The “Quantent” Quandary
Last week, I had the opportunity to attend a teleconference highlighting IBM Watson’s success stories over the past year. Most of them are under NDA, so I can’t go into the details, but I will say they covered an incredibly broad range of use cases. One use case that I was hoping they would cover and didn’t was content analytics for marketing, aka “quantent.”
In the customer analytics arena, we often talk about “getting the right message to the right customer at the right time.” This is only partly true. Well-built and rigorously tested propensity models will deliver you the right customer and the right time. Behavioral segmentation models may even specify the best channel to use to deliver the message. But that still leaves the message itself. Whatis the right message?
Content analytics begins with entirely different data than customer analytics, and the two analytical streams merge just prior to the point of action. Whereas customer data contains information about customer profiles, transactions, and behaviors, data about content characterizes tone, length, wording, dates, products mentioned, type of offer (if applicable), and other key themes within the content itself. Most importantly, content that has been subject to A/B testing also creates data about the success of the message on an individual customer basis.
Quantifying the essence of content into data is no easy task. Until recently, it would have taken human beings to do so and would likely have produced disgruntled employees and inconsistent results. This is where IBM Watson and similar AI platforms have a huge opportunity. Watson’s cognitive computing capabilities, resting on a foundation of natural language processing and deep learning, make it a ready candidate for the onerous task of turning folders of content into analyzable “quantent.” Once this is accomplished, companies can truly begin to match the right message with the right customer at the right time by delivering the optimal content to a customer by matching content characteristics with the customer’s preferences and patterns of behavior.
The question that remains is whether cognitive platforms will soon create deliverable content on their own. That day is bound to come, but until then, marketers should focus on transforming existing content into “quantent.”