How Cognitive Computing Can Help You Shop Smarter

Share this post:

If you have kids, you probably know about the simulation game, Minecraft. It‘s a multiplatform game that enables players to build fantasy worlds out of textured cubes in a 3D space. This game has been a favorite of kids and teens since its debut in November 2011.

Yet unless you’re a Minecraft maven you might not have heard of Gameband, the first-ever Minecraft wearable, which wraps around your wrist and lets you take your worlds with you and play Minecraft anywhere, anytime – on any computer. Gameband is driving a surge of interest in Minecraft.

That’s is one of the surprising insights surfaced by a new free mobile app and Web service that we’re launching today—IBM Watson Trend. (iOS and IBMWatsonTrend.com) Our goal is to put the power of Watson’s cognitive computing in the hands of shoppers everywhere.

The app and Web site offer visitors an up-to-the-minute ranking of the top trending products or categories in retailing. You can browse trends by categories and save and share your favorites.

The app not only tells shoppers what products are trending, but it explains why, who the trend applies to and what people are saying about it. That’s how it surfaced the story behind Minecraft’s resurgence. Armed with these insights, people should be able to more quickly and easily focus in on gifts that they’ll want to get family members and friends.

I envision a big multi-generational family scattering around the house after Thanksgiving dinner. They’re loosening belts, kicking off their shoes, and settling into the couches with their iPhones and iPads. Time to plan their holiday shopping. And Watson Trend is there for them—showing off what cognitive computing can do.

We are starting off this year by focusing Watson Trend on three categories: consumer electronics, toys and health—but we’ll expand into new categories next year.

We throw a wide net in search of our featured trends—searching through millions of conversations on social media, blogs, forums, ratings and reviews. We then use Watson’s ability to understand natural language and reason to identify, measure and predict trends.

The Trend app and Web site are designed for use by consumers, but we believe that in the future, retailers and consumer packaged goods companies will use our data to analyze market trends, develop merchandizing and marketing campaigns, and plan future products and services.

How does Watson Trend work? We tapped into a handful of Watson cognitive technologies and added a new one of our own.

Step 1: Watson collects months of chatter off the Web, identifies individual conversations related to purchase intent and purchase experience, aggregates conversations into trends, and measures the sentiment surrounding each trend.

Step 2: Watson creates a trend forecast. We measure the impact and momentum of each trend and use that analysis to calculate a trend score and ranking for each day. We run the historical trend scores through hundreds of predictive models and use cognitive technology to select the best model(s) for each trend, each day. Then we forecast the trend score over the next three weeks. The idea for the predictive technology came from the data science genius on the team, Geoff Pofahl.

Step 3: Watson examines share-of-voice and cohort data to provide additional analysis—what we call “the story behind the trend.” We present app users with snippets and quotes that help capture the spirit of conversations about each trend.

This whole adventure started when IBM Commerce staged an internal innovation contest to bring new, high-value offerings to life. I pitched a trend service for retailers and consumer products companies to a panel of six IBM executives. A few weeks later, they asked me to not only build the trend service but to develop the consumer app and web site as well.

Crazy!

The IBM Watson Retail Trend development team

I am thrilled that I got to lead a team of ten designers and developers who brought the Watson Trend app to life over the past three months. They are an amazing team of IBMers! I have been a business consultant nearly all of my professional life, so running a product development team has been a kick. One of the things I like most about IBM is you get to move around and follow your dreams.

It has been a wild ride. We built the app in three, short months using the agile software development model—which is all the rage at IBM these days. Now it is done, and I can catch my breath and think about what comes next.