The New York Times had an interesting piece titled A Data Explosion Remakes Retailing. The gist of it was that retailers are now collecting all kinds of information about shoppers and shopping patterns and utilizing this information to sell more:
Retailing is emerging as a real-world incubator for testing how computer firepower and smart software can be applied to social science — in this case, how variables like household economics and human behavior affect shopping.
To be sure, major retailers like Wal-Mart Stores have long been sifting through in-store sales and demographic information to aim goods at different stores and to tightly manage supplies.
But what is changing, experts say, is the rapid surge in the amount and types of digital data that retailers can now tap, and the improved computing tools to try to make sense of it. The data explosion spans internal sources including point-of-sale and shipment-tracking information, as well as census data and syndicated services. Companies also track online visitors to Web commerce sites, members of social networks like Facebook and browsers using smartphones.
The better tools, they say, are ever cheaper and faster computers and so-called business intelligence or analytic software for finding useful information and patterns in that data.
Retailers are increasingly mining vast troves of digital information to improve the decisions they make about pricing, shelf-stocking and product offerings. “This huge and growing ecosystem of data is an asset that some retailers are really beginning to exploit for competitive advantage,” said Thomas H. Davenport, a professor of information technology and management at Babson College. “It brings more science into the business. Relying on gut feel is yesterday’s strategy in retailing.”
The article didn’t specifically deal with food retailing, though one reason the industry was so concerned about Tesco’s opening in America was because its facility with data, honed through its relationship with Dunnhumby, was legion. In fact, in an early exchange between yours truly and Bryan Silbermann, President of the Produce Marketing Association, about Tesco’s intentions to open in the US, it was to this data usage that Bryan pointed. As it happens, Tesco couldn’t use Dunnhumby in the US because Kroger has that relationship in the states. Kroger executives, who partner with Dunnhumby in the United States, also say the data is powerful.
Our sense is that though enormous amounts of data are being collected, it is used mostly in a rather “macro” manner.
We mentioned the other day that here at the Pundit we had a new sister publication, CHEESE CONNOISSEUR … Shameless self-promotion: You can purchase subscriptions here and gift subscriptions here.
When we introduced the magazine at the cheese counters of fine food retailers, we mostly did tests. We assumed that the many sophisticated retailers doing these tests would not merely look at whether the magazine sold but at the relationship between purchase of the magazine and other variables. Surely they would want to know if the purchaser of a copy of CHEESE CONNOISSEUR also bought high volumes of specialty cheese. Perhaps they would want to look at specific articles we ran — did a cover story on Italian Specialty cheeses boost sales of those items? Did a piece preaching the joys of pairing bosc pears with Parmigiano-Reggiano and Prosecco, a dry, white, sparkling wine from Veneto, boost sales of all three items?
Did the average CHEESE CONNOISSEUR reader purchase more high profit prepared foods, fine wines, specialty foods and expensive fruit at the start of the seasons?
All this data is already collected but someone has to decide to run the reports, analyze the data and then act on it and, virtually without exception, nobody had the time or inclination to do this for CHEESE CONNOISSEUR, at least not on the item level that one magazine represents.
It is understandable. After all, manpower is expensive and in short supply, but it does mean that we are not as effective in stocking our stores as we could be.
Right now data is used in the roughest way. For example, when Wal-Mart wanted to sell more organic produce, it looked at the stores where organic milk was selling well as the places to start. The challenge, though, is to understand the interplay between different departments, so, perhaps, carrying a full range of greens, even beyond what sales would seem to justify, leads to sales of the most expensive and profitable balsamic vinegars.
The answer is unlikely to be more people, so the key is software that can analyze the data in a useful way. This article in The New York Times may have been a little too quick to associate the collection of data with its actual use. Still, the old world of trained experts relying on their well-honed gut is setting; the future is data and its manipulation. Those who want to get ahead of the curve need to be focusing on this now.