Key takeaways- Installs are cheap to count and easy to inflate. Bookings, purchases and paid sign-ups are the numbers that show whether the app earns money.
- Write down the three to five in-app events that equal revenue before anyone opens Firebase, and name them consistently on Android and iOS.
- When the buyer leaves for a UPI app or gateway page, confirm the payment on your server and log the purchase from there, not from a screen the user may never return to.
- On iOS especially, expect partial data. Plan reports around what you can see: totals, trends and funnel steps.
Your app has thousands of installs and a monthly report that celebrates them. Yet bookings are flat and nobody can say which campaign produced the paying users. An install means someone downloaded the app. It does not mean they registered, paid, booked a slot or came back.
This guide shows how to define the events that equal revenue, set them up in Firebase and GA4, handle the UPI and gateway hand-off, connect results to Google Ads, and review them monthly. It is written for business owners and product managers working with a developer, so each section ends with something you can ask for or check.
Why install counts mislead
Installs are the first step of a long funnel, and most of the funnel is invisible if you only count them. Three things make them unreliable as a success measure.
- Cheap installs are easy to buy. A campaign can optimize for the lowest cost per install and fill the report with people who open the app once.
- Installs say nothing about intent. A user who installs for a discount code and one who books a service look identical in the store console.
- Revenue often happens outside the app. In India the final step is frequently a hand-off to a UPI app, a bank page or a gateway screen. If the app never learns that the payment worked, the sale is not in your data.
So the question changes from “how many installs?” to “how many people completed the action that pays us, and where did they come from?”
Defining the in-app events that equal revenue
Before any setup, list your funnel in plain words. Here is an example for a salon booking app and a shopping app.
| Stage | Booking app event | Shopping app event |
|---|
| Interest | view_item (service page) | view_item |
| Intent | begin_checkout (slot chosen) | add_to_cart, begin_checkout |
| Payment started | add_payment_info | add_payment_info |
| Revenue | purchase (booking paid or confirmed) | purchase (order paid or COD confirmed) |
| Retention | repeat booking within a set period | second purchase |
Decide for each one what counts. Is a booking a revenue event when it is paid, or when it is confirmed with a pay-at-venue option? Is a COD order a purchase when it is placed or when the courier delivers it? Write the rule down. Different answers for Android and iOS produce reports that cannot be compared.
A checklist for naming events
- Use Google’s recommended event names where one exists (for example purchase, begin_checkout, generate_lead, sign_up), because GA4 reports and Google Ads recognize them. Invent a custom name only when nothing fits.
- Send value, currency (INR) and a transaction or booking ID with every purchase.
- Keep the same names and parameters on Android, iOS and the website.
- Mark only the revenue and lead events as key events in GA4; marking everything dilutes the reports.
Setting up Firebase and GA4 events correctly
Google Analytics for apps runs through Firebase. The developer adds the Firebase SDK, which starts collecting a set of automatic events such as first open and session start. Your own events are added in code. The setup has a few steps worth checking.
- Create or confirm the Firebase project and link it to a GA4 property. Check that the account belongs to the business, not to a freelancer who may leave.
- Add the Android and iOS apps to the same project so reports can be combined.
- Log the revenue events from the agreed list, with parameters. See Google’s Analytics developer documentation for event and parameter formats.
- Test with DebugView in Firebase or GA4, using a real test build, and confirm each event arrives once with the right value.
- Mark the revenue events as key events and set the reporting currency to INR.
- Add the website to the same property if you have one, so web and app journeys can be compared. Our GA4 and Google Tag Manager setup service covers the website side.
If you sell digital items through Google Play billing or Apple’s in-app purchase, check what Firebase records automatically for those and compare it with your own payment records before trusting it. Physical goods, services and bookings paid through a gateway are not store purchases, so you must log those events yourself.
Tracking hand-offs to UPI apps and payment gateways
This is the weak point in many Indian apps. When the user taps pay, your app may open a UPI app (PhonePe, Google Pay and others) or a gateway page, and the user returns afterwards, or does not. If you log the purchase when the user taps Pay, you count failed payments. If you log it only when the user returns to a success screen, you miss those who paid and closed the app.
The reliable pattern is:
- Log begin_checkout and add_payment_info in the app, when those steps happen.
- Let the payment gateway notify your server of the result through its webhook (Razorpay, PayU and Cashfree all offer this).
- Mark the order as paid on the server, and log the purchase from there, or let the app log it only after it has fetched the confirmed status from your server. Sending from the server uses the GA4 Measurement Protocol, which needs the app instance ID, so have the app send that ID to your server with each order.
- Use the order or booking ID as the transaction ID so the same sale is never counted twice.
- Handle the return: when the app reopens, it should check the order status and show the right screen, including “payment pending”.
This is part of payment integration, not just analytics. Our payment and API integration work covers the webhook side, and the app maintenance and support team can audit an existing app for these gaps.
Connecting app results with Google Ads and wider reports
Once the events are reliable, connect them to where you spend money.
- Link GA4 to Google Ads and import the key events as conversions, or link Firebase directly, depending on how your account is set up.
- For app campaigns, choose a goal based on in-app actions rather than installs, once you have enough conversions recorded for the campaign to learn from. Start with an event further up the funnel if purchases are rare, then move to purchase.
- Use one source of truth. If the same sale is imported from GA4 and also tracked by a separate SDK, it will be counted twice. Check the conversion list for duplicates, as covered in our post on duplicate conversions.
- Remember the store listing. Many organic installs come from search inside Play and the App Store. Better app store optimization improves them, and the same events tell you whether those users convert.
Because one GA4 property can hold both app and web data streams, key events from the app and the website appear side by side in the same reports. Attribution across devices still depends on signed-in or consented users, so treat cross-channel comparisons as directional. Google changes how app conversions appear in its reports from time to time, so read the current notes in Analytics Help and check your own property before comparing channels.
Privacy limits on iOS and Android, and what you can still see
You will not get a complete, user-level picture, especially on iPhones.
- On iOS, the App Tracking Transparency prompt means many users do not allow cross-app tracking. For those users, ad networks rely on Apple’s aggregated attribution frameworks (SKAdNetwork and its successor AdAttributionKit), which report installs and a limited set of post-install signals with a delay and no user-level detail.
- On Android, users can reset or delete their advertising ID, and on Android 13 and later an app must declare the AD_ID permission to read it at all. Check that your app and its SDKs handle a missing ID.
- In both cases, you can still see how many purchases happened, which funnel steps lose people, and which campaigns correlate with more revenue at an aggregate level.
Plan reports around totals, rates and trends instead of one customer’s journey. Have your developer review the app against current Apple privacy requirements and the data safety section of Google Play, and make sure the consent wording matches what the app actually collects.
A monthly app performance review
- List installs, first opens and registrations, and calculate the share that reaches each next step.
- Report purchases or bookings and their value, checked against your payment gateway records for the same dates.
- Calculate cost per booking or purchase by campaign, not only cost per install.
- Look at day 7 and day 30 return rates for users who completed a purchase against those who did not. If both are low, the problem is retention rather than acquisition; our guide on why users uninstall apps covers the usual causes.
- Review crashes and slow screens on the checkout path. A payment flow that fails on common phones will look like a marketing problem.
- Note any app or SDK release, because tracking breaks quietly after updates, and re-test the revenue events.
Frequently asked questions
How do I track purchases in my mobile app?
Log a purchase event with value, currency and a transaction ID using Firebase or GA4 after the payment is confirmed. For gateway or UPI payments, confirm the result on your server through the gateway webhook, then log the purchase from there. Test with a real build and DebugView before relying on the numbers.
What are the most important events for app conversion tracking?
Start with the events that equal revenue: purchase or booking confirmed, plus sign-up or generate_lead where relevant. Add steps that show where users drop off, such as add_to_cart, begin_checkout and add_payment_info. Keep names identical across Android, iOS and the website so reports can be compared.
Why do my installs not match my sales?
Installs count downloads, while sales count completed payments. Low-cost campaigns bring many people who never register, and some buyers pay through a UPI app and never return to your success screen. Tracking server-confirmed purchases and comparing them with gateway records reveals the real figures.
Can I track iOS app conversions despite privacy restrictions?
Yes, but with less detail. You can still count purchases, funnel steps and trends in your own analytics. Ad attribution on iOS is aggregated and delayed because many users decline tracking, so judge campaigns by totals and cost per purchase rather than by individual user paths.
Measure bookings and purchases, not just installs
We can audit your app events, payment hand-offs and Google Ads links, and fix what is missing.
Request an app tracking audit →