Loyalty program metrics: the four numbers worth watching
Member count is the number everyone quotes and the one that means least. Four others tell you whether the program is earning its keep, or quietly leaking.
Ask a shop owner how the loyalty card is going and you will usually get a member count. It is the easiest number to find and close to the least informative one. A thousand members who each joined once and never returned is not a program; it is a list.
Four numbers tell you what is actually happening, and each one fails in a distinctive way that points at its own fix.
Key takeaways
- Enrollment rate: joins divided by transactions. Gates everything else.
- Redemption rate: the share of earned rewards actually claimed. Reads as a thermometer for whether the card is reachable.
- Repeat-visit rate: whether members come more often than they used to. The only number that proves the program changed anything.
- Reward cost against member revenue: the cost side, and almost always far smaller than owners fear.
- Give it two full card cycles before any of the four means much.
1. Enrollment rate
Joins ÷ transactions, over the same period.
This is the gate. Every other number is computed on the members you have, so a program that is not recruiting cannot be rescued by better campaigns.
A passive poster with no staff involvement typically converts a low single-digit percentage of transactions. Staff asking (one sentence, every time, at payment) moves it several times higher, and it is the cheapest intervention available to you. If your enrollment rate is disappointing, the fix is almost never the card and almost always the code by the till and the sentence next to it.
When it is low: the code is in the wrong place, the sign describes a program rather than a reward, or nobody is mentioning it.
2. Redemption rate
Rewards claimed ÷ rewards earned.
The most diagnostic of the four, because it fails informatively in both directions.
- Very low (people earn rewards and do not claim them) means the reward is not worth crossing the room for, or the card is so long that people who finished have forgotten why they were collecting.
- Very high, with nothing else moving, means you are rewarding people who were coming anyway. The program is working as a discount rather than as an incentive.
A healthy figure sits between those, and the direction of travel matters more than the absolute number. What you want to see is redemption rate holding steady while visit frequency rises.
An unredeemed reward is not a saving. It is a customer who stopped believing the card was going to pay out.
When it is low: shorten the card, or make the reward something people actually want. Both are covered in stamps, points or vouchers.
3. Repeat-visit rate
How often members visit, compared with how often they visited before they joined: or, failing that, compared with your non-member average.
This is the only number that demonstrates the program caused something. The others describe the machinery; this one asks whether the machinery did any work.
The comparison people reach for (members visit more than non-members) is real but overstates the effect badly, because the customers who join a loyalty card are the ones who were already coming most. That is selection, not causation. Two better readings:
- Before and after, same people. Take members with enough history and compare their visit interval in the eight weeks before joining with the eight weeks after.
- Interval drift over time. Watch whether your members’ average gap between visits is shortening, holding, or quietly lengthening.
When it is flat: the card is recording loyalty rather than creating it. Usually the reward threshold is too far away to change a decision, or you are not using campaigns to move people at all.
4. Reward cost against member revenue
What the rewards cost you ÷ what members spent.
The number owners most fear and most overestimate, because they think about the reward at its menu price rather than what it costs to make.
On a ten-stamp card, one coffee in eleven is free. That sounds like giving away roughly 9% of revenue. But you give away a coffee’s cost, not its price, and on a £3.20 coffee costing perhaps 64p to make, the actual figure is nearer 2%.
Two things follow. The reward is cheaper than it feels, so being slightly generous costs less than being slightly mean costs you in enthusiasm. And the curve flattens: past about ten stamps, lengthening the card saves you very little while making it much harder to finish.
When it is high: the reward is priced against retail rather than margin, or you are stacking a standing discount on top of the card.
What not to watch
- Total members. Accumulates whatever happens and never goes down.
- Stamps issued. A measure of how busy you were.
- Campaign open rates alone. A message everyone opens and nobody acts on is a well-written failure. This matters most when you send on more than one channel: a text will out-open an email every time, which tells you about the medium and nothing about whether the offer was any good. Judge each channel on visits it produced, not on opens.
- Anything measured in week two. A six-to-eight-week card needs two cycles before the numbers are stable: for a café on a twenty-stamp card, about three months.
Reading them together
The four interact, and the combination localises the problem better than any one of them alone.
| What you see | What it usually means |
|---|---|
| Low enrollment, everything else fine | A counter problem, not a program problem |
| Good enrollment, low redemption | Card too long, or reward too small |
| Good redemption, flat repeat visits | Rewarding people who were coming anyway |
| Everything healthy, reward cost climbing | Working as intended: check margin, not mechanics |
| Members rising, visits flat | You built a list, not a program |
The bottom row is the common one, and the fix is usually not the card at all. It is that nobody is being messaged, or everybody is being messaged the same thing. Sorting that out is loyalty segmentation.
Frequently asked questions
How often should I look at these? Monthly. Weekly is noise at small-shop volumes, and it tempts you into changing the card before a cycle has finished.
What is a good redemption rate? There is no universal figure: it depends on card length and reward. Judge yours against its own trend, and be suspicious of both extremes.
Can I measure this without a POS integration? Yes. Everything above comes from card scans and join records, which is why a loyalty card can be run without touching the till.
How long before I judge the program? Two card cycles. Changing the card length or the reward before then means you never find out what the first version did.
What if members visit more but I make less? Check the reward against margin rather than price, and check you are not running a discount alongside the card. Two overlapping schemes is the usual culprit.
Where to go next
Metrics tell you which decision to revisit. If it is the card mechanic, that is stamps, points or vouchers. If it is who hears from you, that is loyalty segmentation. All four decisions, in order, are in the complete guide.
Passumo reports visits, redemptions and redemption rate per campaign and per card, so the four numbers above come out of the dashboard rather than a spreadsheet.