Learning from the past is good, but predicting the future is better.
Retailers are leveraging predictive technology tools to discover the competitive power of data for a variety of applications ? customer-facing and operational ? and they are reaping the benefits.
What is predictive analytics? A very basic definition is mining data for information to be used in predicting trends and behavior patterns, or as Linh Peters, VP of marketing at SpartanNash, in Grand Rapids, Mich., puts it, ?predicting the future using data from the past.? But there?s much more to it, she notes.
?There are many tactics, tools and capabilities that would fall under this definition,? Peters says. ?At the end of the day, the goal is to make smarter, more strategic and informed business decisions that meet consumer needs.?
For example, predictive analytics tools not only take into account past data, such as product sales, but also the forces that will shape them in the future. And it?s on a ?per-individual level,? according to Eric Siegel, a former Columbia University professor who founded the Predictive Analytics World business conference, based in Santa Barbara, Calif., and authored ?Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie or Die.?
?It is technology that learns from data to render predictions per individual, which in the case of retail, is the individual consumer or shopper,? Siegel notes. ?That?s what differentiates it and makes it by definition more actionable, more directly applicable, in rendering mass-scale operations more effective than other forms of business intelligence.?
Uses Across the Enterprise
Among the many uses supermarkets are finding for predictive analytics are promotions and coupons; campaign management; category management; assortments; inventory planning; resource and staff planning; shopper patterns, including attrition, pricing, e-commerce and mobile marketing; and various aspects of the supply chain.
?We are using it for pricing. We are using it in promotion. We are using it in our campaign management,? says an IT executive with a major supermarket chain, noting that it?s too early to know the results, as the retailer has been using predictive analytics for only six months.
?We are investing heavily in software as service types of applications that help us do predictive analytics tied into supply chain, marketing, pricing,? adds the executive, who asked not to be named. ?We are doing a lot of work in that space using solutions, so we are not creating our own. There are apps out there that we can use. We see a lot of value there.?
Like many organizations, SpartanNash leverages predictive analytics to determine sales and volume forecasting, according to Peters. ?From a consumer standpoint, SpartanNash has made investments in technology and resources to ensure that we have the tools and capabilities in place that allow us to access and leverage data more broadly across the organization,? she says.
The company sees significant opportunity as it relates to making more informed decisions in areas such as assortment, planograms and CRM strategy, Peters adds. ?This not only benefits the shoppers of our corporate retail stores, but also provides meaningful data and insights for our distribution customers that will help them compete and differentiate their businesses.
?SpartanNash believes the potential for predictive analytics is significant,? she continues. ?Equally significant is the plethora of information that is available to harness. Many retailers are able to capture information about their shoppers either through a loyalty program or analysis of credit card data. But with the growth of social media [and] mobile and digital channels, understanding how consumers think, feel and behave has become even more complex. The more information available, the more difficult it is to ?predict the future.??
As an example, SpartanNash has been using predictive analytics to better understand shopper preferences for promotions and products. ?By leveraging our loyalty program data, we have been able to deliver more meaningful and relevant communications and offers to our shoppers,? Peters notes. ?We have seen positive response both in sales and engagement from the consumer.?
Kroger?s Aggressive Stance
The Kroger Co. is taking an aggressive stance on predictive analytics, which is typical of its use of technology. The Cincinnati-based grocer is using a Queue Management solution from U.K.-based provider Irisys, to make sure shoppers never have more than one person ahead of them at checkout.
Meanwhile, Harris Teeter is using a Big Data analytics platform from Tresata, based in Charlotte, N.C., ?to dynamically understand its product, customer and channel behaviors, in an effort to provide its customers better value across its online, mobile, social and brick-and-mortar channels.? Matthews, N.C.-based Harris Teeter, now a wholly owned subsidiary of Kroger, began the program last year.
When online merchant FreshDirect, based in Long Island City, N.Y., was shopping for a new market, ?they used predictive analytics to help them decide where to go, and ultimately chose Philadelphia as their second location,? notes Alan Lipson, global retail industry strategist at Cary, N.C.-based SAS.
Paul Scorza, CIO at Ahold USA, which operates from Carlisle, Pa., and Quincy, Mass., points out two current product examples where predictive analytics were used to inform current manufacturing and promotional efforts. One is the rapid growth of pedometers, based on an increased interest in fitness and the emergence of wristband units. ?I believe that prediction is extremely accurate,? he says.
Another product more typically found on supermarket shelves is coconut oil in its many forms. ?People put it on their skin; they put it on their food. There are all these uses for coconut oil, and a year ago, I didn?t even know what it was,? Scorza says. ?Now you can?t keep it on the shelves. That?s an example of how a supermarket retailer would have used predictive analytics a year ago.?
Scorza pegs the general accuracy of predictive analytics at about 50 percent, remarking, ?That?s pretty good.? It may be somewhat lower in the retail environment, because of the difficulty in predicting consumers? buying patterns, and the emergence of an entirely new cohort of shoppers, the Millenials.
?To me, the only predictive analytics that I would have a lot of faith in would be ones that are near-term, that are predicting current trends that are going on.? That is, predictive analytics based on three- to six-month-old data will be the most accurate for retailers, he notes.
Predictive analytics is one of the highest levels of business analytics. ?That?s the end goal everybody is trying to get to when they do business analytics,? Scorza says.
But the majority of retailers aren?t doing true predictive analytics, because they don?t have good, clean data, he asserts: ?Predictive analytics is becoming a reality, but first you have to have good data. ? [It?s] something that can be very powerful for retailers. It allows you to see the future and tells you what?s coming, and it?s all based on historical information, as well as current information that isn?t always so structured.?
?A Requirement of the Business?
Lipson of SAS asserts: ?You?ve got to be using analytics. There is just no question. It is, in fact, a requirement of the business. Given the speed at which retailing is moving, your customers are moving quickly. They are on social [media]. They are able to do the price-matching. They are able gather all that investigative information. So you need to know what has happened in the past and integrate that into your processes. Things are moving so fast that you accomplish things manually any more, or even with Excel spreadsheets.?
Because the volume of data is only going to increase from this point forward, with the Internet of Things and event-stream processing coming along, it?s just going to get faster and faster, he adds. ?It is a great opportunity for retailers to continue down that road, to see what they can leverage out of the vast data stores that they have,? Lipson notes.
The chains that are most effectively leveraging predictive analytics are those applying experimentation across all aspects of their business, says Jeff Campbell, VP of client services at Applied Predictive Technologies, in Arlington, Va. ?These chains know that rapidly and rigorously testing new ideas enables them to stay ahead of the competition and keep up with evolving consumer preferences,? he notes.
For retailers that incorporate predictive analytics to drive activation programs, ?we see a strong incremental volume response, typically in the range of 2X or more,? says Bob Tomei, president of Chicago-based IRI Consumer & Shopper Marketing.
?Insight-driven activation platforms ? programs, data, analytics ? are really the future of marketing and sales in our industry. From my perspective, those that get it right and operationalize it will be the winners, and those that for some reason cannot do it will be the laggards in our industry. The future is now,? he notes.
Shoppers? expectations are on the rise, notes Danny Silverman, VP of Cambridge, Mass.-based Clavis Insight. ?Whereas there was a time when a supermarket had a captive audience within their geographical sphere, the acceleration of grocery delivery via the likes of Peapod and Instacart has suddenly given the shopper an immediate alternative,? Silverman says. ?These online-only businesses live and breathe Big Data and predictive analytics to the point where it?s a standard process, not a project to be undertaken. Thus, shoppers often discover a more intuitive and convenient shopping experience. Supermarket retailers can combat this with an excellent in-store experience and outstanding customer service, both of which require the power of Big Data and predictive analytics.?
Retailers need to know that the use of predicative analytics ?is essential to establishing and maintaining a competitive advantage, and that assistance is available to help them integrate this data-driven function into their marketing and sales operations,? says John Ross, chief marketing officer of Inmar and president of Inmar Analytics, in Winston-Salem, N.C.
?As shoppers are demanding more personalized engagement and are more inclined to do business with those who meet this demand,? he adds, ?retailers must have predictive analytics in place in order to engage at this level, particularly with members of their loyalty programs.?
?The potential for predictive analytics is significant. Equally significant is the plethora of information that is available to harness.?
?Linh Peters, SpartanNash
?Predictive analytics is becoming a reality, but first you have to have good data.?
?Paul Scorza, Ahold USA