At a glance
| Field | Detail |
|---|---|
| What | Choose a trial or subscription event for app acquisition bidding. |
| Who it matters to | App buyers whose subscription journey includes a free trial. |
| Where | Google Ads App campaigns and TikTok app event optimization; iOS signal conditions require separate review. |
| When it applies | Before selecting or changing the event used for bidding. |
| Action | Check the event definition, available signal, and downstream subscription measurement together. |
App subscription optimization starts with deciding which action the campaign should pursue and which outcome will judge its value. Google recommends considering proxy events for a desired outcome, while TikTok lists Start Trial and Subscribe as distinct supported optimization events. Our suggested approach is to evaluate trial acquisition against subsequent subscription outcomes before choosing the bidding event. Google’s campaign guidance and TikTok’s event list provide the platform premises.
Define the subscription milestone first
TikTok defines Start Trial as starting a trial of an app’s services and Subscribe as subscribing to a channel or service. That Subscribe definition does not specify whether the event represents a first payment, renewal, or another subscription milestone. Check the actual trigger in your integration before treating its count as newly paying subscribers.
Adjust’s subscription documentation separates trials, activations, expirations, and cancellations. It also describes measuring free-trial signups and conversions to paying subscribers, with subscription revenue analysis covering renewals and cancellations by channel. Those distinctions support a measurement plan that follows the journey beyond the event selected for bidding. Adjust’s subscription documentation
Our interpretation: write down the business meaning and payment status of each event before comparing campaign costs. A label alone does not establish that the platform event and the subscription report describe the same outcome.
Compare the available approaches
The following checks combine platform documentation with our suggested interpretation; they are not universal eligibility rules.
| Decision or approach | Source-backed capability | Limitation | Suggested check |
|---|---|---|---|
| Google Ads: consider a proxy action | Google recommends considering events that act as proxies for the desired event. | A proxy records a different action from the final outcome. | Check whether trial acquisition corresponds to subsequent paying subscriptions in your reports. |
| Google Ads: optimize for in-app actions | Google recommends choosing an action completed by at least 10 different users per day in the campaign. | This is campaign-level guidance, not a cross-platform threshold. | Review the selected event’s user volume within the campaign. |
| TikTok: select Start Trial or Subscribe | TikTok lists both events as supported for app event optimization. | The event list does not establish the payment timing of your implementation. | Verify each trigger against the subscription lifecycle. |
| TikTok: use iOS dedicated campaigns | TikTok lists Start Trial and Subscription among supported events. | Its documented AEO eligibility condition uses a corresponding SKAN event at the TikTok App ID level. | Check iOS signal eligibility separately from event naming. |
The Google volume recommendation comes from its App campaign setup guide. TikTok’s iOS conditions are described in its bidding and optimization documentation.
Separate event choice from signal readiness
Google’s recommendation concerns different users completing the chosen action per campaign per day. TikTok’s cited iOS documentation says app event optimization becomes available after at least 1 corresponding SKAN event at the TikTok App ID level. These statements address different scopes and should not be treated as equivalent evidence of readiness.
For a trial-based app, our suggested check is to inspect trial and paid-subscription signals separately. Do not assume that availability of one event establishes availability or sufficient volume for the other. Account-specific availability beyond the documented conditions is not specified in the cited documentation.
TikTok also describes SKAdNetwork postback delays and privacy-withheld conversion values. Review those reporting conditions before interpreting an absent subscription signal as an absent business outcome.
Account for the time to payment
Google explains that conversion windows should reflect how long the selected action takes. Longer windows can capture more conversions but may include organic conversions that are less incremental; shorter windows may miss incremental conversions. Its Days to conversions segmentation and conversion-lag reporting can inform that review. Google’s conversion-window guidance
For trial-based acquisition, our suggested check is to compare the actual time to the intended paid milestone with the configured measurement window. Avoid assigning a standard trial duration or assuming all subscription products convert on the same schedule.
Google also recommends aligning conversion windows with third-party measurement and deciding whether to count the first occurrence or every occurrence. Review that setting against your intended distinction between acquiring a subscriber and recording subsequent subscription activity. Google’s event measurement guidance
What to review
- Check whether trial, first paid subscription, and renewal have distinct definitions in your event dictionary.
- Compare the selected Google Ads action’s campaign-level user volume with Google’s guidance.
- Check TikTok’s iOS event eligibility and SKAdNetwork reporting conditions separately.
- Review conversion windows and event-counting settings against the paid milestone you intend to measure.
- Compare trial acquisition with downstream subscription conversions, revenue, renewals, and cancellations where your measurement setup supports them.
Use a measurement data-path review to connect these definitions to the events and reports used for the buying decision.



