Retail Apps: Loyalty or Friction

A loyalty programme has exactly one moment where it succeeds or fails, and it is not the sign-up.

It is the four seconds at the counter when the customer has to be identified. If that takes too long, staff stop asking, customers stop bothering, and within a month you are paying a monthly fee for a database of people who joined once.

Every comparison of loyalty platforms is written about points structures, reward tiers and email automation. Here is the same market compared on the four seconds.

The identification step, and why it decides everything

There are four ways a customer gets recognised, and they are not close to equivalent.

They open an app and show a code. Requires the app installed, the phone unlocked, the app found, and the code loaded. Realistically eight to fifteen seconds with a queue behind them. This is the highest-friction option and it is the one most custom-built loyalty apps default to.

They scan a QR code on the counter. The customer scans, rather than staff scanning them. Fast, needs no app installed, and works from the camera. This is what most modern small-retail platforms use.

They give a phone number or email. No technology on the customer’s side at all. Slower to type but universally available, and it works for the customer who has left their phone in the car.

The POS recognises the payment card. Invisible to everyone. The best experience by a distance, and only available if your point-of-sale platform supports it.

The pattern is clear enough to be a rule: the options requiring an app are the slowest, and the ones requiring nothing from the customer are the fastest. Which is awkward, because an app is what most retailers think they are buying.

★ Insight ------------------------------------- The reason app-based loyalty underperforms in small retail is that app install is a cost paid by the customer for a benefit that arrives much later. A coffee shop asking for an install to start a ten-stamp card is asking for a commitment worth more than the reward. QR and phone-number identification work because they defer the commitment until after the value has been demonstrated — the customer collects three stamps first and installs something later, if ever. -------------------------------------------------

The platforms, compared on that step

Stamp Me is Australian and built around exactly this problem: digital punch cards identified by QR and NFC rather than deep point-of-sale integration. That design choice is why it deploys fast — it does not need to talk to your till — and it is a good fit for cafés, salons and single-site retailers running a simple stamp model.

Marsello originates in New Zealand with a strong presence across this region, and its differentiator is the data model rather than the counter interaction: a customer who buys both online and in-store accumulates points and history into one profile. If you sell through both channels and currently treat them as two businesses, that is the thing worth paying for.

LoyaltyLion sits further up-market, aimed at established e-commerce with analytics and benchmarking, with pricing to match.

Pricing across the category is usually tiered on active customers or order volume and it changes regularly — check current pricing with the vendor rather than trusting any comparison page, this one included.

What the platforms are really selling

Once you have chosen an identification method, the platforms differentiate on something else entirely, and it is worth naming so you are comparing the right thing.

The database, not the card. The stamp card is the visible part and the least valuable. What you are buying is a list of customers with purchase history attached, and the ability to contact them. A programme that collects stamps and no contact detail has given away margin for nothing.

Automated re-engagement. The single feature with the clearest return is the one that contacts a customer who has not been in for a while, without anyone remembering to do it. Birthday and lapsed-customer campaigns are standard across the category now, and they are the reason a platform beats a paper card by more than convenience.

Whether online and in-store are one customer. If you sell through both, the question is whether a person who buys online and in store is one profile or two. Two profiles means two histories, two segments, and a programme that cannot see half of what someone spends.

The order of those matters. A retailer choosing on points flexibility and reward tiers is optimising the part customers care least about. Ask instead what you will be able to do with the list in twelve months.

The number that tells you if it is working

One metric, and it is not sign-ups.

The share of transactions that are identified. Of every sale you put through in a week, how many were attached to a known customer. That single number captures whether staff are asking, whether the identification step is fast enough, and whether the offer is worth joining — all of which are invisible in a sign-up count.

Below about a fifth, the programme is decoration. Above about half, it is a genuine asset and the data is worth acting on. Most small retailers we see who have never measured it are surprised by how low it is, and the cause is almost always the four seconds at the counter rather than the offer.

When a custom app is the right call

We build apps, and for most small retailers we would advise against building this one. Three cases change that.

The app is a genuine second channel, not a loyalty card. Ordering ahead, a table booking, a service history, a stock check. Loyalty then rides along on something the customer already opens for its own reasons — which is the only way an installed app works in retail.

You have multiple sites and the programme is part of the brand. Consistency across locations, a single balance, and an experience you control rather than one carrying a vendor’s design.

The programme rules are genuinely unusual. Tiering by product category, partner redemption across businesses, anything involving membership or entitlement rather than points. Generic platforms model points and stamps; anything structurally different gets worked around manually, and the manual step is where programmes die.

Outside those, the correct build is not an app. It is a fast identification method plugged into a platform that already exists, and the money spent on making the offer worth joining.

One rule about the offer itself

Worth stating because it is the most common reason a technically sound programme fails: the first reward has to arrive early enough to be believed.

A ten-visit card for a business a customer visits monthly is a promise with a ten-month horizon. Nobody is persuaded by that. Something small at visit two, even trivially small, converts a sign-up into a habit — and the habit is what you are actually buying.

The check worth running

Go and stand at your own counter at your busiest time and watch ten transactions.

Count how many customers are asked about the programme, and of those, how many complete the identification step without the queue noticeably slowing.

If staff are not asking, the friction is already too high and no platform change will fix the offer. If they are asking and customers are fumbling, you have chosen the wrong identification method — and that is a fixable problem that has nothing to do with which vendor you picked.

Vendor capabilities taken from publicly available documentation, September 2026. Products and pricing change; confirm with the vendor before buying.

---

Awesome Apps builds custom iOS and Android apps for Australian businesses. See how we work and what we have built. Retail websites and online stores come from Cosmos Web Tech, IT and cloud from Cloud Geeks, and we are part of Ganda Tech Services.