Case Study: Publicis Groupe achieves a centralized, scalable data platform and cuts model run times from hours to minutes with Qubole

Publicis Groupe - Customer Case Study

Publicis Media, formed in 2016 from a merger of multiple agency teams, needed a centralized data and analytics platform to "democratize data" and eliminate the operational overhead of managing disparate, bespoke infrastructure (Redshift, on‑prem servers and legacy tools). The challenge was to provide agency teams easy, secure access to large, varied data sets so they could focus on analytics rather than data ops.

The team built an AWS‑based global data lake and ingestion framework and adopted Qubole to unify processing (Spark, Presto), querying (Tableau/SQL/Python/R/Scala) and access control. The solution handles nearly a petabyte of data, delivers multi‑region low‑latency access, automatically scales compute (cutting model run times from ~6 hours to minutes), supports ~10,000 queries/month, and lets a small admin team support hundreds of users while meeting security and compliance needs.

Publicis Groupe

Darren Smith

Engineering and Data Teams Leader