I’m about to walk you through a quick, pragmatic RFM test I use with clients when they’re considering a tiered loyalty relaunch. The goal is simple: spot whether moving customers into a new higher tier will primarily reward customers who would have bought anyway — that’s cannibalisation — rather than genuinely increasing incremental spend or engagement. You don’t need a fancy data warehouse or a PhD in statistics for this. A clean RFM table, a few filters and the right comparisons will surface the red flags fast.
Why RFM for spotting tier cannibalisation?
RFM (Recency, Frequency, Monetary) gives you a compact, behaviour-first view of your customer base. When you’re introducing or reworking tiers — especially when the differences are value-based (discounts, points multipliers, free shipping) — the risk is that the new benefits just accelerate purchases from your already-best customers without growing overall revenue or retention. RFM helps identify who those “already-best” customers are, how they behave, and how much of the targeted uplift would likely be incremental.
What I want you to prepare
Before we start, pull a dataset with these minimum fields for the past 12 months (you can use 6 months for faster checks, but 12 gives a better view):
If you have more — channel, product category, acquisition source, promo flags — include them. But the test will work on the basics.
The test in four steps
We’ll create an RFM score, define target segments that would be eligible for the new tier, and simulate what happens if we upgrade them. I keep the procedure intentionally simple so you can do it in Excel, Google Sheets, or any BI tool.
Calculate for each customer:
Then assign quintile scores (1 lowest — 5 highest) for each metric. Reverse Recency so a recent buyer gets a high score. Combine into a composite RFM score (e.g., R_score + F_score + M_score). I often use RFM sums from 3–15.
Define who would be moved into the higher tier under your relaunch rules. Typical criteria:
Create two sets: those already in the current top tier (if you have one) and those who would newly qualify under the relaunch.
Here’s where the cannibalisation signal lives. Look at average order frequency, average order value (AOV), repeat rate, and share of wallet. Key comparisons:
If newly eligible customers are statistically indistinguishable from current top-tier members — particularly in frequency and spend — you’re likely looking at cannibalisation risk: you’ll be taking value from the brand to provide rewards to customers who were already high-value.
Run a simple uplift simulation. Assume the tier gets a benefit that increases spend or frequency by a conservative 5–10% (you can base this on past promotions or modest A/B test results). Multiply that uplift against the newly eligible group and compare it to the cost of the benefit for that group.
| Metric | Current Top Tier | Newly Eligible |
| Customers | 1,200 | 3,400 |
| Avg annual spend | £420 | £390 |
| Projected uplift (5%) | £21 | £19.50 |
| Cost of benefit per customer | £12 | £12 |
| Incremental margin (est.) | £9 | £7.50 |
If the incremental margin for the newly eligible is small or negative after benefit costs, the relaunch is likely redistributing existing value rather than creating new revenue.
Red flags to watch for
When I audit programmes, certain patterns repeatedly predict cannibalisation:
Any of these suggest that simply moving customers up a tier will reward activity that was going to happen anyway.
How to reduce cannibalisation before you launch
If your test shows risk, try one or more of these mitigations I use with clients:
A quick real-world example
I worked with a mid-size apparel brand that planned to reduce the spend threshold and automatically upgrade members. My RFM test showed the newly eligible group had similar frequency and AOV to existing VIPs and a high percentage of promo-driven purchases. We recommended a staged rollout with a behavioural gate: customers qualified only after a second purchase post-qualification. That change reduced immediate upgrade numbers by 40% but preserved most of the incremental value — and the brand avoided a costly permanent uplift in benefit liabilities.
If you want, I can provide a spreadsheet template with the steps above pre-built so you can drop your export in and run the analysis in an afternoon. It’s the exact template I use when I audit loyalty programmes for SMEs — lightweight, transparent and actionable.
