Big data practices are a natural fit for retailers, which already parse inventory information from ERP systems, multiple databases and transactions from registers. But they haven’t seen anything yet, as RFID returns and retailers must confront wave upon wave of Twitter and Facebook data.
A recent survey by Enterprise Strategy Group found that more than a third of retailers surveyed deploy more than 200 production databases. Across the rest of the industries surveyed, only 15 percent could make that claim.
Steve Stone, former CIO of Lowe’s Home Improvement (s LOW), can vouch for that abundance of data. Lowe’s processed “something like 46 million line items per week as well as 15 to 16 million unique invoices, he said. “Retailers [do] tremendous analysis around inventory. That isn’t as intuitive — with invoices you can see a sale or transaction but with inventory you need to track it over time to understand if you’re running out of stock. Inventory is probably the largest data set.”
A return of RFID?
Now, as more stores want to allow unattended check-outs a la Apple(s aapl), they will need better loss prevention, which could mean a resurgence of radio frequency identification (RFID) technology. RFID was one of those next-big things in IT that didn’t really pan out because the readers were expensive and the read-rate failure was high. (The reader had to be really close to the tag). But with stakes getting higher, that calculus might change as retailers revisit the idea of tagging merchandise to enable easy payment and inventory monitoring.
More RFID tags and readers mean even more data. Those tags contain information that can be incorporated into store analysis, said Stone, who is now SVP of cloud intelligence at MicroStrategy(s MSTR). The business intelligence company brought Stone, a former customer, aboard to build out its cloud infrastructure to enable analytics on big data.
Big-box stores may well mimic the Apple model to enable this frictionless commerce. A consumer could use her smartphone to find a leaf blower or washing machine at the store, scan it with the phone for payment and schlep it to the car — all without waiting in a checkout line (or an unpleasant encounter with store security).
Analyzing customer scans in real time, a store could make sure the customer found and accessed the product before leaving in frustration. “If you’re a big store with high racks, and they scan the item they want, you want the item to tell you ‘I’ve been scanned and paid for,'” Stone said. “There’s no other way to do it.”
That information also tells the retailer what products are selling in real time so they can be restocked.
More big data tricks for retailers
Most companies dip their toe into big data when they start analyzing traffic patterns and behaviors on their web sites, Stone said. The next step is usually to run text analytics against Twitter to find and interact with customers who had a bad experience with the business. But Facebook is the motherload of data for retailers.
Facebook gives a retailer all sorts of affinity information about Facebook users who like or follow the company. “You get real tokens for the likes/dislikes of your followers and their friends’ likes and dislikes,” Stone said. All that psychographic (likes and dislikes) and demographic data (location, age, etc.) is invaluable. (MicroStrategy fields a Facebook Gateway analytics tool that runs against that Facebook data.)
“If you know how many of those who like Lowe’s love music, or that 95 percent of them love NASCAR, that’s the kind of thing marketers love. They can use that to target advertising. If you spend your ad dollars on Everyone Loves Raymond and no one watches it, that’s wasted money,” he said.
Jeff Bedell, CTO of MicroStrategy, also sees huge opportunity in Facebook for retail customers: “Twitter is fine for straight sentiment analysis, but what you get with Facebook is [things like] what colleges [your followers] were most likely to attend, the top books read by your people who read The New York Times (s nyt) … it’s very much about the network, analyzing the network versus analyzing the feed. The bulk of marketing activity now is happening through Facebook on the social networking side.”
Stone’s first Facebook promotion for Lowe’s was a huge success as gauged by customer reaction, although it wasn’t without problems. “We did it last year as a precursor to [Black] Friday [and] it nearly brought down our system.” But the net result was positive: Lowes.com business rose a whopping 600 to 700 percent compared to the previous year. That uptick gave it plenty more data to analyze and feed future promotions.