DEV Community

Cover image for The App Store Fee Nobody Budgets For Is the Calendar
Sonia Bobrik
Sonia Bobrik

Posted on

The App Store Fee Nobody Budgets For Is the Calendar

Ask an iOS developer what the App Store costs and you'll get a percentage: 30, or 15 if you're small. Ask how long their money spends in transit and you'll usually get a shrug, even though that second number quietly decides how fast an app can grow. A subscription sold on the first day of one of Apple's five-week fiscal months typically reaches your bank about 68 days later, and by then you've already paid for the ads, servers and API calls that produced it. A recent essay on the cash conversion economy argues that growth only counts once it turns into cash a business can actually use, and app stores make that idea uncomfortably concrete. So I modeled the gap for a small subscription app, and the result changed how I think about the whole in-app purchase versus web checkout debate.

How a Tap Becomes a Deposit

Apple doesn't pay on the calendar you use. It runs a fiscal calendar in which every month is exactly four or five weeks long, and its developer documentation commits to sending proceeds within 45 days of the last day of the fiscal month in which a transaction completed, bundled into one payment per currency. In practice, developers have generally seen the deposit about 33 days after a fiscal month closes. A sale on the last day of a fiscal month therefore waits roughly 34 days, while a sale on the first day of a five-week month waits about 68. Averaged over a quarter, the wait comes out to almost 49 days.

Renewals can stretch it further. When a subscriber's card fails, Apple keeps retrying for up to 60 days, and a renewal that finally goes through is paid on the schedule of whichever fiscal month it settles in. RevenueCat once worked through an example in which a December 26 purchase took 40 days to collect and reached the developer 96 days after the purchase date.

Google Play keeps things simpler: everything processed in a calendar month is paid around the 15th of the following month, and the transfer can take up to five more business days to settle, so the wait runs from about two weeks to about seven. A checkout you run yourself on Stripe is a different animal. In the US, standard payouts arrive two business days after the charge, roughly three calendar days once weekends are counted.

I Simulated the Gap

Fees are easy to compare on a napkin. Timing isn't, because it compounds with growth. So here is a deliberately small model. The app spends $200 a day on ads, acquires subscribers at $25 each, charges $9.99 a month and loses 8% of its subscribers at every renewal. On the App Store path it pays the Small Business Program's 15% commission and gets paid on Apple's calendar, starting on the first day of a five-week fiscal month and using the typical 33-day lag rather than the full 45 days Apple reserves. On the web path it pays Stripe's 2.9% plus 30 cents and gets paid three days later. Taxes and refunds are ignored, and both paths are assumed to convert equally, which they won't.

// cash-valley.js: how deep does your bank balance dip before payouts catch up?
// Run with: node cash-valley.js

const DAYS = 365;

// Apple: fiscal months of 5, 4 and 4 weeks; the money lands ~33 days after one closes.
function appleMonthEnd(day) {
  const lengths = [35, 28, 28];
  let end = 0;
  for (let i = 0; end <= day; i++) end += lengths[i % 3];
  return end;
}

const channels = {
  appStore: { net: (p) => p * 0.85, payoutDay: (d) => appleMonthEnd(d) + 33 },
  webCheckout: { net: (p) => p * 0.971 - 0.3, payoutDay: (d) => d + 3 },
};

function simulate({ net, payoutDay }, { adSpend = 200, cac = 25, price = 9.99, churn = 0.08 } = {}) {
  const payouts = new Float64Array(DAYS + 120);
  let cash = 0, valley = 0, valleyDay = 0, lastNegativeDay = 0;
  for (let d = 0; d < DAYS; d++) {
    // each daily cohort pays on day 0 and every 30 days after, minus churn
    for (let start = d % 30; start <= d; start += 30) {
      const subscribers = (adSpend / cac) * (1 - churn) ** ((d - start) / 30);
      payouts[payoutDay(d)] += subscribers * net(price);
    }
    cash += payouts[d] - adSpend;
    if (cash < valley) { valley = cash; valleyDay = d; }
    if (cash < 0) lastNegativeDay = d;
  }
  return { valley: Math.round(valley), valleyDay, positiveForGood: lastNegativeDay + 1, day365: Math.round(cash) };
}

for (const [name, channel] of Object.entries(channels)) {
  console.log(name.padEnd(12), simulate(channel));
}
Enter fullscreen mode Exit fullscreen mode

Both versions collect the same $145,000 from customers over the year. The App Store version bottoms out almost $20,000 in the hole on day 158 and doesn't stay above zero until day 278. The web version bottoms out about $6,000 in the hole on day 62 and stays positive from day 153.

The obvious suspect is the commission, so I swapped the pieces. Apple's 15% with web-style timing digs a hole of about $7,000. Stripe's fee on Apple's calendar digs one of about $18,700. Even a hypothetical zero-commission App Store, paying on the same calendar, would need about $17,900 of cash before the app could fund itself, three times what the web path needs. The commission mostly shows up in your profit. The calendar shows up in how much money you need to survive long enough to earn it. At the end of the simulated year, about $30,000 that customers had already paid was still on its way from Apple, against less than $2,000 on the web path.

Why the Hole Matters More Than the Fee

Double the ad budget and the hole doubles too, to roughly $40,000. That isn't a quirk of the code. It's one of the oldest lessons in small-business finance. In Harvard Business Review's classic analysis of how fast a company can afford to grow, Neil Churchill and John Mullins showed that a profitable company can still run out of money by growing faster than its own cash cycle can fund. Their self-financeable growth rate depends on how long cash stays tied up in each operating cycle, how much cash every dollar of sales consumes and how much it gives back. When the market would let you grow faster than that, they point to three levers: speed up the cash flow, cut costs or raise prices.

For an app, the payout calendar is a piece of the operating cycle that nobody can negotiate. Apple won't pay early because a cohort is performing well. That leaves you covering the gap with savings, a credit line, a financing deal that charges you for exactly this wait, or a slower acquisition ramp that lets the money catch up. None of these is a disaster. It's still better to pick one on purpose than to discover it while staring at your bank app.

The US Link-Out Window Is Open, for Now

This is why the legal fight over external purchase links matters to anyone with a subscription app, not just to Epic. In April 2025, Judge Yvonne Gonzalez Rogers found Apple in contempt of her 2021 anti-steering injunction and barred it from charging any commission on purchases US users make through links out of an app. Spotify submitted an update the next day and had it approved the day after that. In December 2025, the Ninth Circuit upheld the contempt finding but ruled that a total ban went too far, allowing a commission limited to the costs genuinely and reasonably necessary for coordinating those links, and sent the rate back to the district court. Apple has since proposed up to 15% for standard apps, with lower tiers for qualifying programs, and Epic is fighting it. Meanwhile, the Supreme Court agreed in June to decide whether Apple can be held in contempt for violating the spirit of an injunction rather than its text. Oral argument is set for January 11, 2027, and the justices aren't ruling on the fee itself.

As of mid-November 2026, US link-out purchases still carry no Apple commission. On cash timing, that makes the web path close to a free lunch. On fees, it's tighter than it looks. Stripe's fixed 30 cents makes the card fee about 5.9% of a $9.99 charge, international cards add 1.5% and currency conversion another 1%, and every dispute costs $15. Sales tax becomes your problem unless you pay a merchant of record to absorb it. On a $2.99 purchase, the card fee alone is almost 13%, which leaves little room for the conversions you lose by sending someone to a browser. Spotify later said its update drove a significant increase in free-to-Premium conversions, but Spotify is a brand people already trust with a card. Your paywall should earn that conclusion with its own A/B test.

Build It So You Can Change Your Mind in a Day

With a fee that could move twice in 2027, the worst architecture is one that hardcodes today's answer. A few patterns keep the decision cheap:

  • One entitlement record, several payment rails. App Store Server Notifications and Stripe webhooks should update the same subscription table, and the app should ask your server what a user owns instead of trusting whatever the device can see.
  • Ship both purchase paths and choose on the server. Link-out rules differ by country and the US rules may shift after January, so the choice should live in config keyed by storefront, not in a binary waiting on review.
  • Store an expected payout date on every transaction. Channel, net amount, fiscal month and expected deposit date turn your cash forecast into a query instead of a monthly spreadsheet ritual.
  • Reconcile against the stores' own reports. Apple publishes financial reports for the previous fiscal month by the first Friday of the current one, and Google settles by calendar month. Compare real deposits with expected ones and alert on drift.
  • Key every ledger write on the provider's transaction ID. Webhooks retry and notifications can arrive twice, so idempotency is what stops a retry from turning into a double credit.

Measure Days From Tap to Bank

We already put p95 latency and crash-free sessions on dashboards. Add one more line: days from tap to bank, per payment channel, weighted by revenue. It tells you how much growth your current cash can fund, what a financing gap will cost and what a change in payment routing is really worth. Plug your own CAC, price and churn into the script before deciding anything, because the shape of the curve matters more than my numbers. Revenue is the request. Cash is the response, and only the response pays for your next experiment.

Top comments (0)