Your annual plan is a loan, not a windfall

The first annual subscription is the best-feeling notification an indie developer gets. One tap, and a number lands that usually takes a month of monthly signups to assemble. It is worth understanding what actually happened: someone handed you a year of cash up front, and you took on a year of obligation to deliver against it. That is a loan with a service repayment schedule, and it behaves like one — right down to having a due date you probably haven't put in the calendar.

The discount is bigger than it looks

Annual pricing is almost always quoted as a percentage off, which is the least useful way to hold it. Convert it into months instead: divide the annual price by the monthly price and you get the only number that matters, which is how many months of monthly revenue the year costs.

Our own web plans are $12.99 a month or $49.99 a year. That is 3.85 — an annual subscriber pays us what a monthly subscriber pays in under four months, for twelve months of service. The common "two months free" framing works out to 10. A $6.99/$49.99 pair is 7.2. These are wildly different businesses wearing the same word. At 10 months, the annual plan is a modest prepayment discount. At 3.85, it is a bet that most of those people would have cancelled before month four anyway — which may well be true, but you should know you're making the bet.

That ratio is also your break-even line. An annual subscriber is worth more than a monthly one only if the same person, on monthly, would have churned before the ratio month. If your monthly cohort reliably sticks around for nine months, selling them a 3.85-month annual plan converts your best customers into your cheapest ones.

What you owe, and when you owe it

The cash arrives on day one. The costs arrive in twelve monthly instalments: support email, OS updates that break something, a store policy change that forces a build, and — if the product does anything with AI — a per-call bill that accrues every time that subscriber opens the app.

Spread that $49.99 out properly. After the store's 15% cut, you keep about $42.49, which is $3.54 for each month of service you owe. A subscriber who generates a dollar of model calls a month is consuming a third of their own contribution — and doing it after the money is spent. This is exactly why accountants insist on deferred revenue: an annual sale isn't earned until the service is delivered, and treating it as earned on day one is how a profitable-looking quarter turns into a cash problem in month eight.

There's a smaller version of the same trap in refunds. A year charged in one line is a more attractive thing for a customer to dispute than a $12.99 monthly charge, and a refund granted in month five takes back all twelve months of cash — including the months you already spent serving them.

The renewal cliff nobody schedules

Monthly subscriptions tell you the truth constantly. Every subscriber makes eleven more keep-or-cancel decisions over a year, and you watch each one. Annual subscribers make one, and it happens twelve months after you last learned anything about them.

So a launch month that converts 100 annual subscribers produces a full year of flat, reassuring revenue, and then a single week the following year in which you discover the renewal rate. If it's 60%, forty people leave at once, and the data that would have told you why is a year stale. Engagement analytics help, but they are a proxy — nobody has ever cancelled a subscription in your product analytics.

The practical defence is to treat the anniversary as an event you own rather than one that happens to you. Know which month your annual cohorts renew in, and make sure something genuinely good shipped in the eight weeks before it. Renewal decisions are made on recent memory, not on the twelve months of value you carefully delivered in the first quarter.

A year of revenue can hang on an expired card

Renewals fail for boring reasons — a replaced card, a bank's fraud heuristic, a larger charge tripping a limit that a $12.99 charge never did. Apple attempts to recover a failed renewal for 60 days, and App Store Connect has a Billing Grace Period setting, off by default, that keeps the subscriber's access alive while that recovery is attempted. You choose 3, 16, or 28 days.

On a monthly plan, a failed renewal costs you a month. On an annual plan, the same failure puts the whole year at risk, and a subscriber who loses access on the day of the failure has been handed a reason to not bother fixing it. Turning grace periods on is a two-minute setting that matters far more to an annual business than a monthly one.

One more asymmetry worth knowing: Apple pays you 70% of the subscription price during a subscriber's first year and 85% after they accumulate a year of paid service, and free trial days explicitly don't count toward that total. Annual and monthly plans reach the better rate at the same point — but the annual subscriber reaches it exactly at the renewal you might be about to lose. (The rest of the money path, including how tax comes off before commission does, is in when Apple and Google pay you.)

The option that splits the difference

In April 2026 Apple added a third shape: monthly subscriptions with a 12-month commitment. The subscriber pays monthly at annual-plan pricing, can see how many of their twelve payments are done, and gets email and push reminders before each renewal. Cancelling stops the subscription from renewing after the committed payments are complete rather than immediately.

For a developer, that trades the up-front cash for twelve monthly signals instead of one, at a price the subscriber accepted for a year. Two caveats before you plan around it: it's an App Store feature, with no Google Play equivalent, and it launched worldwide except the United States and Singapore, on iOS 26.4 and later with the full rollout in 26.5. If most of your revenue is American, this is not yet your answer.

So when should you sell a year?

Annual makes sense when you already have retention data showing subscribers stay past your ratio month, when marginal cost per user is low and predictable, when acquisition costs real money that up-front cash pays back, or when the product's value is seasonal and lumpy — a developer who ships two apps a year genuinely wants an annual plan for a store-listing tool, because a monthly one gets cancelled in the ten quiet months.

It makes much less sense before you know your churn curve, when each user carries real variable cost, or when the product is changing fast enough that today's promise isn't what you'll be delivering in month nine. In those cases the honest move is to keep the monthly plan front and centre, and let the year be something a happy customer asks for — not a discount you lead with because everyone else does.

None of this argues against annual plans. It argues for pricing one deliberately: work out the ratio, spread the cash across the twelve months it belongs to, put the renewal month in your calendar, and ship something worth renewing for shortly before it arrives.