Analytics Blog

The End of the Analytics Black Box

Havas Media is one of the largest media companies in the world in over 126 markets and generating over $2 Billion in annual revenue last year.  Sylvain Le Borgne, EVP Data Platforms, shared with us the secret to their success was creating a data driven culture by moving from strategic analytics to tactical. Specifically, Havas Media had always been a data driven culture, but it was only in the last few years that they embedded analytics into every aspect of their client services operations. Havas Media accomplished this by creating an open data platform called Artemis with built in collaborative, visual, and self service functionality.

black box, analytics

At the core, Havas takes the huge volumes of media data streaming in from digital and traditional channels and makes it accessible without the need for coding, scripting, or IT requests. Havas has taken the advanced analytics out of the black box with Alpine and is exposing the intricate details of attribution, customer segmentation, lifetime value, and all of their library of algorithms to the client. By doing so, clients can sit with the account managers and data science teams to make changes to campaigns or programs on the fly.

Not only does Havas make the analysis transparent, the platform is also scalable as Alpine is web-based and compatible with Hadoop and Big Data. Havas takes advantage of the collaboration interface to ensure all clients consistently experience a high level of service they have come to depend on. As clients continue to demand more from their data, there is a need to be able to drill into the analysis using advanced exploratory analysis visually and with statistically backed findings. Alpine has made it dead simple to conduct exploratory analysis to find hidden segments, key signals in the noise and correlations in the data and then visualize with Tableau. For example, you may have a distribution of data were it is somewhat clear there are natural clusters of data (Figure 1). With Alpine and Tableau, you can quickly create a clustering algorithm to visualize the inherent clusters in the data (Figure 2).

Screen Shot 2013-12-08 at 4.53.55 PM Screen Shot 2013-12-08 at 4.55.18 PM

For more information about this use case, read here the full case study, or if you’d like to start collaborating and leveraging data to drive your business, click here for a personalized demo!