A loyalty card is working when four things are true: customers keep joining at a steady trickle, members come back and get stamped, rewards are handed over at roughly the pace you designed the card for, and members make up a steady or growing share of your trade. Check those four numbers once a month, using your till and your members list, and within a quarter you will know whether the card has earned its place on the counter. Leave spend per visit alone for now. It moves for too many reasons to tell you anything about the card.
Set up the counting before you need it
Most of the work is done once, in ten minutes, and then runs itself.
Ring every reward through the till as its own line. A $0 item for a free product, or a discount key for money off, labelled "card reward". Your till's sales report will then count rewards by day, by hour and by staff member without anyone keeping a tally.
Pick one sample hour a week, the same hour each time, and have whoever is on the till make a mark for each sale: member or not. A sheet of paper by the register is enough. One busy hour and one quiet hour gives a fairer picture than either alone.
Keep a one-page log. Each month, write down the number of members, how many joined that month, rewards rung through the till, and the member share from the sample hours. Four numbers on one line. After three months the trend is plain.
Number one: joins
Count new members each month and set them against your number of sales. In the first weeks, joins come in a rush as regulars sign up. After a couple of months they settle into a trickle of newer customers. What matters is the trend after that. A steady trickle means staff are still asking and new faces are still finding the card. Joins falling close to zero usually means one of two things: the asking has stopped, or the QR has been moved, covered or knocked behind the tip jar.
Number two: members who come back
If every new member is stamped on the day they join and your rule is one stamp per visit, any member on two or more stamps has been back at least once. That gives you a repeat rate without any special report.
Take members who joined at least five weeks ago and look at forty of them. Count how many have two or more stamps. Most of them should. If fewer than half do, either members are not coming back or they are coming back and not being stamped, and the next number will usually tell you which.
The Members page shows each member's stamps and the day they joined, newest first, but not the date of their last stamp, so this sample is the quickest way to read it; the Members page help describes each column.
Number three: rewards handed over
Work out roughly how many rewards your card should produce in a month. Estimate member visits from your sample-hour share: if members are 35 percent of sales and you make 6,000 sales a month, that is about 2,100 member visits. Divide the stamps those visits should earn by your target, and you have the rewards you should expect.
Then compare it with the till's count of "card reward" lines. A count well below the estimate means stamps are being missed or rewards are not being offered. A count well above it means the stamping rule is being applied loosely, or someone is pressing for friends. A count close to the estimate is a card doing what you designed it to do.
Count rewards at the till, not from the list. A code leaves the Rewards waiting column when someone presses Mark used, as the reward codes help explains, so the column is only as good as the habit of pressing it.
Number four: member share of trade
The sample-hour tally, taken the same hour each month, shows how much of your trade the card reaches. A rising share means more of your regulars have joined. A flat share after six months is fine; it means the card has reached most of the people it is going to. A falling share is worth a look, because it can mean regulars are drifting off or new customers are not being asked.
A worked example: Mulberry Fold Grocer
Mulberry Fold Grocer is a fictional greengrocer making about 1,400 sales a week, or 6,000 a month. Its card is ten stamps, one per visit over $10, with $10 off as the reward. Four months in, the owner sits down with the log.
Joins: 41 new members that month, taking the club to 380. That is about 0.7 joins per 100 sales, down from the opening rush and steady for the last two months. Healthy.
Return: of 40 members who joined in the first two months, 33 have two or more stamps, which is 83 percent. Healthy.
Member share: the Saturday 9 to 10 tally shows 52 member sales out of 130, or 40 percent. The Wednesday 3 to 4 tally shows 18 out of 70, about 26 percent. Across the week the owner settles on about 35 percent, or roughly 2,100 member sales a month. About nine in ten of those are over $10, so members should be earning about 1,900 stamps a month, and with a ten-stamp card that means about 190 rewards.
Rewards: the till shows 64.
That gap is the finding. The owner spends a Saturday morning watching the counter. Produce is weighed at the scales by one person and paid for at the till by another, and in the rush the person on the till is not asking "are you in the club?" because they assume the scales already did. Nobody is stamping. The fix is a single rule: the till asks, every time, and the QR moves from the scales to beside the EFTPOS terminal.
The next month the till counts 151 rewards. At $10 off after ten visits averaging about $28, each reward is roughly 3.6 percent of what that member spent to earn it, which is a cost the owner had already decided was worth paying. The card had been working all along on paper; now it works at the counter as well.
What not to measure yet
Spend per visit. It moves with the weather, the season, prices and what is in stock, and in the first months the card's effect is lost in the noise.
The result of a single message. One campaign that brings in eleven people, or none, says more about that week than about the card.
Week-to-week swings. Monthly is often enough. Weekly numbers will have you changing things that were never broken.
Mistakes to avoid
Changing two things at once. If you change the target and the reward in the same month, you will never know which one moved the numbers.
Measuring only the busy hour. Saturday morning flatters the member share and hides a weekday problem.
Treating every lapsed face as failure. Some regulars move away every year. A trickle of lapsed members is ordinary; a sudden cluster is the thing to investigate.
What to do this week
- Add a "card reward" line to your till, as a $0 item or a discount key.
- Choose one busy and one quiet sample hour, and put a tally sheet by the register.
- Count your members and start the one-page log.
- Work out how many rewards a month your card should produce at your current member share.
- Put a reminder in the diary for the same day next month, and compare.