The worlds of Digital Analytics and Marketing Analytics have frequently led somewhat independent lives – with the Digital Analyst spending time looking at digital channels (web/mobile/social), reading out metrics, understanding conversion rates, focused on conversion funnels, A/B and multi-variate testing and the like while the Marketing Analyst was more concerned with Survey Analysis, developing “What-If” simulators for product features and concerning themselves with ROI from campaigns.

There has been an inevitability in the growth in popularity of the digital medium even as more and more content was consumed through digital channels – and quite naturally, the marketing and advertising dollars followed suit. This graphic from IAB captures this rapid growth: 

Understandably, a significantly greater focus of marketing activities now involve the digital medium – not as a “nice to do” – but as the leading medium to engage with clients, potential customers and other stakeholders. And just as inexorably, marketing analytics activities have followed through to involve the digital medium. 

Ranging from online surveys to understanding lead generation activities across various digital channels, from sentiment analysis impacting sales pipelines to understanding digital footprints – Marketing Analytics has embraced the digital medium completely and forms a beautiful extension to the traditional focus areas of Digital Analytics. To be sure, the actual techniques of Marketing Analytics Modeling have not significantly changed – continuing to rely on Frequentist and Bayesian techniques as well as other traditional data mining methods to develop the actual models – but the source of data, and the medium through which the model predictions are effected has taken on a distinctly digital edge.

The medium does present interesting challenges however – the anonymous visitor over the registered user, unreliable survey participation, the ever present challenge of spam, the smoldering issue of privacy – the issues of digital analytics are now also becoming the issues of marketing analytics. Marketing models have traditionally relied on a lifecyle of 18 months before typically requiring a refresh (owing to data drift etc.) – and it remains to be seen whether models developed with the digital medium as the primary mode of interaction continue to have that life-span or not!



Sri Srikanth

Advanced Data Analytics & Strategy, Senior Data Scientist, Cisco Digial