An install is a distribution milestone. It does not prove that a person understood the product, reached its main benefit, or is likely to pay. If the app earns value only after a user completes a workflow, optimizing acquisition around the install can reward the wrong outcome: more people entering the funnel, without evidence that they progress through it.
Google App campaigns support objectives for installs, in-app actions, and conversion value. The practical work is to define a reliable event that represents progress, connect it to the campaign, and keep judging users by what happens after the platform records that event.
Choose the event that marks a real product win
Do not start with the event that is easiest to fire. Start with the user action that makes the product’s value concrete, then check whether it happens often enough and soon enough to guide acquisition. The ladder below separates acquisition milestones from evidence of value.
| Event layer | Example | What it tells you | Common trap |
|---|---|---|---|
| Install / first open | The app launched after download | The campaign acquired a reachable user | Calling the install a success when setup never finishes |
| Setup completed | Account or workspace is ready | The user cleared initial friction | Optimizing for account creation even when accounts stay empty |
| Activation | The user completed a core workflow for the first time | The product delivered its first intended value | Counting a screen view instead of a completed result |
| Commercial event | Trial started, appointment booked, or first purchase completed | The user crossed a business-relevant threshold | Choosing an event so rare that the campaign has little feedback |
Score a candidate event before making it the goal
- Meaning: Can product and finance explain why this action predicts a useful customer, without referring to the ad platform’s default setup?
- Specificity: Does the event fire only after the meaningful outcome is complete? A button tap or screen view is often intent, not completion.
- Reliability: Is it recorded once, with consistent names and parameters, across the versions and operating systems the campaign targets?
- Learning readiness: Does it happen frequently enough and quickly enough for the campaign to receive feedback? A high-value event with very sparse volume may be better as a downstream quality check at first.
- Predictive value: In historical cohorts, do users who reach it retain, purchase, or generate contribution margin at a better rate than users who do not?
This is a product decision before it is a bidding decision. A subscription app might define activation as completing a first guided session and saving a plan. A marketplace might choose a first qualified booking rather than a profile completion. The event name should capture the outcome; its parameters should preserve context needed for analysis. A clear event tracking plan helps teams keep those definitions stable.
Connect event quality to the amount you can afford
When you optimize for an in-app action, the target cost should reflect the expected economics of users who complete that action. One planning estimate is:
Allowable cost per activated user ≈ contribution margin per paying user × activated-to-paid conversion rate × acquisition share.
For example, if a paying customer contributes $120 over the chosen horizon, 10% of activated users become paying customers, and the business allows half of that expected contribution for acquisition, the illustrative ceiling is $6 per activated user ($120 × 10% × 50%). The inputs are assumptions to validate by cohort, not universal benchmarks. Include refunds, platform fees, variable service costs, and the time needed to earn the margin. For customer-acquisition definitions, see the guide to what belongs in CAC.
Wire the event before increasing spend
- In Firebase or a supported app attribution partner, define the event at the point the user completes the outcome—not when the flow starts.
- Link the app analytics property and Google Ads account, then select the intended event as a conversion action. Google documents both importing Analytics events and tracking custom app actions.
- Test the event on a production-like build for each supported platform. Check that one completed action creates one event, event values are correct where used, and test traffic is excluded from the live performance readout.
- Keep install, activation, purchase, and retention events available for diagnostics. Choose one primary optimization event for the first focused campaign so the team can explain what the bidding system is being asked to find.
- Align ad assets and the store listing with the action. If the campaign promises a feature, the first session should make that feature easy to reach.
Google Ads describes bidding options that include maximizing conversions for installs, in-app actions, or conversion value. Firebase documentation explains that linked Analytics events can be imported into Ads as conversion actions. The available options and requirements can change, so verify the current setup in the account before building the campaign.
Read performance on two clocks
Platform reporting answers whether the campaign is finding users likely to complete its selected event under the platform’s measurement rules. Product and finance reporting answer whether those users become durable, profitable customers. Review both.
| Acquisition view | Business-quality view |
|---|---|
| Cost per selected in-app action | Cost per activated user using the product’s own event definition |
| Action volume by campaign and asset | Activation-to-paid conversion by install cohort |
| Campaign conversion value, when configured | Day-7 / day-30 retention and contribution margin after variable costs |
| Conversion timing and reporting lag | Incrementality or holdout evidence when volume supports a clean test |
Let cohorts mature before comparing them. An install can be recorded immediately; a meaningful action may occur days later. Changing the event, creative, and target cost at the same time also makes the result hard to interpret. Keep a dated record of each change and review the event’s predictive value again as the product or onboarding flow changes. The existing product activation rate guide can help define the downstream rate consistently.
When installs are still the right objective
Install optimization is appropriate when the app is new, a measurement path is not yet reliable, or a post-install event does not have enough clean observations to guide a campaign. Treat it as a deliberate learning stage: instrument the product, understand which cohorts activate, and change the acquisition objective when the event signal is trustworthy. Do not set a low-value event merely to give the campaign more conversions, and do not call an in-app-action campaign successful until activation translates into the customer outcomes the business needs.
Platform references
- Google Ads Help: About bidding in App campaigns
- Google Ads Help: Mobile app conversion tracking
- Firebase: Google Ads integration
How did this article land?
Choose one reaction. You can change it anytime.