The most common way loyalty programs get judged is the number of cards issued. “We have seven hundred!” But seven hundred cards don’t mean seven hundred extra customers – they mean seven hundred people who scanned a QR code. How many of them came back because of it is not in that number.

Here are four metrics that do say it, and how to get from them to a return.

What it is: what percentage of issued cards saw activity in the last 90 days.

This is the health check for the whole program. A card that hasn’t gained a stamp in three months is dead – the customer has it on their phone but isn’t coming in.

How to read it:

Under 30% – the program didn’t catch on. Either the reward is too far away, or nobody is scanning.
30 to 50% – normal for a business that doesn’t run campaigns.
Over 50% – the program works and deserves more investment.

The most common cause of a low number isn’t the customer, it’s the staff. Before you start changing rewards, check whether anyone is scanning at all: compare the number of stamps with the number of receipts for the same day. If it comes out one to five, the problem is at the counter, not in the settings.

2. Gap between visits

What it is: how many days pass on average between two visits by the same customer.

This is the one metric a loyalty program is supposed to move directly. Everything else is derived.

How to read it: compare cardholders against how they behaved before the program launched, not against guests without a card. People who took a card were more loyal to begin with – that is why they took it. Comparing them with a random guest gives you an effect the program never caused.

The realistic impact of a well-tuned program is shortening the gap by 5 to 15%. Anyone promising more is selling.

3. Reward completion rate

What it is: what percentage of customers who got a first stamp made it all the way to the reward.

This number exposes a badly set reward faster than anything else.

Under 10% – the reward is too far away. The customer loses motivation after the second stamp and the card goes to sleep. Shorten the path.
Over 60% – the reward is too close. You are handing it to people who would have come anyway.
20 to 40% – a sensible range for a stamp card.

It also helps to watch which stamp people stall on. If most drop off between the third and fourth, it doesn’t matter whether the reward is at ten or twenty – the problem is that there is no milestone halfway.

4. Revenue lift per cardholder

What it is: how much a cardholder spends per month compared with before they had the card.

This is the number that turns loyalty into money. Calculate it as average spend × visits per month, measured on the same group of people before and after.

Watch out for two distortions:

Season. The December bump wasn’t the loyalty program.
The cost of the reward. Subtract it from the lift. If a customer spends €8 more a month but gets a reward costing you €2.40 every month, the net gain is €5.60.

How to turn that into a return

The return is then simple:

(active cards × net monthly gain per card) − monthly plan

A worked example: 300 active cards, a net gain of €3.60 per card per month, a €40 plan.

300 × 3.60 = €1,080
1,080 − 40 = €1,040 per month

This calculation is deliberately rough – it is about the order of magnitude, not accuracy to the cent. If it comes out negative, the program is badly set up, not proof that loyalty doesn’t work. To play with your own numbers, there is the ROI calculator.

When to evaluate

Not before three months. A shorter window doesn’t have enough data: it takes most businesses six to eight weeks before cards spread among guests and become a habit.

A sensible rhythm:

Weekly: new cards and number of scans. Just a check that scanning is happening.
Monthly: share of active cards and reward completion rate. This is what you tune the settings by.
Quarterly: gap between visits and revenue lift. This is what you decide the program’s future by.

A basic overview of these numbers is in the statistics with no manual counting; deeper breakdowns come out of the customer database.

What not to trust

Three numbers that look good in a screenshot and mean nothing:

Cards issued. It measures how many people scanned a QR code. Nothing more.

Push notification open rate. For a lock-screen notification, an “open” is loosely defined and the numbers tend to be inflated. Watch how many people came in during the week after a campaign instead – that is measurable and decisive.

Customer satisfaction with the program. In a survey people will say they love your loyalty program. Then they don’t show up. Behaviour is the only feedback that counts.

If you are still building the program, start with the reward – three of the four metrics above rest on it. The how-to is in the guide to loyalty programs for cafés and restaurants, and costs and return models are covered in what a loyalty program costs.

1. Share of active cards