Paper vs digital loyalty cards: one real number

Paper vs digital loyalty cards: what each side actually claims

I read the first page of results on 29 September 2026 and wrote down every rate, along with whatever each page offered as a source.

The gaps are not rounding. One page says roughly four in ten paper cards get finished, another says roughly two in ten, and the digital column runs from 34% to 85% with nothing underneath most of it.

Stampeo is the one page that names a study, which is worth more than the number it draws from it. Both the 19% and the 34% in that row came from cardboard cards stamped by hand, and the thing the experiment varied was not whether customers could see their progress. It is still the right study to reach for, so let's read it.

Follow the 39% and the trail stops at a paywall

Two of the pages above carry the same figure: 39% of customers abandon a loyalty program because they lost the card. FaveCard attributes it to Statista without naming a report, which is still more than the others manage, since they print the number and move on. So I went looking for the report.

Statista does hold two relevant items. One is a June 2015 survey of 402 US millennials asked why they abandoned loyalty programs, with multiple answers allowed. The other is a 2013 UK survey of 2,159 people asked, among other things, whether they find it hard to remember to bring their loyalty card shopping. Both sit behind Statista's paywall, so the percentages are not readable without an account.

Neither one counted a paper card. Neither one measured a completion rate. They asked people to pick reasons from a list, eleven and thirteen years ago respectively, and a stated reason for quitting a supermarket points scheme is not a measurement of how many cafe stamp cards end up in the wash.

One field experiment, 300 cards, nine months

The car wash study is real and you can read it. Joseph Nunes and Xavier Drèze ran it for a paper published in the Journal of Consumer Research in February 2006, The Endowed Progress Effect. Study 1 is the part that matters here.

On the second and third Saturdays of April 2004, the car wash randomly handed out 300 loyalty cards. Half required eight stamps and arrived empty. Half required ten stamps and arrived with two already affixed, described to customers as a promotion that day. Both groups therefore needed exactly eight more paid washes. Over the next nine months the proprietors stamped each card with the date of the visit, and 80 cards came back for a free wash between 5 June and 17 September 2004, recording 720 visits in total.

The eight-stamp cards were redeemed by 19% of their holders. The ten-stamp cards with the head start were redeemed by 34%, a gap the authors report as statistically significant at better than the one percent level. Same eight washes, same reward.

Hold those two numbers next to the table above. A paper card in a real business, measured rather than estimated, landed at 19% when it started empty and 34% when it started two stamps in. The pages that put paper at 18-25% are close to the empty-card case. The ones that put paper at 40% are close to the head-start case. Neither says which it means, and neither cites this study. Note also what the two conditions had in common: both were cardboard, both were stamped by hand, and the 15-point gap between them had nothing to do with the medium.

The only measured digital rates come from one company's own platform

One vendor in this category has published its method. It happens to be a competitor of ours. FaveCard's 23,000-stamp study, covering 23,296 stamp events from 10,690 customers across 1,013 small businesses between 20 January and 18 May 2026, puts the average completion rate at 5% to 8%, with its best-designed cards reaching 11.4%.

That is roughly one tenth of the 65-85% the other pages promise for digital.

FaveCard is candid about the limits: the data covers its own platform only, and the observation window is four months, so the slowest cards are undercounted, and the company says plainly that the absolute numbers will differ elsewhere. All true. It is still the only digital figure in this whole category that arrived with a described dataset rather than an adjective, and a 5-8% number from a vendor with every commercial reason to publish a bigger one is worth more than a 75% number from a vendor with no dataset at all.

The research comparing print with digital points both ways

The academic literature would settle this quickly if paper were simply the weaker medium, and the two field studies closest to the question would point the same direction. They point in opposite directions.

A field experiment run with Fujifilm on 7,500 registered customers, published in the *Journal of Advertising Research* on 1 December 2024, sent each customer two coupons a month apart for a photo album service. The group that got a postcard first and an email second redeemed at 20.7%. Postcard second, email first: 17.8%. Email both times: 3.5%. Getting a postcard even once lifted redemption more than fivefold. Lead author Taku Togawa attributes that to cognitive engagement, and the follow-up survey backs him up: the postcard recipients understood what the offer actually was better than the email recipients did.

Pull the other way and there is Son, Oh, Han and Park's 2020 study in Information Systems Research, which tracked 7,712 South Korean rewards members over seven months. Half collected points on a plastic card, half on the company's app. App users spent more and redeemed points more actively. They also visited more stores and spent less at any one of them, which is the opposite of what a single cafe wants. Members chose their own format rather than being assigned one, so the comparison is correlational, a caveat Notre Dame's own writeup of the paper repeats.

So the best available evidence says print gets read more carefully and digital gets redeemed more often, which is an awkward result for anybody selling either one. Call it a split decision. It is nowhere near the rout the comparison tables describe.

What has actually been measured is the distance to the reward

Both sides of the research agree on one thing, and it is not the format.

Ran Kivetz, Oleg Urminsky and Yuhuang Zheng ran the cafe version of the car wash experiment for The Goal-Gradient Hypothesis Resurrected, in the Journal of Marketing Research in 2006. Customers offered a 10-stamp card finished their ten purchases in 15.6 days on average. Customers offered a 12-stamp card carrying two free stamps finished the same ten purchases in 12.7 days, nearly three days faster, with the median dropping from 15 days to 10.

FaveCard's platform data finds the same shape at a much larger scale. Sorted by how long the card takes to finish, its completion rates were:

• 2 to 4 weeks: 11.4%
• 4 to 8 weeks: 9.3%
• 8 to 16 weeks: 4.0%
• 16 to 32 weeks: 1.2%
• 32 weeks or more: 0.8%

A card that takes eight months to finish completes at about a fourteenth of the rate of one that takes a month. Nothing about paper or plastic explains that curve. It is the wait, and by the company's own stamp-count breakdown, 5-stamp cards were completed by 27.9% of customers against 5.6% for 10-stamp cards.

If your card currently takes a customer four months to finish, moving it into Apple Wallet will not fix it, and cutting it from ten stamps to six probably will. Do that first. Then pick a format.

Three things the format does decide

Completion rate is the wrong place to look for the difference between the two formats, because the honest answer is that nobody has measured it properly. Three other differences are not in dispute. They are matters of fact rather than statistics.

You can see who is close. A stack of stamped cardboard tells you nothing about how many customers are two stamps from a reward, and a digital card tells you exactly, which matters given everything above about the distance to the reward.

You can send a reminder, up to a point. A pass in Apple Wallet or Google Wallet can push a notification, and Google's documentation caps that at three push-triggering messages per 24 hours. We have written separately on how often to actually send them. The answer is far below the cap.

The card survives a lost wallet. This is the claim the comparison tables inflate hardest, and it is still true in its unexaggerated form: a pass restores with the phone, a piece of card does not.

When paper is still the right answer

Paper is still the right answer in a few cases. A market stall or a one-day pop-up, where handing someone a card beats explaining a QR code to them. A regular clientele that is older and would rather carry the card, a preference the Fujifilm result suggests is less foolish than it usually gets treated as.

The third case is the one I would push hardest. If you have never run a program at all, fifty printed cards and a rubber stamp will tell you in a fortnight whether people want the reward you picked, which is the question underneath whether a loyalty program works for you and the one no platform can answer on your behalf. Switching afterwards is cheap. Moving to a virtual punch card is a smaller job than choosing the reward was.