the only loyalty number worth watching

Loyalty software loves a dashboard. Stamps issued, cards started, rewards redeemed, scans this week, scans last week — a wall of numbers that all move roughly together and none of which tell you whether to do anything differently tomorrow.

There is one number underneath all of that which is worth more than the rest combined: what share of your customers came back.

why this one and not the others

Every other loyalty figure grows when you are busy. Stamps issued goes up on a good Saturday whether or not a single new person decided you were their café. It measures traffic, and you already know when you were busy.

The share of customers who came back is different, because it is a ratio rather than a count. It does not move because the weather was good. It moves when something about the experience made people choose you again — which is the only thing a loyalty program is actually for.

Counts tell you how busy you were. A ratio tells you whether you are building something.

how to define it so it stays honest

The definition matters more than the measurement, because a loose definition drifts in your favour over time. Ours is deliberately narrow: a returning customer is somebody who collected a stamp on two or more different days.

Two things follow from that, and both are worth understanding before you quote the number to anyone.

• Different days, not different transactions. Somebody who buys three coffees in one afternoon has visited once, and a definition that counted three would flatter you.
• It counts people who collect stamps, not everybody who walks in. A customer who came back twice but never started a card is invisible to it.

estimate it by hand, this month

You do not need software to start. You need one number written down consistently for four weeks, which is more than most cafés have.

1. 01Pick one hour you are always open — say 8am to 9am on weekdays — and keep to it. A sample at a fixed time beats a guess across the whole day.
2. 02Each day in that hour, count the customers whose face or order you recognise from the same hour in the previous fortnight.
3. 03Write two numbers in a notebook: total customers in the hour, and recognised ones.
4. 04At the end of four weeks, divide the second by the first. That is a rough returning rate, and its direction over the next four weeks is what you actually care about.

It is crude, it undercounts, and your staff will disagree about who counts as recognised. It is still worth doing, because a crude number that moves is more useful than an exact number you never collect.

what to do once you have it

Change one thing at a time and watch the ratio, not the counts. If your returning share rises while total customers stay flat, the change worked and it is worth keeping — you are converting more of the same traffic into regulars. If both rise together, you got busier and learned nothing about loyalty.

That is the whole discipline. It is also, exactly, the number our dashboard puts at the top of the page in the largest type on the screen — because if we showed you six numbers of equal size, we would be doing the thing this post is complaining about.