A Game in a Loyalty App Buys Frequency, Not Affection

Published on August 14, 2026

Games belong in a loyalty programme when they buy app opens, consented preference data and rewards that cost less than a discount. Judge them on those three lines and the brief writes itself. Judge them on "engagement" and you will ship a slot machine nobody opens twice, then quietly retire it two quarters later while the agency invoice is still clearing.

💸 The Budget Arrived Before the Mechanics Did

The money in this category is no longer speculative. Antavo's Global Customer Loyalty Report 2026, published in February 2026 off a survey of 3,000 marketers and 10,000 consumers, puts loyalty and CRM at 51.5% of total marketing budget among programme owners, with an average reported programme ROI of 5.3x and 92.7% of programmes reporting a positive return. Nearly 60% of those marketers said they would move even more budget out of short-term promotions if they could.

That is the context a game proposal lands in, and it cuts both ways. There is real money to spend. There is also a very well-defended incumbent use for it: the discount. When you propose a game, you are not competing against doing nothing. You are competing against another 10% off, which is boring, easy to model, and works.

So the question that decides the project is not "will people play it". People will play almost anything free. The question is whether the play produces something the discount cannot.

🎯 The Three Things a Game Actually Buys

Only three outcomes have ever justified a game inside a loyalty programme, in my experience of watching these briefs come and go.

Frequency

A points balance gives a member a reason to open the app when they are about to buy. A daily or weekly game gives them a reason on the other twenty-eight days. For a grocery chain with a weekly shop, that is a fourfold increase in touchpoints. For a bank, where the natural cadence might be monthly, it is a much bigger multiple — and a much harder sell internally, because nobody at a bank has a KPI called "app opens".

Frequency only counts if it is attributable. If your analytics cannot separate "opened the app because of the game" from "opened the app anyway and also played", you have not measured anything.

Declared preference data

This is the underrated one. A game is a socially acceptable place to ask questions that a form cannot ask. Pick your team. Choose a flavour. Which of these three do you want as the prize? Every answer is a member telling you something voluntarily, in a context where answering is the fun part rather than a tax on the fun part.

Post-cookie, that data is the durable asset. The game is the collection mechanism. Brief it that way and each round should be designed to return at least one field you did not have before — not as an interstitial survey bolted on afterwards, but as the choice inside the round.

Reward efficiency

A prize distributed through a game costs less per member than the same value distributed flat, because not everyone wins. That is the entire economic case, and it is also where the legal exposure lives — see below. Done properly, a game lets you offer a headline reward big enough to be interesting at a per-member cost small enough to be affordable. Done carelessly, it is a promotion whose liability you have not modelled.

📉 Treat the Engagement-Lift Numbers You Will Be Shown With Suspicion

If you go looking for evidence, you will find a dense layer of figures circulating on vendor blogs — retention up 47%, loyalty up 22%, participation rates 75% higher in gamified programmes. Chase them to their origin and most terminate in another blog post. No sample, no control, no method, frequently no named study at all. Do not put those in a business case you will be held to.

What does hold up is smaller and duller. Euromonitor's analysis of gamified loyalty cites its Voice of the Consumer: Loyalty Survey 2025, in which 27% of Gen Z and 25% of millennials described loyalty programmes as "not fun" — a diagnosis rather than a promise. The same piece reports Marriott Bonvoy's May 2025 "Travel Roulette" campaign delivering a 6.5% rise in app downloads and a 2.3% rise in daily active users during the campaign window.

Read those two numbers properly. They are positive, they are modest, and they are campaign-scoped. A 2.3% DAU lift on a base the size of Bonvoy's is a genuinely good outcome that would nonetheless get a project cancelled if the deck had promised 47%. Set expectations at the level real published results sit at, and the programme survives its first review.

On mechanics rather than outcomes, Open Loyalty's Loyalty Program Trends 2026, drawn from 170-plus loyalty professionals across 19 industries, reports 81% of consumers expressing interest in visual progress bars and 75% of businesses prioritising real-time rewards. Both point the same way: visible progress and immediate settlement beat elaborate gameplay. That is a useful steer for anyone about to over-specify.

🧩 Match the Mechanic to the Purchase Cycle, Not to the Mood Board

The most common briefing error is picking a game because someone liked it, rather than because its natural session rhythm matches how often the customer actually transacts.

Daily cadence (telecom, convenience, coffee, transit): instant-win and streak mechanics. One round, under thirty seconds, resolves immediately. The streak is the product; the game is the wrapper.
Weekly cadence (grocery, fuel, pharmacy): collection and set-completion. Members accumulate pieces across visits and complete something. This tolerates a gap between sessions in a way a streak does not.
Monthly or quarterly cadence (banking, insurance, telco billing, travel): progression and milestone mechanics tied to real account behaviour. Do not try to manufacture a daily habit around a product nobody thinks about daily. You will pay for retention you cannot use.
Event-bounded (seasonal retail, sports sponsorship, product launch): tournament or predictor formats with a hard end date. These are the easiest to justify because the cost is capped and the measurement window is obvious.

Skill-based arcade titles work as the entertainment layer on top of any of these, and they are what most members will actually spend time inside. But the loyalty mechanic — the thing that writes to the points ledger — should be the simple one. Keep them separate in the brief so the reward logic does not depend on a leaderboard you will later need to defend against cheating.

⚖️ The Prize Mechanic Is a Legal Design Decision, Not a Legal Review

If your game awards prizes by chance, the design of the draw is a compliance question that has to be settled before the build, not after it.

In the UK, the ASA's CAP guidance on promotional marketing and prize draws sets out obligations that land directly on the game's UI. Rule 8.17 requires significant conditions — how to enter, the closing date, the nature and number of prizes, eligibility restrictions — to appear in the initial marketing, not buried in a terms link. Rule 8.24 requires draws to be conducted according to the laws of chance, using either a verifiable random process with documented evidence or an independent observer who is not a member of staff. Rule 8.20 prohibits misleading members about their chances of winning, and 8.21.1 bars charging consumers to claim a prize. CAP's own guidance also tells promoters to take legal advice specifically to confirm they are not running an unlicensed lottery.

Three practical consequences for the brief:

The odds have to be a specified, documented parameter, owned by someone, logged on every award, and reproducible months later. "The developer tuned it" is not an answer when a regulator asks.
Prize inventory and liability are modelled before launch. Every instant-win outcome is a commitment you have to honour, including on the day the campaign goes unexpectedly well.
The rules differ by market. A mechanic that is a clean prize draw in one jurisdiction can require a free entry route, a skill element, or a licence in another. If you are running a multi-country programme, the mechanic gets designed for the strictest market in scope or gets regionalised deliberately — those are the only two options, and picking neither is what produces an emergency takedown.

None of this is exotic. It is just work that has to happen at brief stage, because it constrains the game design.

🛠️ What "Integration" Actually Means Once You Get Past the Demo

A licensed HTML5 game drops into a native loyalty app through a webview, or into the web programme through an iframe. That part is genuinely easy. The parts that consume the timeline are these:

Identity. The game has to know who is playing without a second login. That means passing a signed token from the app session into the webview, and validating it server-side before any award is written.
The points ledger. Awards must be written server-authoritatively. A game that decides in the client how many points you won is a game that will be modified in the client. Assume someone will try, because at any meaningful scale someone always does.
Idempotency. Mobile connections drop mid-award constantly. Without idempotency keys on the write, you will double-credit some members and zero-credit others, and both generate support tickets.
Weight and cold start. A loyalty app member is not a gaming-portal visitor; they arrived to check a balance. If the game takes several seconds to reach a first playable frame on a mid-range Android device, a large share of them are gone before it loads. Budget the first frame, not the total download.
Rotation. One game forever is the single most reliable way to kill this. Novelty is a real part of why the frequency lift exists, and it decays. A catalogue you can rotate on a schedule is worth more than one bespoke title, which is why most programmes that run this well license breadth and build depth only where the brand demands it.

That last point is the honest argument for licensing over commissioning, and it is a capacity argument rather than a quality one. Bespoke development gives you exactly the game you specified. It does not give you eleven more next quarter.

📊 Instrument These Six Things, and Ignore "Plays"

Total plays is a vanity number that goes up whether or not the programme is working. Measure instead:

Incremental app opens attributable to the game, against a holdout group that does not see it.
Return rate at day 7 and day 30 for players versus matched non-players in the same tier.
New data fields captured per active player — the actual output of the data play, and the easiest thing to forget to log.
Cost per consented profile, compared with what you pay for the same field through any other channel.
Redemption rate change, because a game that raises engagement while lowering redemption is moving members away from the behaviour that makes the programme pay.
Incremental margin versus the discount control. The comparison that decides renewal. Run it from day one or you will be arguing from anecdote when the budget review comes.

If you cannot hold a control group out for political reasons, say so explicitly in the plan and accept that your results will be directional. Pretending a pre/post comparison is causal is worse than admitting the limitation.

🚫 Five Ways These Programmes Fail

Shipping one game and calling it a channel. Play decays within weeks without rotation, and the post-mortem blames "gamification" rather than the content plan that never existed.
Letting the game sit outside the points ledger. If playing does not touch the currency members already care about, it is a free arcade attached to a loyalty app, and it competes with the programme instead of feeding it.
Not modelling prize liability. The campaign that overperforms is the one that hurts. Cap exposure in the mechanic, not in the terms and conditions.
Ignoring who is actually playing. Casual games attract under-18s regardless of who the programme was designed for. If minors can reach it, age gating, data minimisation and prize eligibility all need answering before launch, not after a complaint.
Treating launch as the finish line. Odds tuning, prize mix, rotation cadence and the data questions asked in-round are all live parameters. Programmes that assign an owner to those parameters outperform programmes that assign a launch date.

🎮 Where a Licensed Catalogue Fits

Forestry Games has been licensing games since 2017 and maintains a catalogue of 1,049 titles across HTML5 and Android, with in-house HTML5 development for cases where an off-the-shelf title will not carry the brand. For loyalty work, the relevant part is usually breadth and rotation rather than any single title — a licensed catalogue gives a programme something new to put in front of members every month without a new development cycle each time. Where a campaign needs its own mechanic, branding or prize logic, that is custom marketing game territory, and the same catalogue can sit underneath it as the everyday layer. Programmes running their own player-facing destination alongside the app tend to end up looking at white-label portals for the same reason.

Forestry Games works with branded IP and has brand partnerships including Disney, Nickelodeon, Cartoon Network and Warner Bros. What any specific brand will approve for a specific loyalty programme is a matter for that brand's approval process, and any agency telling you otherwise before the submission is guessing.

🧭 What to Do Next

Before you brief anything, write down the three numbers this game is supposed to move — an app-open target, a data-fields-per-player target, and a cost-per-reward target — and get the loyalty owner to sign that page rather than the creative deck. Then pick the mechanic from the purchase cycle, settle the prize structure with legal before the build, and reserve budget for rotation from the start rather than treating it as a phase two nobody funds.

If those three numbers are not obviously better than what the same money does as a straight discount, the honest answer is to skip the game this cycle. That answer is available, it is cheap, and it is a great deal less embarrassing than retiring a game six months after the launch press release.