But further down the road you may want to perform more advanced tests when you try out new features. Testing involves creating installs and events that aren't "real" and would corrupt your data.
What is the best practice for performing these tests, then?
Is Creating a Test App a Good Idea?
There are two possible methods for you to choose from, which we will cover here.
1. Create a Test App
1.1 Android and iOS test apps
You can copy your Android app and create an identical app with a different package name on AppsFlyer's dashboard. This is the same method as you would use for publishing your app on out-of-store markets.
For iOS advertisers can create a test app on Apple. As opposed to Android, this test app shouldn't be defined as a separate app on AppsFlyer's dashboard. It can be tested using OneLink (details below).
- Keeping the "real" app's data clean from the tests.
- Requires creating a new app on Android or a test app on iOS, which costs valuable development time.
- Test app and real app may differ by mistake. If so then tests on the test app don't necessarily apply on the real app.
Advertisers may use a separate build for testing their iOS app. In this case, the bundle ID and/or the App ID prefix differ between the production app and the test app. When testing the debug iOS app with OneLink, you need to setup the "Test App Configuration" settings under OneLink's setup, as seen in the capture below.
When clicking on this OneLink with an iOS device Universal links should open up the test app. If the production app is also installed on the device it gets precedence over the test app.
2. Create a "Test" Media Source
The best practice for testing is by using your real apps and not any debug apps.
To do this you can simply perform all tests on your test devices (which must be white listed) under the media source "test" (or any other name signifying test).
The custom tracking link you'll prepare for these devices should have the "&pid=test" macro to achieve this.
- No set up is required. Just create your custom tracking link and start testing immediately!
- Can be used with BOTH android and iOS apps.
- Test results are valid since the tests are performed with your real apps.
- Installs and subsequent events from these devices are seen along the data from real campaigns.
Although the "test" source alters the attribution numbers during the period of time it is used please note that:
- It's totally negligible - tests tend to produce very low numbers of installs, which compared with real campaigns have very small impact on your averages.
- They "fade away" quickly - shortly after you conclude your tests there are no new installs or events performed by the "test" media source. Therefore they soon disappear from the dashboard anymore.