Many small apps end up with both ads and a subscription. Ads earn something from people who will never pay. The subscription earns more from the people who use the app most. The trouble is that each one can work against the other. Aggressive ads can push people to uninstall before they ever see the paywall, and a weak subscription offer can leave you depending on ad revenue you never planned around.
This post is a practical setup for running both without the conflict. It covers what the subscription should actually sell, how to wire entitlements so ads switch off reliably, where to place ads and upgrade prompts, the billing edge cases that cause support tickets, and how to measure the two revenue lines together.
Decide What the Subscription Actually Sells
Paying only to remove ads is a weak offer on its own terms. It turns the paywall into a ransom note: you made the experience worse, and now you charge to undo it. Some users will pay, but you also give everyone a reason to resent the ads.
A stronger structure is a subscription that sells more of the app, with no ads included as a side benefit. Ask yourself:
- What would your most engaged users pay for even if there were no ads at all? Think extra content, sync, export, advanced tools or higher limits.
- Can a free user get real, lasting value without paying? If not, the ads are carrying the whole free tier, and you will feel pressure to show more of them.
- Is there a feature that rewarded ads can unlock for a while, and the subscription unlocks permanently? This creates a natural bridge between the two models.
Write the answer down as one sentence, for example: Free: full core features with ads. Premium: no ads, plus cloud backup and unlimited projects. Every later decision should be checked against that sentence.
One Source of Truth for Entitlements
The most common hybrid bug is a paying user seeing an ad. It usually happens because several places in the code check premium status in their own ways, or because the check runs before the store has answered.
Put entitlement state in one place, cache the last known answer, and have every ad call ask that one place. A minimal Swift sketch:
struct EntitlementSnapshot: Codable {
let isPremium: Bool
let expiresAt: Date?
let checkedAt: Date
}
final class Entitlements {
static let shared = Entitlements()
private(set) var snapshot: EntitlementSnapshot? = loadCachedSnapshot()
// Short tolerance so a renewal that has not synced yet
// does not flash ads at a paying user.
private let renewalTolerance: TimeInterval = 3 * 24 * 3600
var adsAllowed: Bool {
guard let s = snapshot, s.isPremium else { return true }
if let exp = s.expiresAt,
exp.addingTimeInterval(renewalTolerance) < Date() {
return true
}
return false
}
func update(_ new: EntitlementSnapshot) {
snapshot = new
saveCachedSnapshot(new)
NotificationCenter.default.post(name: .entitlementsChanged, object: nil)
}
}The important parts:
- Cache the result. Launch is when the store is slowest to respond and when you are most tempted to load a banner. A cached snapshot gives you a sensible answer straight away.
- Add a short tolerance window. Renewals can arrive late. Briefly over-serving premium is cheaper than showing an ad to someone who just paid.
- Broadcast changes. When a purchase completes, visible banners should disappear right away, without a restart.
Do Not Initialize the Ad SDK for Subscribers
Hiding ad views is not enough. If adsAllowed is false, skip initializing the ad SDK and its mediation adapters altogether. That brings three benefits:
- Less network traffic and less startup work for your best users.
- Less data collection for people who have no ad relationship with you, which makes your privacy disclosures easier to keep honest.
- No reason to show an ad-related consent form or tracking prompt to someone who will never see an ad.
This affects the order of your consent flow. Check entitlements first, then decide whether ad consent is needed at all. The order of those steps is covered in more detail in Consent and ATT Prompts Done Right. With a hybrid model, the extra branch is simple: a premium user skips the ad-consent path, and a user who lapses enters it the next time ads would load.
One caveat: if your analytics or attribution SDKs also need consent, they still need their own flow. Keep the two concerns separate in code, even if they share a screen.
Place Ads So They Do Not Sabotage the Paywall
An ad placement can earn well and still cost you subscribers. A few rules help:
- No interstitial right after the paywall is dismissed. A user who just said not now and then gets a full-screen ad learns that declining is punished. Add a cooldown.
- Skip interstitials in the first sessions. New users are still deciding whether the app is worth keeping. Let them reach the core value first.
- Cap frequency by time, not by action count. Action counts punish your most active users, who are also your most likely subscribers.
- Keep banners out of the main workflow area. A banner near a primary button causes accidental taps, which annoy users and can cause policy problems with networks.
Keep these values in remote config so you can tune them without shipping a release:
{
"interstitial_min_seconds_between": 180,
"interstitial_skip_first_sessions": 2,
"no_interstitial_after_paywall_seconds": 120,
"rewarded_premium_unlock_hours": 24
}These numbers are only starting points for your own experiments, not recommendations. The right values depend on your session length and audience, so measure them as described below.
Use Rewarded Ads as a Bridge, Not a Substitute
Rewarded ads fit naturally into a hybrid model. A user can watch an ad to unlock a premium feature for a limited time, try it, and then decide whether to subscribe. The risk is that the rewarded path becomes so convenient that nobody needs to pay.
To keep the bridge working:
- Make rewarded unlocks temporary and scoped, for example one export or one day of a feature, rather than ongoing access.
- When the unlock expires, show a quiet message that the feature is part of Premium. Do not show a hard paywall straight away.
- Never require a rewarded ad for something that used to be free.
For more on reward design itself, see Rewarded Ads Design Patterns That Respect Your Players.
Where the Upgrade Prompt Belongs
Upgrade prompts work best when they match what the user is trying to do at that moment. Good places:
- A small Remove ads link next to the banner, which opens the full Premium offer rather than a remove-ads-only screen.
- The moment a user reaches a free-tier limit.
- Settings, where people who are already thinking about paying go looking.
Avoid adding an upgrade pitch after every interstitial. Pairing the two teaches users that the paywall and the annoyance go together.
Billing Edge Cases That Cause Support Tickets
Most hybrid support email comes from a small set of situations. Handle each one on purpose:
- Reinstall or new device. The cache is empty, so ads are allowed by default. Run a silent entitlement refresh at launch and delay the first ad load until it completes or times out. Keep a visible Restore Purchases button.
- Billing retry and grace periods. Both stores have states where payment has failed but access continues for a while. Treat those as premium, and show a gentle message about updating payment details.
- Refunds and expiry. When premium ends, turn ads back on from the next session, not in the middle of one. Starting the ad SDK partway through a session can be jarring and may also need a consent step.
- Family sharing or shared entitlements. If your store setup allows them, make sure your entitlement check counts them. Otherwise, a family member who has access will see ads.
Measure Both Revenue Lines Together
Judging ads and subscriptions separately leads to bad decisions. A change that raises ad revenue may quietly lower subscription conversion, and you would never see the trade-off.
Track these per install cohort, for example per week of install:
- Ad revenue per free user per day, from your ad network report divided by daily active free users.
- Conversion to paid within a fixed window after install, using the same window for every cohort.
- Retention at a few fixed days, split by free and paid users.
- Total revenue per install over the same window: ads plus subscriptions, net of store fees.
When you change an ad setting, compare total revenue per install between cohorts, or better, run a remote-config split test. If ad revenue goes up while conversion and retention go down, take the change seriously before celebrating. Give subscription effects enough time to show up, because trial conversions and renewals arrive later than ad revenue.
A Short Checklist
- The subscription offers real value beyond removing ads.
- One entitlement service with a cached snapshot and change notifications.
- The ad SDK is never initialized for premium users.
- The consent flow branches on entitlement status.
- An interstitial cooldown after paywall dismissal, plus a new-user grace period.
- Rewarded unlocks are temporary and scoped.
- Restore, grace period, refund and shared-entitlement cases have been tested.
- Revenue is reported per cohort as ads plus subscriptions combined.
Conclusion
Ads and subscriptions can work together if each has a clear job. Ads support the free tier, and the subscription sells a better version of the app. Most of the work is plain engineering: one entitlement source, no ad SDK for paying users, sensible cooldowns and careful handling of billing states. After that, judge every change by its effect on total revenue per install, not by either revenue line alone, and tune from your own data.