Retail is fundamentally a reactive business; as consumer demand changes, retailers react to meet that demand. What has changed in the last 50 years is in how quickly retailers must and can react to sudden shifts in demand. Retailers’ response time has gone from (potentially) months, to weeks, to days, and even intra-day. What has made it possible for retailers to compress their time-to-reaction is the ability to analyze a huge amount of information, about what’s available to sell, what’s selling, and where it’s selling, very quickly. And what made that possible was the realization of Moore’s Law, a 1965 prediction by a young researcher named Gordon Moore that the number of components on a single computer chip would continue to double every year, while the cost per chip would remain constant (he later revised that prediction to every two years).

In the last decade, new information sources have emerged (along with incredibly powerful but affordable computing technology) that can help retailers glean demand much more quickly; this is all the non-transactional “big data ” that RSR and just about everybody else been talking about for several years. The idea is that if retailers can capture and analyze consumer behavioral signals from the digital world (searches, click-throughs, social media “likes “, etc.), they can respond much more quickly, and indeed even predict how consumers will respond. Today, winning in retail all about being able to close the gap between something happening in the marketplace, and reacting to it. But tomorrow’s winners are likely to be those retailers who are good a predicting demand before it exists.

That at least is the logic that should drive retailers to invest in new-generation data analytics designed to deliver consumer insights very early on in the demand creation/fulfillment cycle. But change is hard, and retailers are struggling to modernize their businesses from being slow-reactive to fast-reactive, let alone predictive. In an in-progress RSR study on growth strategies, we have learned that “predictive analytics and visual presentation of data ” is mostly the focus of over-performing retailers but even those retailers are struggling to implement predictive analytics (less than 25% of over-performers said that they have implemented and are satisfied with such capabilities).

And that, in a nutshell, is the challenge that technology giant IBM is faced with, as it tries to prove the value of its cognitive computing technology, Watson.

‘Siri For Business’

IBM’s Watson technology has been in the marketplace for 5 years (it was in 2011 that the cognitive computer won at Jeopardy!). Someone recently described the Watson technology to me as “Siri for business “, and that on the surface isn’t far off. A user can ask Watson a question in natural language and get an answer back in natural language. The science of Watson is far too complex for simple explanations, but basically the system can develop and test hypotheses based on massive amounts of data. A simple example: if you ask Watson what “four by four ” is, it will analyze occurrences of that phrase in its huge bank of data to determine if the answer most likely is a (1) truck, (2) a piece of wood, or (3) “16 “.

The problem for IBM is, how can such a capability be focused on practical stuff? The first big use of the Watson technology was in the medical field by Wellpoint (renamed to Anthem in 2014). I had a chance to see a demonstration of this a few years ago at IBM’s White Plains, NY facility. The software could analyze all kinds of complex patient attributes (age, weight, diet, blood chemistry, DNA data, etc.) to help determine the efficacy of various lung cancer treatments based on a big collection of past patient histories. I remember my schizophrenic reaction: on the one hand it was amazing, and on the other hand it was truly frightening, as in, “I sure a heck don’t want a machine to prompt my insurance company to tell my doctor how I should be treated! ”

Amplify 2016: Your ‘Learned Colleague’

IBM has been trying to develop compelling sets of use cases that would justify a company’s interest – and that’s where The IBM Commerce suite comes into the picture. IBM has focused the Watson technology for the new generation of marketing, e-commerce, and merchandising practitioners. At the Amplify 2016 conference, recently held in Tampa, Florida, the focus of the event was “cognitive commerce “, and IBM was ready to show how cognitive capabilities can be used to augment (not replace) human decision making when it comes to developing marketing and merchandising plans.

My favorite moment of the conference was during the opening day keynote presentations, when Kareem Yusuf, IBM’s VP of Offering Management and Development, demonstrated the Watson Cognitive Rules Advisor. Referencing a recent IBM television commercial featuring Bob Dylan where Watson analyzed his songs for content and tone, the technologist challenged the audience to imagine their own workflows with such a “learned colleague ” at the ready.

To demonstrate, Melanie Butcher, the head of IBM’s User Experience Design Studio, played out a scenario as a marketer for a sporting goods company that wants to promote a “commuter bicycle ” event, to promote the joys of commuting to work via bike. Melanie asked, “Watson, I’d like to target a new audience for our paid social interaction ” to promote the event. She then proceeded to ask a series of qualification questions, such as, “Watson, consider customer, social media, and 3d party data and suggest relative attributes for targeting. ”

You can see where this goes – the Watson Cognitive Rules Advisor helped the marketer put together a targeted social promotion based on customer behavioral attributes to deliver a much more focused promotion. It was an impressive demonstration of the power of “big data “, analytics delivered via “the Cloud “, and (of course) cognitive computing.

More, Please!

As the data from RSR’s upcoming study on global growth strategies shows, it’s still early days for retailer adoption of predictive analytics with data visualization capabilities. IBM takes it one step beyond that with Watson’s cognitive analytics with natural language capabilities. That’s a big leap forward for most retailers, and IBM has a lot of work to do to develop a wide and compelling set of use cases to move the technology out of the realm of “how cool is that? ” to “I absolutely need this for my business “.

IBM is being sensible by focusing, not on the science behind cognitive computing, but in its ability to help real people harvest value from the massive amounts of data that are inundating them in their jobs. The biggest issue for IBM will be to make the compelling case before retailers invest too heavily in (perhaps) less daring but still very modern predictive analytics technologies – because the other thing that RSR has seen in its past studies of retailers’ investments in business intelligence & analytics systems is that they tend to be almost generational. After all, many retailers are still using the “data cubes ” developed in the 1990’s to analyze product movement information to make their marketing and merchandising decisions.