The Rise of Serendipitous Analytics

For those of us who work with analytic data as part of our marketing efforts, being able to slice and dice user data to create targetable cohorts is a routine part of our day-to-day. Nearly all of this segmentation has focused on analyzing data to meet specific customer requirements; who is most likely to want to upgrade their existing service plan, who is interested in similar products, who is ready to replace an aging product, etc.

DataXu recently released a report that correlated user behavior across normally non-correlated variables to create a serendipitous profile of users, connecting dots that most people would never think to connect (e.g. what type of soda do free checking account customers prefer?). This type of information is gathered through a Programmatic Marketing Platform that tracks user interaction with advertisements and websites as people more through their digital day. Now, of course, the initial reaction to the earlier soda question is, who cares? And the answer to that question would be, soft drink companies, who just picked up an unexpected angle on their customers.

So while this is weird, interesting, and somewhat cool, think about applying the same type of correlative analysis to social media. Programmatic platforms are great at high-volume tracking of consumer behavior, but the data is very light, very brief. If a company like DataXu is able to extract that kind of serendipitous correlation from surface-level transactional behavior, imagine what they (or someone like them) could do with the vast, deep, and rich trove of user data that people upload into social media sites on a non-stop basis.

One of the consequences of social media is continuous bifurcation of people into groups of common interest. This happens routinely on Facebook, and the reason this matters is once you’re in a focused cohort, there tends to be a lot more depth of discussion, and the contextual frame of reference becomes a lot more focused. This level of focus and detail, the nuances of consumer interest (rather than transactional behavior), will be what drives serendipitous analytics to a whole new level of cool, opening up opportunities that would not have occurred to most of us.