Which mobile tracking system is better to use when building end-to-end analytics through GCP

Among the paid systems, the most popular are Appsflyer and Adjust (Amplitude is still quite common, but it is more suitable for product applications).

Among the free \ shareware - Firebase. There is also AppMetrica from Yandex and MyTracker from Mail.ru Group, but we do not take them into account.

With the approach of building end-to-end analytics through BigQuery, mobile application tracking is usually used mainly to collect information. Therefore, it doesn’t matter what report algorithms and what interface the service uses, since all reports will be built on the BigQuery side, apply the source data, their algorithms, and then be displayed as their reports in Google Data Studio or another data visualization system. This approach involves reducing the budget to the required minimum for the use of tracking systems.

We recommend Firebase as the main tracking system because:
  • chargeless
  • has a native connector to BigQuery
  • allows you to create deeplinks, send push notifications, do A/B testing, etc.
  • when used in conjunction with BigQuery, sampling is avoided (that is, we get all data on users in BigQuery, and not averages).

Firebase works quite reliably with an audience of up to 5 million users, it is used by most banks in Ukraine.

The disadvantage of Firebase when tracking is that the platform does not support certain traffic channels (for example, Facebook or Apple Search after the last update).

Traffic channels are very important, especially for building marketing reports.
And there are two options for solving the problem:

1) If the budget is limited, then you need to separately connect all traffic sources in BigQuery in order to correctly identify traffic channels.

2) Ideally, additionally connect Adjust (or Appsflyer). They have a good system of integration with traffic channels, which is one of the main advantages. For this reason, systems take budgets for the amount of traffic (for example: one installation of an application with traffic detection costs 10 cents).

We recommend Adjust as it has an easier way to integrate a new traffic channel.

Adjust, for example, can cover those traffic channels that failed to connect to BigQuery, or it can track all traffic channels in general to simplify data collection. Based on the decision on how many traffic channels to track through Adjust, the next step is to connect the necessary events.

We recommend Firebase as the main tracking system, Adjust - for certain events to determine traffic.

Vadim Smolkin, web-analyst
Made on