In brief
A paid membership works when the customer can recover the fee inside a month they can picture, and visits often enough that recovery is not hypothetical. Below roughly weekly, the arithmetic fails for both sides.
The short answer
Charge for membership when the customer can recover the fee inside a period they can picture, and when they already visit often enough that recovery is obvious rather than theoretical. A £5 monthly membership giving one free filter coffee a week hands the customer about £14 of retail value and costs you £3.44 in beans, milk and cups.
Both sides win because the numbers are not close. If the fee needs almost every possible visit to break even, the customer will do that arithmetic too, and decline.
What a loyalty program member actually is
The word does two jobs, and the difference is the whole subject of this post.
A member of a free program is an entry in a ledger. They gave you a way to be identified, they collect stamps or points, and they owe you nothing. Enrolment costs them a scan. That is the sense used on the what is a loyalty program page and in most software.
A paid member is a subscriber. Money leaves their account on a date whether they visit or not, and in exchange something is permanently true about how you treat them. That is a contract, not a ledger entry, and it changes who is carrying the risk.
In a free program you carry the risk. You promise a reward, they earn it over nine visits, and you pay on the tenth. In a paid program the customer pays first and hopes to earn it back. The direction of the promise is reversed.
The arithmetic, with the assumptions on the table
Assumptions, all yours to change:
- Membership costs £5 a month
- It includes one filter coffee a week
- That coffee sells for £3.20 and costs £0.80 in beans, milk and cup
- An average month is 4.3 weeks
| Line | Per member, per month |
|---|---|
| What the customer pays | £5.00 |
| Retail value they collect | £13.76 (4.3 × £3.20) |
| What it costs you | £3.44 (4.3 × £0.80) |
| What you keep | £1.56 |
The customer sees nearly £14 of coffee for £5. You see £5 of cash for £3.44 of stock. Neither party is being clever at the other’s expense, which is why this shape survives contact with reality.
Sixty members at £1.56 is £93.60 a month before anyone buys anything else, and the anything else is where the real money is. A member collecting a free filter coffee is standing at a counter with pastries on it. They pay full price for the pastry.
The recovery test
One division decides whether a membership is worth offering.
Fee ÷ retail value of the included item per visit = visits needed to break even.
At £5 and £3.20 that is 1.6 visits. A customer who comes weekly makes 4.3 in a month, so they clear the fee in the second week and spend the rest of the month ahead. They can feel that without doing any sums.
Price it at £15 and the same division gives 4.7 visits. That is more than a weekly customer makes. The membership now requires them to change their behaviour before it pays anything, which is exactly the failure mode of a threshold set too far away.
Set the fee so a customer at their existing visit rate clears it in the first half of the month. Anything tighter and you are selling a bet.
Why paid members visit more
They have already spent the money, and money already spent is uncomfortable to waste.
A free stamp card asks the customer to invest visits before they get anything, so walking away costs them nothing they can feel. A membership inverts that. On the Thursday they might have gone somewhere else, the £5 is sitting in the account of the place they did not go to, and that is enough to move a decision that was close.
This is the only real advantage a paid program has over a free one, and it is worth being precise about it. You are not buying loyalty. You are buying a small, recurring reason for the customer to resolve a coin-flip in your favour.
The churn problem
Monthly billing means a monthly decision, and that is the cost of charging.
A free program lapses silently. Someone stops coming, the stamps sit there, and nobody has to conclude anything. A membership lapses out loud. The customer opens their banking app, sees your name on a month they barely visited, and cancels. The quiet drift that costs you nothing in a free program becomes an explicit judgement about you.
Three things follow.
Every month needs a reason to still be a member. Not a new perk each time, but something that happened: a member-only item in the second week, first access to a batch, an early booking window. A membership where nothing ever happens is a direct debit waiting to be noticed.
A member who has not visited in three weeks is about to cancel. That is not a prediction, it is the same arithmetic they will do. A short message before the billing date, offering to pause, keeps a customer who would otherwise leave permanently. Cancelling ends the relationship. Pausing is a diary entry.
The first month is not evidence. People sign up for anything once. Look at whether the second and third months hold, which means you need to be able to see visits per member over time. If you cannot read that from your own dashboard, do not launch a subscription yet.
What to include and what to hold back
Include things with a low cost of goods, high frequency and predictable supply. The filter coffee qualifies on all three. A £14 retail value costs you £3.44, restocks itself, and never runs out in a way that turns a member into an argument at the counter.
Hold back four kinds of thing.
Anything unlimited with a real cost of goods. Unlimited coffee is priced by whoever drinks the most, and the person who drinks the most will find you.
Guest privileges. A member bringing three friends who all collect the included item turns a £5 subscription into a £20 giveaway, and refusing on the day is a scene.
Anything requiring staff judgement. If honouring the membership needs someone to decide whether this counts, it will be honoured differently on different shifts, and members compare notes.
Your best thing. Members should get more of what you sell, not the only good version of it. A shop where the walk-in customer gets the second-tier experience has trained everyone who is not a member to go elsewhere.
The reward-choosing logic is the same as for a free program, and the longer version is in how to reward customer loyalty.
Where paid membership does not work
Two variables decide it: how much a visit is worth and how often visits happen. Frequency matters more.
| Typical ticket | Visit frequency | Paid membership |
|---|---|---|
| £3 to £6 | Three or more a week | Works well. The fee clears in days |
| £3 to £6 | About weekly | Works, if the fee stays under two visits’ value |
| £8 to £15 | Weekly lunch | Works, with a fee small against a single visit |
| £25 to £45 | Every five or six weeks | No. Sell a prepaid block instead |
| £40 plus | A few times a year | No. The customer forgets they are paying |
| £100 plus | Annual or seasonal | No. Sell a service agreement, which is a different product |
The barber row is the one owners argue with, so here is the reasoning. At a cut every six weeks, a member makes eight or nine visits a year. A fee small enough to be fair against that frequency is too small to be worth billing, and a fee large enough to matter needs the included benefit to be a significant part of a £30 service. Meanwhile the customer is paying in the five weeks they are not there, which they will notice.
The right product at that frequency is prepayment, not subscription: six cuts bought up front, the seventh free. The money arrives early, which is most of what a membership was for, and there is no monthly decision to lose. It is a stamp card with the cash collected at the start, and it needs no billing system at all. The counting side is the same as any visit-based program.
Before you price it
Two numbers, both from your own records rather than from anyone’s benchmark.
The first is the median number of visits a regular makes in a month. Not the average, which one enthusiast distorts, and not the number you would like it to be. The second is the cost of goods on the item you plan to include.
Multiply them, add the margin you want, and you have a fee. If you do not have the first number, run the free version for three months and count. A membership priced on a guess about frequency is a guess about whether you are giving away £3.44 or £14, and the cost model will not save you from getting that one wrong.