Building an analytics system for a cross-platform project for a VoD media service (Firebase, Google Analytics web + apps, BigQuery, Google Data Studio, and Power BI)

Details

Client:

VoD-media service – provides the ability to watch TV channels and movies through an application for smartphones, applications for SMART-TVs, a website.

 

Service:

We are setting up a unified analytics system for all touchpoints with the client.

 

A task:

Build an analytics system that will allow you to: analyze marketing communications, track user interactions with content, segment users for Remarketing, analyze the product, and improve it.

 

Result:

As a result of the work, a unified analytics system for data analysis was built from scratch (based on Firebase, Google Analytics, BigQuery, Google Data Studio, and Power BI tools).

 

The author of the case: Elena Berezovskaya, head of the web analytics department at the agency

 

Using the Firebase and Google Analytics web + app, data was collected and imported into BigQuery. Then, data from BigQuery is visualized in Google Data Studio, Power BI, or the report is built in the form of a table. 

 

Thanks to this, it became possible to:

 

  • Use app notifications for marketing purposes: stimulate additional sales, increase loyalty, and user engagement.
  • Optimize advertising campaigns for the main business activities.
  • Use information about user behavior to improve the product.
  • Track, analyze, and eliminate the errors in the application.

 

A little more detail about tools

pic_1.jpg

Google Analytics is a system for analyzing online data.

 

Firebase is a mobile platform that helps you quickly develop high-quality applications, analyze and optimize advertising campaigns, and grow your business. Has functionality for creating notifications, messages within applications, A/B testing, error analysis, etc.

 

BigQuery is a cloud-based database that allows you to store your customers’ data and easily integrates with other Google products such as Google Analytics, Firebase, and others.

Google Data Studio is a tool for data visualization, its advantages: ease of use, integration with Google products. And it’s free.

 

Power BI is a tool for business analysis that allows you to get, transform, and visualize data. More flexible and functional than Data Studio, but more complex and has several features when working with Google services. You have to pay for it.

 

Everything seems to be simple, but at the stage of implementation, there were many pitfalls that we had to face.

 

The problem

Implementation pitfalls can be divided into two categories: technical and communicational.

 

The technical features are related to the selected technologies, application architecture.

 

Communication features were associated with a large number of teams that participated in the project.

 

Solution

Among the technical nuances, we can distinguish two as the main ones:

 

  • All the most essential actions do not take place in the application. To register, connect packages, pay for the user’s tariff, you need to go either to the site or to the site pages that are integrated into the application. In both cases, a session break occurs, which made it challenging to analyze ad campaigns. To solve the problem, we had to modify the site and transfer additional data from the server at the time of registration. 
  • Not all smart TVs support Firebase and the standard Google Analytics implementation. Since at the time of the work on the project, there is still no official Measurement Protocol for the Google Analytics app + web, so we had to create a standard auxiliary resource for integrating data into a single system. 

 

Working on an application differs from working on a website not only at the technical level but also at the communication level. In our case, we needed to communicate with five teams:

 

  • TV application development team;
  • smartphone application development team;
  • website development team;
  • product team;
  • marketing team.

 

When working with multiple teams, it is essential to remember the following:

 

  • Each team has its area of ​​responsibility. And it is vital to stipulate who is responsible for what at the beginning to speed up the implementation process.
  • It is necessary to collect the requirements for the analytics system from all stakeholders in it. As a result, some of the provisions will overlap, and this will avoid unnecessary redundancy and duplication of data.
  • It is essential to speak with everyone in the same terms. As an example, the ID of the subscription option can be different for each department:
  • for a marketeer – the name of the package;
  • for the product manager – the package code;
  • for the developer – the ID in the database.

 

And for everyone to speak the same language, as a result, it is necessary to eliminate all inaccuracies and unambiguously prescribe the values and names of all events and parameters.

 

  • It is vital to have easy access to all tasks, preferably in one place. That way, you can avoid losing assignments during transfer between subdepartments. 
  • It is crucial to track the implementation of tasks in the process. That means that the person responsible for the result needs to control the job at intermediate stages, and not only at the end of the implementation.

 

The Result

As a result of the work, the analytics system for the service was built from scratch. During the assignment:

 

  • Terms of reference for the implementation of analytics have been created and implemented.
  • Joint meetings were held, at which all the requirements were defined and brought together, the nuances of implementation clarified.
  • Documentation has been created for the built analytics system.
  • The training was provided on the use of the system.

 

Since the beginning of the use of data in marketing, the number of installations has grown significantly. The main merit is the work of the promotion teams, but without the data obtained, some of the strategically crucial decisions would have been impossible to make.

The number of installations

 

The end-to-end analytics implementation phase

 

Data usage phase in marketing

 

 

Based on user behavior data, the product was improved, which led to an increase in the number of active users:

pic_3.png

 

Сonclusions

In order for the analytics system to work, all those interested in it must know its functionality and trust the data that is stored in it.

It is also important not to forget that technical implementation is influenced not only by the competence of specialists but also by properly built communication and understanding of business processes.