Subscription App User Acquisition: Paying for Trials

Subscription app user acquisition: which event should you pay for?

For most subscription apps the best paid event is a trial start with a quality condition attached (for example, the trial is still set to renew 48 hours after it began), because it confirms within days and predicts revenue. Installs are cheaper but say little about who will pay. Trial-to-paid and first renewal prove real revenue, but they confirm only when the trial ends or a full billing period later, so fewer partners will run them and those who do want a higher price. Whatever you choose, price it from what a trial is worth to you: your trial-to-paid rate times the lifetime value of a paying subscriber.

Policy and benchmark details below were checked on October 21, 2026.

Install, trial start, trial-to-paid or first renewal

A subscription funnel has four natural events a partner can be paid for, and each one moves risk between you and the partner differently. CPI vs CPA for apps covers the general trade-off.

You do not have to pick one. A multi-event offer can pay a small amount on trial start and more on the first charge, so partners get fast feedback while most of your money follows real payers.

How much is a trial worth? The math

• Value per trial = trial-to-paid rate × LTV of a paying subscriber.
• Value per install = install-to-trial rate × value per trial.
• Maximum bid for any event = value of that event ÷ the return you need (1.3 if you want $1.30 back for every $1 spent).

Use payer LTV net of store commission, taxes and refunds, over a period you can wait for, such as 12 months. CAC vs CPA vs LTV shows how to build it, and the LTV calculator does it from your price and churn. Churn drives LTV more than anything else; see what is churn rate.

Value per trial for a few combinations (sample figures only):

The same trial can be worth $2 or $32, so a “typical cost per trial” from another app says nothing about what you can afford.

For scale only: RevenueCat’s State of Subscription Apps 2026, built on more than 115,000 apps and 2025 data, puts the median trial-to-paid rate at about 32% on both the App Store and Google Play, and the global median realized LTV per payer after year one at $23 ($32 for apps from North American developers). Medians from different apps do not multiply into a benchmark, so use your own rates.

How trial length and the paywall change what you can buy

Trial design changes both the value of a trial and how long a partner waits for results. From the same RevenueCat report:

• Longer trials convert better. Median trial-to-paid was 25.5% for trials of 4 days or less, 37.4% for 5 to 9 days and 42.5% for 17 to 32 days. Yet 46.5% of apps now use trials of 4 days or less.
• Short trials die on day 0. 55% of all 3-day trial cancellations happen on the day the trial starts, which is why a bare trial start overstates value.
• Trials start right away or not at all. In most categories 80% to 89% of trials start on day 0.
• Hard paywalls convert faster, not better in the long run. Median download-to-paid by day 35 was 10.7% for hard paywalls against 2.1% for freemium, with nearly identical retention after one year. For freemium apps, 23% of conversions came 6 or more weeks after download.

For paid acquisition this means: a 30-day trial raises value per trial but makes a first-charge event a month-long wait, so a qualified trial is usually the only practical paid event. A 3-day trial makes paying on the first charge realistic. A freemium app needs a long conversion window on any install-to-paid offer; what is a conversion window explains the trade-off. If you change the trial or paywall mid-campaign, recalculate your bids.

Store rules for trials and introductory offers

Both stores allow free trials and require clear terms before anyone subscribes. Checked on October 21, 2026; read the current pages before launch, and treat this as general information, not legal advice.

Apple App Store

• App Review Guideline 3.1.2 allows auto-renewable subscriptions in any category if they provide ongoing value and last at least seven days, says they may offer a free trial by providing the relevant information in App Store Connect, and says apps that trick users into a subscription under false pretenses will be removed. 3.1.2(c) asks you to describe clearly what the user gets for the price before asking them to subscribe.
• Apple’s subscriptions page lists three introductory offer types (free trial, pay as you go, pay up front), allows one introductory offer per subscription group per customer, and asks that a free trial purchase flow clearly show the trial length and the price billed after it.

Google Play

• Google Play’s Subscriptions policy says that before enrollment you must clearly describe the offer’s duration, pricing and content, how and when a free trial converts to a paid subscription, what it will cost, and how to cancel. It also requires an easy-to-use online way to cancel in your app.
• Google’s subscriptions developer guide says Play verifies a valid payment method before a free trial starts, and that a user who cancels during the trial keeps access until it ends and is not charged.

Partner copy counts as your marketing. If a partner calls your trial “free” without saying it renews into a paid plan, you have the same problem as a misleading paywall. Rewarded campaigns have their own store rules, covered in incentivized installs and Google Play policy.

Measuring trials with an MMP and store server notifications

The most reliable record of a trial comes from the store itself, sent to your server. Use three layers:

1. Attribution. Your mobile measurement partner ties the install to the partner and click. On iOS, ATT limits matching; see SKAdNetwork and ATT explained.
2. Store notifications. Apple’s App Store Server Notifications (version 2) include SUBSCRIBED (subtype INITIAL_BUY for a first purchase, with the offer type in the payload), DID_RENEW, DID_CHANGE_RENEWAL_STATUS (subtype AUTO_RENEW_DISABLED when renewal is switched off), EXPIRED and REFUND. Google Play’s Real-time developer notifications arrive through Cloud Pub/Sub with types such as SUBSCRIPTION_PURCHASED, SUBSCRIPTION_RENEWED, SUBSCRIPTION_CANCELED, SUBSCRIPTION_REVOKED and SUBSCRIPTION_EXPIRED, plus a voided purchase notification for refunds. Google notes these only say the state changed, so your server must call the Play Developer API for details.
3. Postbacks to partners. Your server, not the app, fires the paid event to the partner with their click ID, using server-to-server tracking.

Many MMPs also report subscription events by channel; Adjust’s Subscriptions help page, for example, describes analyzing conversions, renewals and cancellations per channel.

Building a qualified trial event

Define it as a first-time trial (Apple SUBSCRIBED with INITIAL_BUY, or Google SUBSCRIPTION_PURCHASED) with no renewal-off or cancel notification in the next 48 hours. Your server waits out the window, then sends one postback. Share the written definition with every partner before launch.

Refunds and cancellations as reversals

• Refunds and voided purchases of the first payment: reverse, within an agreed window such as the first billing period. RevenueCat found median refund rates of 3% to 4% in most categories (2.5% for Travel to 4.7% for Productivity), and 3.4% for North American developers against 7.7% for India and Southeast Asia.
• Fraud found later, such as duplicate accounts or emulators: reverse.
• Normal cancellation after the event you paid for: do not reverse. That risk belongs in your price or your event definition, not in a later clawback.

Send each reversal by postback with the original click ID and track reversal rates by source. Pay only for confirmed results explains why the confirmation must come from your server.

Trials and rewarded traffic: the risks to plan for

In rewarded (incentivized) traffic, users earn points or currency in another app or site for reaching a goal in yours; incentivized traffic explained covers when it works. For subscriptions, plan for these risks:

• Reward-then-cancel. A user who came for the reward has a reason to start a trial and switch off renewal straight away. Assume a lower trial-to-paid rate from rewarded sources until your own cohorts show otherwise.
• Payment method at the door. Google Play checks for a valid payment method before a trial starts, and an auto-renewing trial is a bigger commitment than the in-app goals rewarded users usually complete.
• Mixed cohorts. If rewarded and paid-social trials share one cohort, a weak source hides behind a strong one. Tag every source and sub ID.

The practical answer is to pay for a qualified trial or an in-app step that shows real use, such as finishing onboarding or a first lesson, and keep a separate value per trial for each source type. How to set a CPA payout for an offerwall campaign turns those values into payouts per step.

Judging sources by cohort ROAS at day 30, 90 and 180

Cohort ROAS is the revenue from users acquired in one period, up to day N after install, divided by what you spent on them. Commentary in RevenueCat’s report warns that freemium apps keep converting into week 6 and beyond, so ROAS pulled at day 7 cuts off revenue that has not landed yet. What each checkpoint shows depends on your plan:

Set a target for each checkpoint from your break-even ROAS, and use the ROAS calculator for the arithmetic. A source at 40% of break-even on day 30 may be on track for a monthly plan if renewals follow, but it is a problem for a yearly plan, where day 30 already holds most first-year revenue. Two more checks:

• Watch the first renewal. RevenueCat found that nearly half of monthly subscribers do not renew even once, so day 30 can flatter a monthly cohort.
• Test incrementality. Some trials would have happened anyway. A holdout by country or period shows what a source adds; see incrementality testing.

For web SaaS trials outside the stores, how to get more SaaS free trial sign-ups covers activation and cost per activated trial.

A launch checklist for a paid trial campaign

1. Calculate value per trial from your own trial-to-paid rate and 12-month payer LTV, by platform and country.
2. Pick the paid event (qualified trial, first charge, or both as steps) and write down when your server counts it.
3. Wire both stores’ server notifications to your back end, fire partner postbacks from there, and agree reversals for refunds and fraud.
4. Check that partner copy states the trial length, renewal price and how to cancel, then start with small caps and scale only sources tracking toward your day 90 and 180 targets.

Frequently asked questions