By Bernard Marr
“We want to know what every product in the world is. We want to know who every person in the world is. And we want to have the ability to connect them together in a transaction.” The words of Walmart’s CEO of global e-commerce in 2013.
Today, retail is driven by data and technology – whether we’re talking about bricks-and-mortar stalwarts or the upstart online-only giants such as Amazon and Ali Baba which have come to dominate.
Old-school, physical shops, stores and supermarkets were experimenting with data collection and analysis before the online boom. Loyalty cards, credit card records and customer feedback were all used to build a picture of who their customers were and what they wanted.
However the catastrophic effects on high street retail certainly spurred them on to find new ways to stay competitive, as well as leveraging advantages that real-world outlets have over virtual storefronts.
Walmart was, by revenue, the world’s largest company last year. It is the largest retailer in the world, employs over two million people and takes in $36 million dollars at its 4,300 US stores every day.
Their attempts to use data to predict customer behavior reportedly date back to at least 2004, when chief information officer Linda Dillman examined sales data after Hurricane Charley to determine what would be needed following the forecasted Hurricane Frances.
As well as the predictable increase in sales of flashlights and emergency equipment, the period saw a totally unexpected increase in demand for beer and strawberry Pop-Tarts. This data was used to inform stocking decisions, and led to strong sales.
This realization of the powers of predictive analysis eventually led to the creation of @WalmartLabs. Since its inception in 2011 it created the Big Fast Data Team with the purpose of finding pioneering uses for data in retail.
One recent development is the Social Genome project, which aims to increase the efficiency of advertising on social networks by guessing what products people are likely to want to buy, based on their conversations with friends. According to their blog its syntax analysis is sophisticated enough to tell from a conversation about “salt” whether the customer is speaking about the movie Salt, or the mineral.
The Shoppycat service suggests gifts that people might like to buy for their friends, based on their interests and Likes, and they also experiment with crowd-sourcing new products with Get On The Shelf. Proposed products are put before a voting public to determine whether they should be stocked by the chain nationally.
They have even developed their own search engine, Polaris, which used sophisticated semantic analysis to work out what a customer wants based on their search terms.
Of course, with the majority of their customers outside of their homes while interacting with their services, mobile is a key priority.
In-store navigation apps steer customers through the aisles to the products they need. Soon they will also suggest routes to the products that Walmart thinks they want based on their social media chatter.
Items can be scanned to provide the casual browser with information about the product, including customer reviews, and if a product isn’t in stock at the store you visit (you could have checked before you arrived!) you can get driving directions to the nearest one that has the product on the shelf.
Real-world retailers such as Walmart, Tesco and Target have shown they have the guts for the fight. But it will be a hard fight. Last year, the amount spent by the average UK consumer shopping online overtook the amount they spent shopping in the high street for the first time, and by 2018, the high street’s share is expected to have fallen to 33%.
However although the e-tailers may have managed to turn some of the bricks’n’mortar’s advantages against them – online execs gloat at the fact their customers use real world stores to try out products before ordering them online – other advantages are trickier to co-opt.
For example the huge number of stores and outlets owned by a company like Walmart or Tesco means they can offer services such as same-day delivery to a far bigger portion of the population.
Sales volumes, it has been discovered, can be subtly manipulated with data in the real world just as they can online. Experiments by UK-based retailer John Lewis involving tracking footfall of customers through stores has found placing salesmen near certain products can lead to a strong increase in sales, whereas their proximity to other products makes little difference. This allows shop managers to make tactical decisions about the placement of staff which have a demonstrable impact on the store’s revenue.
Cameras, as well as newer developments such as the iBeacon mean there are more ways than ever to measure and analyze the behaviour of customers. For example beacons inside mannequins have been used to beam details about the clothes being worn directly to a customer’s smart phone, as well as the directions to where they can find them on the shelves.
Last year Walmart installed lightbulbs manufactured by GE and containing Apple’s iBeacon technology in stores around the world, so although a company spokesman told the Huffington Post in 2013 that the company “doesn’t monitor shoppers in their stores”, that may be about to change.
Bricks and mortar retail has certainly realized that if it wants to compete with others who offer the same service, but without the need for the customer to leave their home, they have to be twice as innovative as their online rivals to stay in the game.
Walmart’s continued dominance of the retail market in the US at least shows there is still strong demand from customers for outlets that they can visit and collect their purchases from at their convenience, rather than having to wait for deliveries.
But the e-tailers are closing the gap fast. We don’t know what the picture will look like in 10 years’ time but the innovations driven by analytics and big data – as well as the continued pressure to drive prices lower – will likely be good for us as customers.
.jpg)
.jpg)
.jpg)
.png)
