Ellie Marsh, who runs Pour & Press on Gloucester Road in Bristol, kept a shoebox behind the till for six months. Every time a customer came in and said they'd lost their stamp card, she dropped the replacement stub in the box. By June she'd counted 312 replacements against 847 original cards issued. That's a 37% loss rate. More interesting: 68% of those lost cards had only one or two stamps on them.
The card wasn't failing at visit nine. It was failing at visit two, which is the visit that actually matters.
The second visit is where programmes live or die#
Most comparison articles argue about the tenth visit: is a free coffee after 9 stamps better than 200 points for a £5 reward? That debate misses the point. If a customer doesn't come back a second time, the reward mechanic is irrelevant.
Stamp cards have a psychological advantage here called the near-completion effect, documented by Nunes and Drèze in their 2006 endowed-progress study. A card with 2 stamps out of 9 feels closer to complete than a points balance of 40 out of 500. The visual gap is smaller. The motivation is immediate.
Points programmes only start feeling worthwhile once the balance accumulates. In a café context, where the average ticket is £3.80 and the customer is out the door in under four minutes, that accumulation takes weeks. Most customers drop off before the points feel real.
"Points felt like homework. Customers would ask how many they had, and my staff would be fumbling with the iPad while the queue built up. With stamps, they look at the card and they know."
Side by side, in a café context#
| Factor | Paper stamp card | Points programme (app) | Digital stamp card (no app) |
|---|---|---|---|
| Signup friction at till | Low (hand over card) | High (app download, email, password) | Low (tap or scan, no download) |
| Lost within 3 visits | 37% (Pour & Press data) | N/A, lives on phone | N/A, lives on phone |
| Drives visit 2-3 | Strong (near-completion effect) | Weak (balance feels small) | Strong (visual progress retained) |
| Cognitive load at redemption | None, customer sees it | High, staff checks balance | None, customer sees progress |
| Cost per month (small café) | £15-30 printing | £80-200 + % of rewards | £20-50 flat |
| Works across multiple sites | No | Yes, if same vendor | Yes |
The app download is where points programmes bleed customers#
Here's the conversion killer no one in the SERP is talking about. If your loyalty programme requires a download, you're asking a customer who is holding a flat white and a pastry, with a queue behind them, to:
- Unlock their phone
- Find the app store
- Search for your app
- Agree to permissions
- Create an account
- Verify an email
- Find you again in the app
- Show the barista a QR code
A 2024 study by Airship put loyalty app signup conversion at the point of sale at under 12%. That means 88 out of every 100 customers who say yes to loyalty never actually join. Your points programme has an 88% leak before anyone earns a single point.
The honest answer: it depends on your average ticket and visit frequency#
A points programme makes sense when the average transaction is high and infrequent (think a wine merchant with a £40 average spend and monthly visits). The balance accumulates to something meaningful quickly, and the maths justifies an app.
A stamp-based system wins when the ticket is low and the visits are frequent, which is exactly the café profile. The visual progress rewards the second visit. The reward feels proportional to the effort. Staff don't have to look anything up.
The real question isn't stamp versus points. It's physical versus digital stamp card. Paper loses 37% of your customers in the first three visits. A digital stamp card with no app download keeps the psychology and removes the churn.
"We ran points for four months and got 94 signups. Switched to a digital stamp card and got 312 in the first six weeks. Same footfall, same staff, same offer."
How many stamps should a café card require?
Between 7 and 9 works best for most cafés. Fewer than 7 and the reward feels too cheap to value. More than 10 and the near-completion effect weakens. Pour & Press settled on 8 after testing 6 and 10.
Won't customers game a digital stamp card?
Rarely, if stamps are added by staff at the till rather than self-claimed. The cost of a fraudulent free coffee is lower than the cost of losing 37% of your customers to paper card churn.
What about older customers who don't use smartphones?
Keep a paper option available for the small minority who prefer it. At Pour & Press this is around 8% of loyalty members. The other 92% prefer the digital card because they can't lose it.
Do digital stamp cards work across multiple café sites?
Yes, if the system isn't locked to a single POS vendor. This is where multi-site operators get caught out. Check before you commit that the card balance syncs across locations without extra fees.
How quickly can we see results?
Repeat visit rate shifts show up within 4 to 6 weeks. If your programme isn't moving the second-visit number in that window, the mechanic or the signup flow is the problem.