A Prepaid Games Bundle Doesn't Churn. It Fails to Bill.

Before you license HTML5 games for a carrier bundle, model the prepaid balance: most loss on an operator games service is failed billing, not churn.

Every games VAS business case I have read models the same way. Gross adds per month, a churn percentage borrowed from a streaming deck, an ARPU number, multiply out, done. It survives about six weeks of live traffic. Then the base flattens while acquisition spend stays where it was, and somebody starts asking whether the catalogue is good enough.

The catalogue is usually fine. What is happening is that the service is billing against a balance that is empty most of the time, and the reporting is filing every failed attempt under a word — churn — that implies a person made a decision. Almost nobody made a decision. There just wasn't fifteen cents on the line at 02:00 when the renewal ran.

📉 "Churn" Is Two Different Losses Wearing One Number

Split the number before you touch anything else. There are two ways a subscriber leaves a prepaid games bundle and they have nothing in common except the row they land in on the dashboard.

• Voluntary exit. The subscriber sent STOP, used the unsubscribe flow, or cancelled through the operator's self-care app. They evaluated the service and rejected it. This is the loss your catalogue and your merchandising can actually influence.
• Billing failure. The renewal attempt returned insufficient funds. The subscriber did nothing. They may not know the service lapsed, and in most flows they get no notification that would tell them.

These demand opposite responses. Voluntary exit says fix the product. Billing failure says fix the cadence, the retry window, and the price point — and possibly the file sizes, which I will come back to, because on prepaid that is a billing problem and not a UX one.

You will not find a published industry benchmark for prepaid billing-success rates on games VAS. I looked. That absence is itself worth sitting with: it means nobody can hand you the number, so the only version that exists for your service is the one you instrument. If your aggregator's reporting collapses "insufficient funds" and "subscriber cancelled" into a single churn field, that is the first thing to change, and it is usually a configuration request rather than a project.

💸 The Same Balance Pays for the Subscription and the Data

Here is the part that gets missed, and it is the reason a games bundle behaves differently from a ringtone bundle or a horoscope SMS.

A games service consumes data. Unless the bundle is zero-rated end to end — and most are not, or are only zero-rated on the portal shell and not on the game assets served from a third-party CDN — every megabyte your catalogue downloads is drawn from the same prepaid balance that has to survive until the next renewal attempt. You are not competing with rival games services for that balance. You are competing with your own subscription fee.

The affordability data explains why the margin is so thin. The ITU's Facts and Figures 2025 affordability release puts the global median price of a data-only mobile broadband basket at 1.4% of GNI per capita, down from 1.5%. That median is comfortable. The distribution is not: subscribers in low-income economies spend roughly 22 times more of their income on the same basket than subscribers in high-income economies, and lower-middle-income subscribers about seven times as much. Of 205 economies with data, 130 meet the 2%-of-GNI affordability target for mobile broadband — but among low- and middle-income economies, only 54, roughly four in ten, meet it for either basket.

In those markets the data your games consume is a material share of what a subscriber can spend on mobile at all. A 40 MB title played twice a day is not a design detail. It is a line item competing with your renewal.

The device side compounds it. GSMA's State of Mobile Internet Connectivity, published 9 September 2025, reports 4.7 billion people using mobile internet — 58% of the world — with 3.1 billion more, 38% of the global population, living inside coverage they do not use. An entry-level internet-enabled handset costs 16% of average monthly income across low- and middle-income countries, and 48% for the poorest fifth. The subscriber you are billing is on a cheap device, on a metered balance, in a market where both are stretched.

🔁 Retry Cadence Is a Product Decision, Not a Billing Setting

Because balances refill unevenly — a top-up on payday, a small one mid-week, nothing for four days — when you attempt the charge matters as much as what you charge.

Most operator platforms expose more control here than the commercial team realises. The levers worth arguing over:

• Cadence. Daily and weekly renewals collect against small balances more often than a monthly charge does. A monthly price point on a prepaid line is a single large ask on one specific day; miss it and the whole month is gone.
• Retry window. A failed attempt is not a verdict. Re-attempting across a spread of hours and days catches top-ups the first attempt missed. What you should refuse to accept is retrying so aggressively that it looks like harvesting a balance the moment it appears — which is both a regulator problem and a trust problem.
• Grace period. Keep access open during retries. Cutting a subscriber off at the first failure and re-acquiring them later costs an acquisition fee to solve a cash-timing problem.
• Time of attempt. Renewals that fire at a fixed hour against a population that tops up at predictable times are leaving collections on the table.

None of this needs a licensing conversation. All of it should be tuned before anyone concludes the games are the problem.

📊 2.3 Billion Accounts, 593 Million Active — Read It as a Warning

The clearest illustration of registered-versus-paying in this market is not from gaming at all. GSMA's State of the Industry Report on Mobile Money 2026, published 24 March 2026, reports 2.3 billion registered mobile money accounts at the end of 2025 and 593 million 30-day actives. That is a monthly active rate of 25.7% — and the report notes this is the highest since 2021. More than $2 trillion moved through those wallets in 2025, double the 2021 figure.

Two things to take from it. First, in prepaid-first markets a three-to-one gap between registered and active is normal across an entire category, not a sign of a broken product. If your games bundle shows something similar, the diagnosis is not automatic. Second, the money is genuinely there and growing fast — which is the argument for fixing collection mechanics rather than concluding the market cannot pay.

It also points at a rail worth asking your operator about. Where mobile money is dominant, a wallet charge and an airtime deduction have different success profiles, and the better platforms will let you attempt both.

⚖️ Catalogue Weight Is a Billing Metric

This is where the licensing decision stops being separate from the billing one.

If data cost suppresses play, and suppressed play suppresses perceived value, and perceived value drives whether someone tops up enough to cover a renewal, then the download weight of your catalogue sits upstream of your collection rate. Most licensors will quote you a catalogue size. Fewer will hand you a per-title asset budget without being asked.

What to ask for before you sign, specifically:

1. Total transferred bytes per first play, per title — not the ZIP size. Measure it on a cold cache, on a mid-range Android device, which is what your subscribers are on.
2. Which titles work offline after first load, and whether service-worker caching survives the browser's storage eviction. A title that re-downloads its assets every session is a recurring charge on your subscriber's balance.
3. Whether assets can be served from your own origin or must be loaded from the licensor's CDN. This decides whether operator zero-rating can cover the games at all, and it is the single highest-leverage clause in a carrier deal.
4. A low-bandwidth build variant, where the licensor offers one — compressed audio, reduced texture sets. Some do; you have to ask.

A 200-title catalogue that averages 6 MB per play will out-earn a 900-title catalogue averaging 25 MB in exactly the markets where prepaid dominates. Sizing the catalogue by title count is the wrong axis. Size it by what a subscriber can afford to load in a month.

🧾 The Reporting You Need Is a Licence Clause, Not a Ticket

Almost none of the above is measurable after the fact if it was not specified up front. Operators routinely discover in month five that they can see plays but not bytes, sessions but not per-title weight, and that the licensor has no obligation to provide either.

Put it in the agreement: per-title transferred size at delivery, a commitment to notify you when a build's weight changes materially at any update, and the right to test a sample of the catalogue on your own devices before acceptance. These are ordinary asks. A direct licensor will agree to them because the data already exists on their side. A reseller two steps removed from the developer often cannot, and that tells you something useful about which supplier you are actually dealing with.

🚫 Five Ways Operators Misread a Prepaid Games Bundle

• Benchmarking churn against postpaid or card subscriptions. Different failure mode entirely. A card has an issuer, an expiry date and sometimes an overdraft. A prepaid balance is zero by design between top-ups.
• Buying more titles to fix a flat base. If the loss is billing failure, catalogue expansion changes nothing and adds cost.
• Zero-rating the portal but not the games. Subscribers hit the shell for free, then pay to load every asset. This produces excellent portal traffic and terrible retention, and it looks like a content problem on every dashboard.
• Treating a failed renewal as a cancellation. Ending the session, revoking access and re-marketing to the same person a week later means paying an acquisition cost to solve a timing problem.
• Pricing monthly because the deck was built monthly. The cadence should follow how the market tops up, not how the spreadsheet was laid out.

🎯 What You Get When You License HTML5 Games From a Direct Licensor

Forestry Games has licensed games since 2017 and its catalogue runs to 1,049 titles across HTML5 and Android APK. For a carrier or portal deal that matters in a practical way: a licence covers HTML5 builds you can host on your own infrastructure — which is what makes operator zero-rating possible — plus APK builds where a native install suits the market better, and source where the deal includes it. Branding, localisation and hosting arrangements are part of the scope conversation, not a separate project.

A catalogue that size is not meant to ship whole. The useful exercise is filtering it to the few hundred titles that match your subscribers' devices, languages and data budgets, and building the subscription games portal around that subset. If you want to see the range first, browse the catalogue, or look at the telecom and operator games section for the carrier-facing view. Where a native build is the better fit — offline play, low-connectivity markets, OEM channels — the same conversation covers Android and APK game licensing rather than starting a second procurement.

🧸 Licensing Branded Games for Campaigns, Portals and Apps

Forestry Games also works with branded IP and has brand partnerships including Disney, Nickelodeon, Cartoon Network and Warner Bros. Businesses can license branded games through Forestry Games for campaigns, operator portals, events and apps. On a carrier bundle, branded titles typically carry their own acquisition pull, which changes the marketing maths on a service that otherwise has to buy every subscriber — worth scoping alongside the generic catalogue rather than after it.

The next step is concrete: ask for a licence scope for your market and device profile, or request a white-label game portal demo you can put in front of the operator's product team.

🧭 The Number to Pull Before the Next Renewal

Go to your billing platform and pull one figure: what share of renewal attempts in the last 30 days returned insufficient funds. Then pull the second: how many of those subscribers topped up within the following 72 hours. If the first number is large and the second is meaningful, your service does not have a churn problem or a catalogue problem. It has a collection-timing problem, and it is cheaper to fix than anything else on your roadmap.

Do that before you renegotiate the catalogue. Then, when you do license HTML5 games for the next phase, you will be buying against a number you actually understand — build weight and per-title data cost, not title count.