End-to-end analytics and infrastructure migration for WiseAdvice-IT
About the client
WiseAdvice-IT is one of Russia’s largest IT integrators. The company delivers large-scale, complex enterprise automation projects, IT support and corporate-system implementation.
The task
Build end-to-end analytics that brings together ad spend, web analytics and sales, and give advertising and SEO specialists convenient dashboards with the ability to view results in different attribution models. A separate task was migrating the entire analytics infrastructure to a new stack after the company's restructuring.
Goals
- →Collect ad spend, web analytics and sales into a single warehouse
- →Stitch user visits to sales for end-to-end analytics
- →Build dashboards for advertising and SEO specialists
- →Allow the attribution model to be selected right inside the dashboard
What was built
- 01
End-to-end analytics on Google BigQuery
Combining data from ad accounts, Google Analytics and sales from the accounting system into unified data marts: the path from ad spend to sales.
- 02
Session stitching by Google Client ID
The Google Client ID is captured from the site in the accounting system, and in BigQuery user visits and sessions are linked to sales by that ID.
- 03
Export via Apache Airflow
Automated export of data from ad accounts into BigQuery, orchestrated with Airflow.
- 04
Selectable attribution models
Several attribution models implemented in the data marts, with the model switchable right inside the dashboard.
- 05
Dashboards in Google Data Studio
Reporting for advertising and SEO specialists.
- 06
Cost optimization
The BigQuery data marts were designed to minimise storage and processing costs.
- 07
Infrastructure migration
After the company's restructuring, all analytics was migrated to a new accounting system, a new Google Cloud account and a new Airflow server, with the dashboards rebuilt for the changed structure; delivered on a tight timeline in close cooperation with the marketing team.
Result
The company gained full-fledged end-to-end analytics with a flexible choice of attribution models and a cost-efficient infrastructure on Google BigQuery. Despite the large-scale restructuring, the analytics was fully migrated to the new stack without losing continuity of reporting.