The one-screen loyalty dashboard every shop owner needs

Open most loyalty products and you get a wall: engagement scores, tier distributions, point liability, a leaderboard of your top customers by lifetime value. It looks impressive in a demo and none of it changes what you do on Tuesday.

A metric earns its place on the screen if, and only if, a plausible value of it would make you do something different. By that test, most loyalty dashboards should be about six numbers long.

The six that matter

1. Busiest days and hours

Every stamp, plotted by day of week and hour. This is the one owners look at longest, because it is the one that costs real money to get wrong. You rota against it, you schedule deliveries against it, and you decide when to run an offer against it: the correct time for a promotion is the trough right before the peak, not the peak itself.

2. Repeat rate

The share of customers who have visited more than once. This is the closest thing to a single score for whether the whole scheme works. If it is low, the problem is not your reward, it is that people are not coming back at all, and no automation will fix that.

3. New against returning, week by week

Two lines on the same chart. New tells you whether acquisition is working. Returning tells you whether retention is. A shop with rising new customers and flat returning has a leaky bucket, and pouring more in will not fix it.

4. How many are one stamp away

The most immediately actionable number on the screen. These people need one nudge, and they are the highest-converting audience you will ever have. This should be automated, and the number is there so you can see the automation working.

5. How many have lapsed

Customers past your win-back threshold. Watching this climb is the early warning that something changed: a new competitor, a price rise that landed badly, a member of staff people liked who left.

6. Stamps per staff member

The audit chart. Not there to catch anyone, there so that a wildly uneven split is visible without you standing behind the counter counting. Ninety-nine times out of a hundred the explanation is that one person works the Saturday rush.

What to leave off

• Point liability. Meaningful for an airline. Not for a shop giving away a cone.
• Engagement scores. A composite of things you cannot act on individually is a thing you cannot act on at all.
• Lifetime value. Interesting once a year, useless weekly, and wildly noisy on small numbers.
• Tier distributions. If you have tiers in a coffee shop, the tiers are the problem.
• Email open rates. They have been unreliable since privacy proxies started opening mail on people's behalf. Measure visits after a send instead.

One screen, and mean it

The discipline is not choosing good metrics; it is refusing to add the seventh. The moment a dashboard needs scrolling and tabs, owners stop opening it, and a dashboard nobody opens is worse than no dashboard because it creates the feeling that the numbers are handled.

If you cannot take it in during the thirty seconds before the shop opens, it is not a dashboard. It is a report.

That is the standard we held our own overview screen to: six numbers, three charts, no scrolling on a laptop.