“Modern marketing is all about making the most compelling offers
to the best customers the fastest.”
to the best customers the fastest.”
Cognitive computing has a big future in marketing. The ability to track virtually everything everyone does on line has elevated the analytical requirements for marketers to the scope of a global stock market. Billions of data points are flowing and changing minute by minute. It is not possible to manage in human time frames or processing capacity. Fortunately machine assistance is rapidly evolving from batch reporting to predictive analytics to automated analytics, natural language processing, natural language generation, machine learniing and cognitive computing. We’re not quite ready to add sentient computing but it will probably happen in our professional lifetimes. While all this sounds like it might squeeze the human value add out of marketing quite the contrary. It creates a cognitive surplus for marketers to spend their time on higher value, “right side of the brain” decision making instead of crunching numbers and clicking software UIs all day. For the foreseeable future humans will be better at applying curiosity and creativity to problem solving.
Virtual telesales is a representative example. As marketers become more sophisticated in understanding customer behavior, they are getting better at managing acquisition cost. High-potential leads get higher-cost resources like a dinner meeting. Low-potential leads go into automated email drip campaigns. But there is a huge midsection in the lead curve that requires a very low cost, highly effective means of further qualification. Traditionally, this is done through telesales services, which can be opaque, inconsistent, and expensive. Enter the virtual sales rep, a digital, rules-based learning algorithm that can replace telesales (via solutions like Conversica. )
For simple email interactions, totally virtual reps are not only indistinguishable from humans, they are often preferred. Why? Because they can be set up to be courteous and respectful and are inherently reliable and scalable. For example, the time it takes several people in marketing and sales to follow-up on a Web visit by a prospect is typically more than a day and can depend on complex scoring, routing, and territorial designations. Studies show that delays of more than a few hours dramatically degrade engagement rates. A totally virtual rep can follow-up on every single Web visit with a personalized message within whatever time period it is programmed to do so. And it can manage several qualified exchanges to the point of setting up a human-to-human sales interaction or providing links for digital commerce. This eliminates delays caused by human availability and preference. The quality, efficiency, and cost of the virtual sales rep are simply compelling.
But the virtual rep is only one beginning for cognitive. The level of complexity in ad optimization and the logical future of personalization (individualization) in multichannel marketing will require millions of decisions to be made in real time. Again, the future is already here. IBM officially added Watson to its marketing cloud solution offerings in 2015. Cognitive services will become embedded in many marketing applications through open APIs, SDKs and community initiatives. They have to potential to solve a great deal of the inaccuracy and latency in how brands interact with customers. The question is not so much about when or how cognitive systems will become mainstream in marketing but whether customers will even notice.