What customer churn costs a small business

Retention arguments usually arrive as a statistic about how much cheaper it is to keep a customer than find a new one. That is true and almost useless, because it does not tell a café owner what to do on Monday. Here is the arithmetic instead.

The worked example

Take a café serving 500 customers a week with an average spend of €6. Of those, 40 per cent currently come back regularly. Annual revenue from repeat customers is 500 × 0.40 × €6 × 52, or €62,400.

Now lift the repeat rate by five percentage points, from 40 to 45 per cent: a realistic outcome for a loyalty programme, and at the conservative end of the 5 to 15% range usually observed.

Against a €24.99 monthly subscription, €300 a year, the programme returns roughly 26 times its cost. Even if you halve the assumed uplift to 2.5 points, it returns about 13 times.

What this calculation leaves out

Two things, and they cut in opposite directions.

Against the programme: reward cost. If a customer earns a free coffee every ten visits, you are giving away roughly ten per cent of the marginal revenue from those customers. On a €6 drink with typical café margins, that is real but small relative to the gain, and it only triggers on customers who returned.

In favour of the programme: this model counts only increased frequency from existing repeat customers. It excludes converting occasional customers into regulars, which is where a well-run programme usually does most of its work, and it excludes the lapsed-customer messaging that costs nothing per send.

Why churn is invisible without a system

The reason this revenue is usually left on the table is not that owners do not care about retention. It is that churn in a local business is silent. A customer who came twice a week for a year and then stopped does not tell anyone, and in a shop serving hundreds of people a day, nobody notices one absence.

The practical value of loyalty software is less the stamp mechanic than the list: knowing that 43 customers who used to visit weekly have not been in for 45 days, and being able to send those 43 people a message that lands on their lock screen.

How do I measure churn in my own business?

You cannot measure churn without a customer list, which is why most local businesses have never calculated it. Once you have one, the measurement is arithmetic: count the customers who were active in the previous period and are not active in this one, and divide by the first number.

Run it monthly on a rolling basis rather than annually. A café that loses four per cent of its regulars every month is losing nearly half of them a year, and monthly is early enough to notice a change and do something about it.

The subtlety is defining "active", because a customer has not left: they have simply not come in yet. The threshold has to reflect your own visit rhythm.

Set the threshold at roughly three times the normal gap. Any tighter and you will be messaging people who were simply away for a week, which annoys them; any looser and you are reaching out long after the habit has transferred somewhere else.

What does a win-back message recover?

A single well-timed message to lapsed customers typically brings back a modest but worthwhile fraction of them, and because the cost of sending it is effectively zero on a wallet-based system, almost any recovery rate is profitable. This is the clearest return in the whole category.

Take the café from the worked example. If 43 regulars have not visited in 45 days, each previously worth roughly €6 twice a week, recovering even a handful of them returns more in a month than the subscription costs in a year. The campaign takes a few minutes to send.

Two things make the difference between a message that works and one that is ignored. Send it once, not repeatedly: a second and third reminder convert almost nobody and get the card deleted. And give a reason to return now rather than a general invitation: a reward already waiting on their card outperforms "we miss you" by a wide margin.

Why is retention cheaper than acquisition?

Acquiring a new customer for a local business means paying to reach strangers, most of whom will never visit, while reaching an existing customer costs nothing and targets someone who has already chosen you once. The asymmetry is structural rather than a marketing slogan.

A boosted social post reaches people who mostly do not live near you, are not in the market today, and have no relationship with the business. A wallet notification reaches only people who have walked through your door, given you their number and installed your card. The conversion rates are not comparable.

This does not mean acquisition is wasted, a business with no new customers shrinks regardless of how well it retains. It means that for most independent businesses the retention side is the one that is unmanaged, and therefore where the cheap gains are.

Which report should I run?

One report, once a month: the list of customers who were regular and have now passed your lapsed threshold. Nothing else in a loyalty dashboard is worth as much attention.

Enrolment counts, total stamps issued and redemption rates are all interesting and none of them prompt an action. The lapsed list does: it is a specific set of named people who used to give you money and have stopped, and you can reach all of them in about two minutes.

If you look at one number alongside it, make it the share of enrolled customers who reached at least half the reward threshold. That tells you whether the programme is changing behaviour or merely recording it.

Run the numbers for your own business

The calculator on the Loonine homepage runs this same model with your figures. It deliberately defaults to the conservative end of the observed range, and it shows the assumptions rather than hiding them, because a retention model that flatters itself is not worth running.