket lunches, not the quarterly tasting-menu loyalists who keep the restaurant viable.

The model isn’t wrong. The configuration is.

What Smart Segmentation does differently

We built four vertical templates: Coffee & QSR, Casual dining, Fine dining, and Retail & service. When a brand admin sets up Smart Segmentation in the Hub, they pick the template that fits their business. The template configures three things:

Monetary metric. Coffee gets total spend (rewards the daily-cortado crowd). Fine dining gets average order value (rewards the quarterly tasting-menu loyalists). Casual gets total. Retail picks.

Lookback window. Coffee gets 6 months (their customers visit weekly — 6 months is plenty of data). Fine dining gets 18 months (their customers visit quarterly — 18 months gives meaningful frequency). Casual gets 12. Retail picks.

Segment mapping. The same RFM code maps differently across templates. A 331 (high recency, high frequency, low monetary) is the textbook coffee Champion. The same 331 at fine dining is unusual — possibly a price-sensitive regular who only orders starters. The mapping table reflects that.

Brand admins don’t see this complexity. They pick “Coffee chain,” and the right model runs. If they want to fine-tune the monetary metric or lookback later, they can — but the baseline is right out of the gate.

Six segments, six clear actions

We thought hard about how many segments to expose. The academic literature suggests 11. Some platforms cap at 9. Both numbers are too many for an operator who runs a restaurant for a living and runs loyalty between lunch and dinner rushes.

We landed on six:

  • Champion — most loyal, most valuable, recent. Focus on keeping, not winning. No discounts.
  • Loyal — stable regulars. Goal: lift frequency. Birthdays, anniversaries, new launches.
  • Promising — on the right track. A few nudges away from Loyal. Easy small rewards.
  • At Risk — used to be a regular, last visit drifted. The single highest-ROI segment to act on.
  • Lost — long since visited. One-shot win-back; if no response, move on.
  • New — first or second visit. Goal: lock in the second visit.

Six fits in a brand admin’s head. Each one comes with an obvious next campaign. No segment is decorative.

Why this lives in Connected Pro

Smart Segmentation runs on basket-level data — every line item of every receipt for every identified member. That’s the gold standard for RFM accuracy, and it’s exactly the data Connected and Connected Pro collect through POS integration.

We made the call to ship it in Connected Pro rather than across all plans. Pro is the tier built for multi-location chains running at scale — the brands with enough customer base and enough campaign volume to make segmentation pay back quickly. Smaller operators don’t typically have the member count to fill all six segments meaningfully.

If you’re on Connected (base) and the segmentation story resonates, moving up to Pro is the path. If you’re on Lite — for salons, boutique cafés, single locations — the simpler “haven’t visited in 6 weeks” automations Lite already includes do the job for now.

How brands actually use it

The fastest payback we’ve modelled is the At Risk segment. A customer who was a Champion six months ago and hasn’t visited in eight weeks is the textbook winnable churner. Catching them before they cross from At Risk to Lost is the highest-ROI campaign a loyalty program can run — and the only way to catch them reliably is to score the whole base every night.

In Smart Campaigns, the segment label (rfm:at_risk) works as a direct filter. You don’t write SQL, you don’t learn a new tool — you drop the tag into the audience builder, write a “we miss you” message with a small reward, and the campaign fires every morning to whoever crossed the line overnight.

The Hub also surfaces a Segment Distribution report — how many Champions you have today, how it’s moving week to week, what share of revenue each segment generates. The strategic question — “is my Champion count growing or shrinking?” — stops being a hunch.

A note on the “Champion churn” problem

A 2025 Grant Thornton study found that 70% of the top 20% of customers in a restaurant loyalty program were different people from one year to the next. That’s a problem if you treat Champions as a stable identity. It’s not a problem if you act on Champions through automated campaigns that fire whenever someone crosses into the segment — which is exactly the point of nightly re-scoring.

Smart Segmentation isn’t a list of names you print out and hand to a marketing manager. It’s a live tag that updates every night and triggers the right action automatically. That’s how it survives the Champion-churn problem the academic literature warns about.

What’s next

Smart Segmentation is included in Connected Pro. Existing Pro customers will see a Hub banner prompting them to pick a template; new Pro customers get the template choice during onboarding.

In the next release we’re adding segment migration reporting (where did your At Risk customers come from? how many of last quarter’s Promising moved to Loyal?) and starter campaign templates for the most-acted-on segments. If you want to be first in line, request a 20-minute demo and we’ll walk you through your vertical’s template on a live dataset.