In brief
Six shifts, written as observations rather than forecasts: wallets replacing apps, the retreat from downloads, paid tiers at small scale, the consent squeeze, AI-generated content flooding search, and rewards moving back from currency to product.
Before the list
These are observations and arguments, not forecasts, and this post is dated deliberately. The blog on this site is the collection that is allowed to go out of date — the loyalty program library is the one that gets corrected. If you are reading this well after September 2026, treat it as a record of what an argument looked like at the time.
No statistics follow, because there are none here worth trusting at second hand. Each shift is stated as a mechanism you can check against your own experience.
1. The wallet is becoming the carrier, not the app
Apple Wallet and Google Wallet are now where most people expect a card-shaped thing to live. Boarding passes and event tickets trained everyone into the habit, and a loyalty card is the same shape.
The consequence for a small business is that the carrier question is largely settled. You are not choosing between building an app and printing paper; you are choosing between a browser link and a wallet pass, and both cost nothing per customer.
What it means for you: stop evaluating loyalty tools on whether they have an app. Evaluate them on whether the pass updates reliably after a scan, which is the part vendors differ on. A pass showing a stale balance is worse than no pass. There is more on the trade-off in how to make loyalty cards.
2. The retreat from downloads
Related but distinct, and broader than loyalty: asking someone to install something in exchange for a small benefit has become a much harder ask than it was a few years ago. Storage is contested, permissions are contested, and people have learned that most single-purpose apps get used twice.
A loyalty program is precisely a small benefit. The download sits between “willing to join” and “joined”, and it loses most of the people standing at the counter with four seconds of patience.
What it means for you: measure enrolment as a share of the people who were asked, not as an absolute number. If it is low and the ask is happening, the friction is the install.
3. Paid tiers appearing at small scale
Paid membership used to be a large-retailer mechanic. It has moved down: independent cafés selling a monthly coffee subscription, gyms adding a priority tier, restaurants selling a supper club.
The mechanism that makes it work is unchanged and worth naming — the customer pays first, which inverts who is carrying the risk and makes them visit more to recover the fee. What changed is that the billing infrastructure to run one became ordinary rather than a project.
What it means for you: worth modelling if your customers are frequent and your ticket is small. Worth ignoring otherwise. The arithmetic is in paid loyalty and membership programs.
4. The consent and data-minimisation squeeze
Regulators and platforms have both narrowed what you can collect and what you can do with it afterwards. The practical effect on a small program is not a compliance project; it is a design constraint that arrives earlier than it used to.
A loyalty enrolment used to be treated as blanket permission to market. It is not, and treating it as one is both a legal exposure and the fastest way to make people unsubscribe.
What it means for you: ask for the minimum that identifies the customer again, state what you will send, and keep the two consents separate — joining a card and agreeing to receive offers are different decisions. Collecting less is also less to be responsible for.
5. AI-generated content flooded the search results
The web filled up with competent, sourceless articles about loyalty programs, most of which say the same six things. That has two consequences for a shop.
The first is that generic advice is now free and infinite, and therefore worth what it costs. The second is more useful: because generic advice is abundant, specific and local information became comparatively more valuable. Your own numbers, your own trade, your own street.
What it means for you: if you write anything at all — a page about your program, an answer to a customer question — the thing that makes it worth reading is the part only you know. What your reward costs you. What your regulars actually do. What you tried that failed.
6. Rewards moving back from currency to product
A points balance is a currency you issue, and a currency loses value when prices rise. A reward denominated in a product does not: “the tenth coffee is free” reprices itself automatically every time you reprice coffee.
After a stretch of unstable pricing, that difference stopped being theoretical. Businesses that ran points found themselves either eroding the reward’s real value by leaving the rate alone, or repricing points and having to explain it.
What it means for you: if you are choosing now, a product-denominated reward removes a maintenance job you would otherwise carry forever. This is also the argument that the Spanish-language Argentina page makes at much greater length, for a market where it has been obvious for years.
What did not change
Worth ending here, because trend lists imply more churn than there is.
The threshold still has to match the customer’s existing visit rate. Redemption still has to work on a Saturday without a manager. Staff still have to ask, and they still stop asking after two weeks unless someone checks. The reward still costs you cost of goods and not menu price.
Every one of those decides more about whether a program works than all six shifts above combined, and none of them is new. If you are choosing what to do this year, how to build a loyalty program is a better use of an hour than this post was.