In September 2022, Pret A Manger restructured Club Pret from a £25 all-you-can-drink subscription to a £30 tier offering five half-price barista drinks a day. Within weeks, the operator that had built the UK's most-copied hospitality subscription lost roughly a third of its net subscribers and watched around 40% of members opt out, according to Fable Data's transaction analysis. The lesson for independent hospitality groups is not that subscription works. It is that frequency, priced wrongly, is the fastest way to cannibalise the customers you already had.
- Club Pret's real achievement was frequency, not revenue: it turned weekly visitors into daily ones.
- The 2022 restructure exposed the cannibalisation risk: high-value customers moved to a discount tier and net spend dropped.
- Fable Data reported ~40% opt-outs and a one-third decline in net subscribers after the price change.
- Chargebee attributes the 44% cancellation rescue rate to billing infrastructure. The real driver is retention interventions, which is programme management, not software.
- For independent groups, the honest question is not 'should we launch a subscription' but 'who is going to run it on Tuesday'.
What Club Pret actually proved (and it isn't what most people quote)#
Most analyses of Club Pret focus on the top-line: Chargebee reports subscribers spent roughly 4x more than non-subscribers and the programme helped drive Pret back to profitability. Those numbers are real. They are also incomplete.
The genuine breakthrough was frequency. A £20 (later £25, later £30) monthly commitment reframed Pret from a sometimes-purchase into a daily habit. Once a customer has paid the subscription, the marginal cost of walking past a competitor is psychological, not financial. That is the mechanic worth studying.
"Subscription doesn't sell coffee. It sells the decision, once, and then removes it from the customer's day."
The cannibalisation problem no one models before launch#
Here is the scenario every independent operator should model before copying the Club Pret playbook. A customer visits your venue three times a week and spends around £8 per visit. That is roughly £96 a month at full price. You launch a £25 subscription that gives them one drink a day. They sign up. Their visit frequency does not change, because they were already visiting often. Your monthly revenue from that customer drops from £96 to £25 plus incidental food attach.
You have not acquired a new habit. You have discounted an existing one.
| Customer type | Pre-subscription monthly spend | Post-subscription monthly spend | Net effect |
|---|---|---|---|
| Occasional visitor (2x/month) | £16 | £25 + attach | Positive: revenue up, frequency up |
| Regular (2x/week) | £64 | £25 + attach | Neutral to negative depending on attach |
| Heavy user (5x/week) | £160+ | £25 + attach | Strongly negative: revenue destroyed |
| New acquisition via subscription | £0 | £25 + attach | Positive incremental |
Pret could absorb this because scale, food attach and volume economics work in its favour. An independent group with three sites and thinner margins cannot. This is why the Elite Business analysis flagged early that the model was structurally optimistic about food attach rates.
The 44% rescue rate isn't a billing story#
Chargebee's case study credits its billing infrastructure with a 44% cancellation rescue rate. Look closer. Rescuing a cancellation requires an intervention: a pause-instead-of-cancel prompt, a timed offer, a personalised message about the credits the customer has not used. Those are programme management decisions. The billing platform just executes them.
This is the operational gap independent groups walk into. They buy a loyalty or subscription platform expecting the tool to run retention. Nobody sends the win-back. Nobody reads the lapse data. Nobody notices when a subscriber's usage drops for three consecutive weeks. That's why most loyalty programmes stall after launch: the software works, the programme doesn't.
"The distribution layer is irrelevant if no one is sending the notifications, running the win-back, or reading the lapse data."
What independent groups should build instead#
The takeaway is not 'don't do subscription'. It is 'don't do subscription as your first move'. For most independent hospitality groups with two to fifteen sites, the stack that compounds looks different from Pret's.
| Layer | Pret's approach | What works for independent groups |
|---|---|---|
| Entry mechanic | Paid subscription at scale | Free stamp or points card, low friction to join |
| Frequency driver | Daily drink allowance | Reward tiers that reward the 4th, 8th, 12th visit |
| Premium tier | Only tier | Optional paid membership for regulars once habit exists |
| Distribution | App download | Wallet-native pass, no app to install |
| Retention | Billing-triggered saves | Scheduled push campaigns, lapse triggers, human review |
Start with a free stamp or points card that captures every visitor. Let the data tell you who is already visiting 3x a week. Then, and only then, offer those customers a paid membership tier that gives them something Pret's subscription doesn't: recognition, priority, a genuinely better experience, not just a discount on what they were already buying.
For a more detailed breakdown of when to use each mechanic, see choosing the right loyalty mechanic for your venue.
Stacking mechanics: subscription plus stamps, not either/or#
The independent operators who get the most from Pret's playbook are the ones who don't copy it wholesale. They stack: a free stamp card for everyone, a paid membership for the top 5% who self-identify through behaviour, and campaign pushes that move the middle group up one tier at a time.
The stacking only works if someone runs it. Push notifications don't schedule themselves. Lapse windows don't get flagged by the platform. This is how Carrott runs the programme for you, and it's the piece the software-vendor coverage of Club Pret consistently misses.
Pricing lessons from Pret's 2022 restructure#
The Club Pret price change is a case study in what happens when you re-price a habit. The customers who churned weren't marginal. They were, by definition, the ones getting the most value at £25. The lesson: if you launch a subscription, price it against the behaviour of your median regular, not your heaviest user. And never raise the price on the cohort that anchored to the original number without a compelling addition, not a subtraction.
Was Club Pret actually a success?
For Pret, yes: it contributed to a return to profitability and lifted subscriber spend to roughly 4x non-subscribers according to Chargebee. But the 2022 restructure showed the model is sensitive to pricing, and Fable Data recorded around 40% opt-outs and a one-third net subscriber decline after the change.
Should an independent cafe group launch a coffee subscription?
Only after modelling cannibalisation against your existing regulars. If a large share of your visitors already come 2-3 times a week, a discounted subscription will reduce their spend, not increase their frequency. Start with a free stamp or points card, learn who your regulars actually are, then layer a paid tier.
What's the real driver of subscription retention?
Retention interventions: pause-instead-of-cancel prompts, timed offers, personalised messaging around unused credits. Billing infrastructure executes them but does not decide them. That is programme management.
Do we need an app to run a subscription or membership programme?
No. Wallet-native passes in Apple Wallet and Google Wallet remove the download barrier entirely. Customers join through a web form and the card saves to their phone.
What's the biggest reason loyalty programmes fail after launch?
Nobody runs them day to day. The software is bought, the card is designed, the launch email goes out, and then no one schedules the pushes, reads the lapse data, or runs win-back campaigns. That's not a tooling problem, it's an operations problem.