Every iOS founder hits the same fork in the road right before launch: should you sell your app once, sell things inside it, or charge every month? The “in app purchase vs subscription” debate gets argued endlessly on Reddit, but most threads mix up two things that are not actually opposites. A subscription is a type of in-app purchase. The real question is which type of in-app purchase matches the value your app delivers over time.
This guide is a decision framework, not a list of definitions. We cover the four StoreKit product types, how Apple’s commission tiers actually hit your net revenue in 2026, which app categories perform best under each model, and the concrete signals that tell you it is time to switch.
First, clear up the vocabulary
Apple offers four in-app purchase product types. Confusing them is the number one reason founders build the wrong paywall.
| Product type | What it does | Restorable? | Typical example |
|---|---|---|---|
| Consumable | Used up, then bought again | No | Coin packs, credits, extra lives, AI generation tokens |
| Non-consumable | Bought once, owned forever | Yes | Pro unlock, remove ads, a filter pack, an extra level world |
| Auto-renewable subscription | Recurring access, renews until cancelled | Yes | Streaming, cloud sync, fitness plans, premium tiers |
| Non-renewing subscription | Fixed-period access, no auto renewal | You handle it | A one-season sports pass, a 6-month course |
So when people say “in-app purchase vs subscription”, they usually mean one-time purchases (consumable or non-consumable) vs auto-renewable subscriptions. That is the comparison we run below.
Side-by-side: one-time IAP vs consumables vs subscriptions
| Criteria | One-time unlock (non-consumable) | Consumables | Auto-renewable subscription |
|---|---|---|---|
| Revenue predictability | Low. Revenue tracks new installs only | Medium, but volatile and whale-dependent | High. MRR compounds, churn is measurable |
| Lifetime value ceiling | Capped at the price you set | Effectively uncapped for top spenders | High, grows with retention |
| Conversion friction | Low. Users like owning things | Very low per transaction | Higher. Commitment anxiety, cancel-first behaviour |
| Ongoing cost coverage | Poor. Servers and AI calls keep billing you | Excellent. Cost scales with usage | Good, if pricing matches heavy users |
| Refund and chargeback exposure | Low | Medium (family sharing, accidental purchases) | Medium to high, especially on trials |
| Apple commission | 15% or 30% | 15% or 30% | 15% or 30% in year one, 15% after 12 months of paid service |
| Support burden | Low | Medium (balance disputes) | High (billing, cancellations, restores) |
| Best fit | Tools with a finished feature set | Games, credit-based AI apps | Apps delivering fresh value weekly |
Apple’s commission tiers in 2026: what actually lands in your bank account
The “Apple takes 30%” line you see in every People Also Ask box is only true for a minority of developers today. Here is the accurate picture:
- App Store Small Business Program: if you earned up to $1 million in proceeds in the previous calendar year, you pay 15% on all in-app purchases, subscriptions included. Most indie and early stage studios sit here.
- Standard rate: 30% once you cross the $1M threshold.
- Subscription loyalty discount: after a subscriber has been paying for more than 12 consecutive months, the commission on that subscriber drops to 15%, even on the standard tier. This is a structural advantage that one-time purchases never get.
- Regional variations: in the EU, the Digital Markets Act business terms let you use alternative distribution and payment options with a different fee structure. In the United States, developers can link out to external purchase flows following the 2025 court ruling. Both routes change the math, but they also add compliance, tax and support work. Check the current App Store business terms for your storefronts before you model anything.
The math founders forget
A $49.99 one-time unlock at 15% nets you about $42.49 once. A $4.99 monthly subscription nets about $4.24 per month. You need roughly 10 months of retention to match the one-time sale, and after month 12 the same subscriber costs you less commission. If your median subscriber churns in month 3, the one-time purchase was the better business.
Rule of thumb: subscriptions win only if you can hold a meaningful share of payers past the 6 to 12 month mark. Otherwise you built recurring billing infrastructure to collect one-time money.
The decision framework: four questions
- Does your app deliver new value after the purchase? Fresh content, updated data, cloud sync, server processing, community, coaching. If value is delivered continuously, a subscription is defensible. If the user got everything on day one, charging monthly feels like rent on a tool they already own.
- Do your costs scale with usage? AI inference, storage, streaming bandwidth, live APIs. Variable costs demand either credits (consumables) or a subscription with usage limits. A one-time price on an AI app is a slow bleed.
- How often does a typical user open the app? Daily or weekly usage supports subscriptions. Monthly or seasonal usage almost always churns out. Occasional-use utilities monetize far better as one-time unlocks.
- Is the value emotional and bursty, or steady? Games, dating boosts and creative surges reward consumables bought in the moment. Steady productivity value rewards subscriptions.
Score it quickly
| If you answered… | Pick this |
|---|---|
| Ongoing value + daily use + variable costs | Auto-renewable subscription (monthly + annual, annual pushed hard) |
| Fixed value + occasional use + near-zero marginal cost | One-time non-consumable unlock |
| Bursty value + wide spending range + usage costs | Consumables / credit packs |
| Ongoing value but heavy spenders exist too | Hybrid: subscription base + consumable top-ups |
Which app categories perform best under each model
Pick a one-time in-app purchase if you build…
- Single-purpose utilities: unit converters, scanner apps, barcode tools, timers, white noise apps.
- Premium local-only tools: a Mac or iPad app that runs entirely on device with no backend.
- Offline reference apps: field guides, chord libraries, recipe collections with a finished catalogue.
- Ad removal: “remove ads forever” is the cleanest non-consumable in existence.
- Pro feature packs: an export module, an advanced filter set, a themes bundle.
Why it wins here: users resent subscribing to something that never changes, and the App Store review pages of subscription-gated calculators are a graveyard of one-star ratings. apple.com goes into the numbers.
Pick consumables if you build…
- Free-to-play games: currency, boosters, cosmetics, energy refills.
- AI generation apps: image, video, voice or document credits where each request has a real GPU cost.
- Dating and social apps: super likes, profile boosts, spotlight placements.
- Marketplace or gig apps: promoted listings, priority placement.
Why it wins here: the top 1 to 5% of users will spend many times the price of any subscription tier you would dare to publish, and the small purchase feels like a decision, not a commitment.
Pick an auto-renewable subscription if you build…
- Health and fitness: training plans, nutrition tracking, coaching.
- Productivity with sync: notes, tasks, habit trackers, password tools.
- Content and media: courses, meditation libraries, news, audio.
- Data-dependent apps: weather, finance, sports, flight tracking, anything with a live feed.
- B2B and prosumer tools: invoicing, CRM companions, analytics dashboards.
Why it wins here: the backend cost is real, the content refreshes, and after 12 months Apple’s commission drops to 15% on those subscribers.
The hybrid model, and when it beats both
The strongest 2026 monetization stacks are rarely pure. Common combinations that work:
- Subscription + consumable top-ups: AI apps give 500 credits a month with the plan, then sell extra packs. Predictable base, uncapped ceiling.
- Free tier + subscription + lifetime option: the lifetime unlock captures subscription-averse users. Price it at roughly 3x your annual plan and cap the promise carefully.
- Consumables + optional battle pass: standard in games, increasingly used in social apps.
The caution: every extra product type multiplies your paywall testing surface and your support tickets. Do not launch a hybrid on day one. Launch one model, learn, then layer.
Signals it is time to switch models
Signals you should move from one-time purchases to subscriptions
- Your server or AI costs per active user are rising faster than new install revenue.
- Revenue is flat or declining while the install base grows, meaning you have monetized everyone once and have nothing left to sell.
- Users keep requesting features that require ongoing infrastructure (sync, sharing, backups, collaboration).
- You are shipping meaningful updates every month and getting paid nothing for them.
- Your team cannot forecast next quarter’s revenue within a reasonable range, which blocks hiring and paid acquisition.
How to switch without a revolt: grandfather existing buyers into a permanent free tier with everything they already paid for, and put only new, clearly-ongoing value behind the subscription. branch.io walks through the specifics.
Signals you should move from subscriptions to one-time or consumables
- Month 3 retention of payers is under roughly 40% and does not improve after onboarding fixes.
- Trial-to-paid conversion is fine, but cancellations spike immediately after the first renewal: users got their value in one session.
- Your reviews repeatedly say “great app, would buy it, will not subscribe”.
- Support time spent on billing and refunds exceeds time spent on product feedback.
- A small group of users hits your plan limits constantly while most never come close: that is a credits business wearing a subscription costume.
Signals your pricing, not your model, is the problem
Before you rebuild everything, check these. Many “model failures” are pricing failures:
- No annual plan, or an annual plan with a discount under 40% versus monthly.
- A paywall shown before the user has experienced any value.
- One single price for wildly different user segments.
- No win-back offer for lapsed subscribers and no promotional offers for churn risk.
A practical launch sequence for new iOS apps
- Weeks 1 to 4: ship free with one clear paid unlock or one subscription. Measure activation, not revenue.
- Weeks 4 to 12: instrument the funnel. Track paywall views, conversion by placement, trial start rate, and day 30 retention of payers.
- Month 3: test price points and plan structure before touching the model itself.
- Month 6: look at the retention curve of payers. If it flattens above a healthy floor, double down on subscriptions. If it collapses, move to one-time or credits.
- Month 12: layer a second product type only if the data points to an unserved segment (heavy users or subscription-averse users).
Frequently asked questions
Is an app free if it says “in-app purchases”?
Yes, the download itself is free. The “In-App Purchases” label on the App Store simply means the app offers paid content, features or subscriptions inside it. You can install and use the app without paying, though some apps gate most of their functionality behind a paywall.
Do I get charged automatically for in-app purchases?
No. Every in-app purchase requires explicit confirmation with Face ID, Touch ID or your Apple Account password. The exception is an auto-renewable subscription: you confirm once, then it renews on schedule until you cancel it in Settings, including at the end of a free trial.
Does Apple really take 30% of in-app purchases?
Not for most developers. Apple takes 15% under the App Store Small Business Program (up to $1M in annual proceeds) and 15% on subscribers who have been paying for more than 12 months. The 30% standard rate applies to larger developers on first-year subscriptions and other in-app purchases. Regional rules in the EU and the US can change this further.
Are subscriptions always more profitable than one-time purchases?
Only when retention supports them. A subscription beats a one-time unlock when the average payer stays long enough for cumulative net revenue to exceed the one-time price, typically past the 6 to 12 month mark. Below that, you are adding billing complexity for less money.
Can I offer both a subscription and a lifetime purchase?
Yes. Apple allows a non-consumable “lifetime” unlock alongside auto-renewable subscriptions. Be precise in your App Store description about what lifetime includes, since “lifetime” means the lifetime of the app or service, and unclear promises generate refunds and negative reviews.
What about consumables versus subscriptions for AI apps?
Most AI apps end up with both. A subscription covers the base cost and creates predictable MRR, while consumable credit packs let heavy users spend more without you having to price the plan for the extreme case. Start with whichever matches your cost curve, then add the other.
Key takeaway
The in-app purchase vs subscription question is really a question about time. If your app delivers value once, sell it once. If it delivers value continuously, charge continuously. If value comes in unpredictable bursts with real costs behind them, sell credits. Pick the model your product actually justifies, watch the payer retention curve for six months, and let the data tell you when to switch.