A customer walks into your suburban brunch site on a Sunday, orders eggs benedict and a flat white at £16, and asks to redeem the free coffee they earned across the road at your espresso bar where their average ticket is £3.20. The barista pauses. Nobody told her what to do. She gives them the coffee. Multiply that across 40 weekends and six sites, and you have a margin leak nobody is tracking, plus a staff team quietly making up the rules as they go.
Multi-site loyalty fails for one reason: rules written for a single cafe get copy-pasted across locations with very different economics, and then staff are left to interpret the gaps. The fix is not a longer rulebook. It is a small set of rules that hold true everywhere, a clearly defined area where managers can flex, and an earn rate that does not punish your best site for being your best site.
- One programme, one set of core rules. Location flex sits inside a tight frame, not outside it.
- Anchor earn rates to spend, not visits, the moment your sites have different price points.
- Decide redemption fairness in advance: either reward value is fixed in £, or it floats with the menu.
- Write rules that are staff-proof, not just customer-friendly. Ambiguity is where complaints start.
- Keep signup friction near zero so a tighter earn rate still converts. No app download is the cheapest lever you have.
Why a single set of rules breaks the moment you open site two#
A single-site cafe can run a buy-9-get-the-10th-free stamp card on the back of a napkin and the maths works. Every visit looks roughly the same. Average ticket is stable. The barista knows every regular by face. None of that is true at three sites.
City-centre espresso bars run on takeaway volume, sub-£4 tickets and 90-second transactions. Suburban brunch sites run on sit-down covers, £12-£18 tickets and a 45 minute dwell. Train station kiosks run on commuters who buy the same thing every weekday for two years and then change jobs. The same stamp-per-visit rule, applied to all three, hands the cheapest site the highest effective discount and rewards your most loyal commuter the least.
| Site type | Avg ticket | Visits/wk per regular | Stamp-per-visit cost as % of revenue |
|---|---|---|---|
| City espresso bar | £3.20 | 4-5 | ~10% |
| Suburban brunch | £14.50 | 1 | ~2.3% |
| Commuter kiosk | £4.80 | 5 | ~7% |
| Roastery flagship | £8.00 | 2 | ~4% |
Read that table the other way and the problem becomes obvious. Your espresso bar regular is getting four times the effective discount of your brunch regular, for spending less money. That is not a loyalty programme. That is a subsidy.
The core decision: one programme, or location-level flexibility?#
There are three workable models for a cafe group. Pick one before you write any rules.
| Model | How it works | Best for | Watch out for |
|---|---|---|---|
| Unified programme | One card, one earn rate, one redemption rule across all sites | Wholly-owned groups with similar formats | Margin leak if price points vary significantly |
| Unified with location modifiers | One card, shared earn rate, but redemption value capped or site-specific reward menu | Mixed-format groups (kiosk + sit-down) | Rules get complex; staff training matters more |
| Federated | Each site runs its own card under a shared brand | Franchise or licensed sites with independent P&Ls | Customer confusion; brand inconsistency |
Most groups under ten sites should run unified with location modifiers. The card is the same, the brand experience is the same, but the redemption rule does the heavy lifting on margin protection. Federated only makes sense when site owners genuinely have separate P&Ls and partial autonomy; if you go that route, you need a written agreement on cross-site redemption (and someone, usually head office, holds the float).
Setting earn rates that feel fair across different price points#
The single biggest lever is shifting from visit-based to spend-based earning. A stamp per visit is simple, but it hard-codes the subsidy problem above. A point per £1 is barely more complex for the customer and instantly fairer across sites.
If you want to keep the warmth of stamps (and many cafes do, because they feel more personal than points) consider a hybrid: stamps for visits, but redemption is for a specific item with a capped value. Your espresso bar regular still gets their 10th coffee. Your brunch regular gets the same coffee, not a free eggs benedict. We covered the underlying tradeoff in choosing between stamps and points for your cafe, and the multi-site version of that decision is heavier on points than the single-site version.
"The simplest fix is the one nobody writes down: define the reward in pounds, not in product. 'Free drink' is ambiguous across sites. '£3.50 off any item' is not."
Redemption rules that do not punish customers for choosing the 'wrong' branch#
Customers do not understand your COGS. They earned a reward; they want to redeem it. The rule has to feel fair to them and protect you. Three patterns work.
- Fixed-£ reward across all sites. '£4 off any purchase.' Simple. The customer always knows what they have. Your brunch site absorbs a bit more cost on transaction value but never the full price of a hot meal.
- Category-locked reward. 'A free drink up to £4.20.' Excludes food. Works well when your espresso bar is the loyalty driver and brunch is a separate occasion.
- Tiered reward menu. Earn 10 stamps for a drink, 25 for a drink + pastry, 60 for brunch. Lets high-value customers self-select into bigger rewards without you eating the cost on every visit.
Handling staff discretion: what your team can override and what they cannot#
Customer complaints in multi-site loyalty almost always trace back to the same sentence: 'the other branch let me do it.' If you have heard that more than twice this month, your rules are not staff-proof. They are staff-interpreted.
The fix is a two-column rulebook every shift lead has access to. Left column: what staff can do without asking. Right column: what requires a manager. Nothing sits in between.
| Staff can do (no approval) | Manager only |
|---|---|
| Add a missed stamp from a same-day receipt | Add stamps without proof of purchase |
| Honour a redemption from any site | Honour an expired reward |
| Replace a lost wallet pass via the join link | Adjust point balances manually |
| Apply the reward to the cheapest item in the order | Stack the reward with another promotion |
| Sign up a customer at the till | Issue goodwill stamps for a complaint |
Notice that 'replace a lost wallet pass' lives on the left. That is deliberate. If your loyalty card lives in Apple Wallet or Google Wallet, there is nothing to physically lose, but customers still switch phones, factory reset, or accidentally delete the pass. We wrote more on what happens when customers lose their punch card; the multi-site version is that any barista at any site should be able to resolve it in 30 seconds, because the card is tied to the customer, not the device.
The no-app advantage for multi-site consistency#
Most multi-site loyalty failures start at signup. A customer at site A is asked to download an app. They are in a queue. They do not. Three weeks later they visit site B, the barista asks the same question, they do not download then either. The programme exists on paper. It does not exist in practice.
When the card is wallet-native (it lives in Apple Wallet or Google Wallet, joined via a short link in under 20 seconds) signup conversion at the till stops being the bottleneck. That matters more across multiple sites than at a single one, because every site is a fresh chance to enrol the same customer who said no last time. We dug into this in why removing the app download barrier lifts signup rates.
The operational point: lower signup friction means you can afford a slightly tighter earn rate (say, 8 stamps for a free drink instead of 6) and still see strong enrolment, because more customers are joining in the first place. That is how you close the margin gap from the stat callout above without making the reward feel mean.
A practical ruleset template for groups running two to ten sites#
Here is a starting point. Adjust the numbers to your prices and margins, but keep the structure.
- Earn: 1 point per £1 spent, all sites, all items including retail. (Or: 1 stamp per visit, drinks only, £2.50 minimum spend.)
- Reward: £4 off any purchase at 50 points. (Or: free drink up to £4.20 at 8 stamps.)
- Cross-site: points and rewards earned at any site, redeemable at any site, no exceptions.
- Combinability: one loyalty reward per transaction, cannot be combined with other promotions or discounts.
- Expiry: points expire 12 months after last earn activity, not after first earn. Rewards once unlocked expire after 90 days.
- Staff override: shift leads can add a same-day missed stamp with a receipt. All other adjustments go to the GM.
- Disputes: 'the other branch said yes' is not a valid reason to deviate. The rule is the rule. Log the complaint, escalate to ops.
Print it. Laminate it. Stick it next to the till at every site. Brief every new starter on it. Review it quarterly with redemption data and adjust the earn threshold, not the rule structure. If you are doing this from scratch, our walkthrough on setting up a loyalty programme for the first time covers the single-site basics; this piece is what you add on top.
Common questions from multi-site operators#
Should each site have its own loyalty card or one shared card?
One shared card, almost always. Customers do not think of your group as separate businesses; they think of it as your brand. The exception is genuine franchise or licensed sites with independent P&Ls, in which case you need a written cross-redemption agreement and a clearing arrangement, because someone has to absorb the cost when a stamp earned at site A is redeemed at site B.
How do we stop customers earning loyalty on top of staff discounts or other promotions?
Write it into the terms and brief staff explicitly: one discount per transaction, customer chooses which. Most disputes come from this rule being unwritten or inconsistently applied. The customer should be told at the point of order, not at the point of payment.
What if our sites have wildly different price points (a kiosk and a sit-down)?
Switch from visits to spend, or cap the redemption value in pounds rather than as a free item. A £4 reward at a £3.50 kiosk feels generous; the same £4 against a £14 brunch feels fair to you. Capping the reward in pounds, not in product, is the simplest cross-site fairness lever.
How do we handle a customer who insists 'the other branch let me do it'?
Apologise, honour it once, log it. Then close the gap by retraining the branch that deviated. The customer-facing rule is firm but humane; the internal rule is firmer. If staff at one site keep deviating, the rule needs to be clearer or the enforcement needs to be tighter, not both.
Can we run different rewards at different sites for limited periods?
Yes, but treat them as campaigns, not rule changes. A site-specific double-points week or a free pastry with first redemption at the new branch is fine and useful. Permanent earn-rate differences between sites are where things get confusing and where complaints start.