The turkey leftovers are long gone and the holiday shopping season is moving into crunch mode. And though Thanksgiving-weekend retail sales were a little disappointing, I’m more intrigued by—and upbeat about—an emerging, high-tech aspect of this annual winter spending orgy: the smart use of big data.
With more purchases happening digitally today, big-data analytics has become a crucial but often little-understood link between the shopper and the retailer, both around the holidays and year-round. The technology is providing shoppers better selection, customer service and more targeted advertising, while helping retailers sell more of those big TVs and hot handbags you want under the tree.
E-commerce providers can now leverage algorithms to target sales and promotions to the groups most likely to buy; they can also promote items that complement specific products a shopper has already purchased, like throw pillows to go with a couch, or jewelry to match a just-purchased dress. Big data has even helped retailers tweak the timing and frequency of their emails to customers. (Guess what: Lots of people read emails from retailers at lunchtime.) These communications are also highly personalized. One flash-sales site I know can send out tens of thousands of variations of its daily email, based on your preferences and past shopping history.
And of course, big data powers the most powerful targeting of all—“re-targeting”—through which a specific item you browsed on a website actually follows you around the Internet, popping up in an ad to remind you it’s still out there, ready for purchase. A woman in my office told me recently about a designer handbag she had seen online showing up a few days later in her Facebook feed. This time, the retailer indicated that the specific bag was 40% off for a limited time. She immediately notified her husband and is hoping he got the hint.
In addition, retailers are using “dynamic pricing” to change the prices for thousands or even hundreds of thousands of goods on the fly, depending on supply and demand as well as what their competitors are charging. This is similar to how airlines price seats. But in the retail world, this means a seller can see immediately that Amazon has dropped its price on, say, a big-screen TV, then slash its own price in a matter of seconds. According to research firm 360pi, Amazon and Sears alter their prices for 15 to 20 percent of their online goods at least once a day.
When it comes to actually fulfilling online orders, data analysis can be critical in figuring out where goods are in the supply chain, which warehouses are under-stocked and how quickly orders can be fulfilled. After the black eye many retailers suffered last year promising Christmas delivery for goods ordered as late as Dec. 23—and not delivering–I’m sure data jockeys have been hard at work all year to prevent a repeat of that disaster. According to customer-service software firm StellaService, several retailers, including Ralph Lauren, Kohl’s and Target, have altered their Christmas cutoff dates in recent days, no doubt as a result of data analysis.
Also on the customer-service front, savvy retailers are now tracking metrics such as call-center wait times and how long it takes reps to resolve issues on the phone or over online chat. Many are studying data that benchmarks them against competitors. If L.L. Bean is answering customer calls in nine seconds, competitors want to know that, so they can adjust their wait times to better compete.
A new metric called “issue resolution” helps retailers understand if they’re actually satisfying the questions being posed by customers who call or email them. If an issue-resolution score is low, a retailer might deduce that call-center reps aren’t adequately trained, or perhaps don’t have enough knowledge about specific products. A high issue-resolution score means fewer repeat calls and more savings. On average, every call answered in a call center costs a retailer $5 to $10.
All this online shopping and customer interaction can’t happen, of course, unless online-retail sites are in tip-top shape technically. They can’t crash as people clamor for those deeply discounted Anna and Elsa dolls, and they need to load quickly before shoppers lose interest and go somewhere else.
According to analytics firm comScore, e-commerce sales on the Monday after Thanksgiving hit $2.04 billion, up 17% compared to a year ago, making it the biggest online shopping day ever (and providing a bright spot when compared with the disappointing retail sales the previous weekend). So when websites go down—like Best Buy’s did for more than three hours on Black Friday—retailers miss out on a significant chunk of revenue.
Web performance-monitoring company Catchpoint recently analyzed the performance of the top 50 e-commerce websites and found that over the Black Friday weekend, e-commerce desktop pages were 20% slower than in 2013. Mobile pages were 57% slower. I thought these Web pages might just be bloated (kilobyte-wise) from eating too much turkey, but Catchpoint’s analysis blamed content-heavy sites and too many steps in the purchase-to-check-out process. However, with the proper insights and analysis, retailers can mitigate these missteps.
Big data doesn’t sound like a sought-after stocking-suffer. But for online retailers this season, it’s one of the best gifts Santa could bring.
Neeraj Agrawal is a partner at Battery Ventures in Boston. This post ran previously in Forbes.
*For a full list of Battery portfolio companies and exits, please click here.