How to tell whether your loyalty programme is working

Cards issued. It only ever goes up, which is exactly why it looks good in a report.

What it measures is how hard your staff pushed signups, not whether the programme works. Issue a thousand cards, have eighty of them in use, and you have nine hundred and twenty dead cards - while the report still says "a thousand customers".

It is not merely useless, it is actively harmful: issuing cards is the easy part, so a team measured on it optimises for the easy part, while all the value sits in cards being *used*.

Four figures describe the state of a programme almost completely.

Active cards - cards used at least once in the last 30 days. This is the programme's basic health signal.

Repeat-visit frequency - how often the average customer comes back. Increasing this number is the entire point of the programme; nothing else is.

Share of transactions carrying a card - what percentage of all sales involve a card being presented. When this is low the problem is usually not the programme but the counter: staff are not asking.

Basket difference - members compared with everyone else. This is what shows the programme's commercial value.

The first tells you whether the programme is alive, the second whether it works, the third whether it is being applied, and the fourth whether it pays.

This is the most commonly skipped step and the only one that cannot be fixed afterwards.

To know what the programme changed you need to know what things were like before it. In the four weeks before launch, measure: daily transaction count, average basket, and if you can, the share of repeat customers. Write those numbers down somewhere permanent.

Once the programme is running you cannot go back and take this measurement. Six months later you will not be able to answer "did it improve?", because the before figure was never recorded. It is the cheapest part of measurement and the part most often missed.

Allow for seasonality too: December always beats November and that has nothing to do with loyalty. Compare with the same month last year where you can.

In Walvero, two of the four numbers are on the Dashboard, in the Usage row:

• Active cards (30 days) - cards used at least once in the last 30 days, shown next to the number of issued cards and compared with the 30 days before
• Average visits and Returning customers - how many visits the average card made in the selected period and what share of cards came back at least twice. A visit is one card's transactions on one day.

Use the period tabs at the top (7 days to 3 years) to compare each figure with the previous period of the same length.

The other two numbers need your till. Share of transactions carrying a card: take the total number of sales from your POS report and compare it with Total Transactions on the Walvero dashboard for the same period. Basket difference: compare the average check of loyalty sales with the average check of all sales in your POS reports.

When the figures stop moving, the cause is almost always one of three things, and they should be checked in this order.

First: the counter. If the share of transactions carrying a card is low, the programme is not failing - it is not being offered. This is the cheapest fix there is: a visible QR code at the till and one sentence for staff to say.

Second: the threshold. If nobody is reaching the reward, the threshold is too high. The signature of this is healthy active-card numbers with a reward count near zero.

Third: the reward itself. If customers are reaching the reward but repeat frequency has not moved, the reward is not compelling enough to change behaviour.

Check those three in that order. Most of the time it is the first, and the first is also the cheapest to fix.

The most common mistake is evaluating a programme after two weeks.

If earning a reward takes eight visits and your customer comes twice a week, one full cycle is four weeks. To see whether behaviour has actually changed you need at least two or three complete cycles - close to three months.

Anything before that is noise. You will see active cards climbing in the first weeks, but that tells you the programme is new, not that it works. Do not make the call before three months - and agree that with yourself before launch, because patience is much harder to find afterwards.

Once a month, on the same date, write down four numbers: active cards, repeat-visit frequency, share of transactions carrying a card, basket difference.

Keep them in one sheet. After six months that sheet will tell you more than any report, because a trend is always more informative than a single figure.

Finish with one question: what is the one thing I will change this month? Visibility at the counter, the threshold, or the reward. Pick one, change it, measure next month. Change all three at once and you will never know which one worked.