Want more Insights from SAS? Subscribe to our Insights newsletter. Or check back often to get more insights on the topics you care about, including analytics, big data, data management, marketing, and risk & fraud.
How Walmart makes data work for its customers
Each week, more than 240 million customers shop at Walmart (online and at its banner stores), making it the world’s largest retailer.
And these days, Walmart is relying on data to power the best shopping experience for its customers, whether they’re buying from the web, via mobile devices or at traditional brick and mortar locations, says Jaya Kolhatkar, vice president of global data for @walmartlabs.
Jaya Kolhatkar, Vice President of Global Data for @walmartlabs
At the National Retail Federation's (NRF) 2016 Big Show conference in NYC, Kolhatkar discussed how @walmartlabs created a structure around its large volume of shopping data and about the culture that enables data scientists and engineers to use data to quickly build and launch new shopping experiences to customers.
To make data work for its customers, Walmart first had to make data work for its eCommerce business. And its’ eCommerce business is huge. $12 billion in sales last year. 12 e-commerce websites around the globe with a constant influx of data. Walmart’s e-commerce branch alone employees more than 3,000 technologists from the Silicon Valley to India, England and South America.
Kolhatkar says, “Our story is one of enormous data, and also one of huge data opportunity. Transactional data, online data, mobile data – it’s what allows us to serve our customers in new ways. But, the data is so large it would be hard to put it to work for us without having the proper underlying infrastructure.”
So how big is the data? Multiterabytes of new data are collected each day. And that is combined with petabytes of historical data. All covering millions of products and hundreds of millions of customers around the world. In addition, more than 100 million keywords are constantly analyzed to know what people near each store are saying on social media.
An infrastructure for making data work
- Data cleansing. “First things first,” says Kolhatkar. “When you have as much customer data as we do, it is absolutely imperative to cleanse it.” But on top of making sure the data is correct and of high quality, they also make sure it is kept appropriately and anonymized. Names, phone numbers and emails are segregated. Other details are encrypted. Those steps enable them to make data available for multiple users without worrying about violating privacy issues. That’s why data cleansing is the first step.
- People make data work. Kolhatkar says that the second key to @walmartlabs success is people. They call it the Big Fast Data Team. The team helps @walmartlab data users - which include developers, data scientists and business analyst - use the data effectively. The Big Fast Fata team acquire data, develop and operate data feeds, analysis tools and implement the infrastructure. It’s a very diverse group of people with a wide set of skills.
- Access. According to Kolhatkar, this is what really makes the magic happen. “Big data democracy is how we like to think about it,” she says. “We don’t have a data bureaucracy or a lot of approver steps. It doesn’t take months to get access to the data. Hundreds of teams can access the anonymized data simultaneously. And that’s how we make the data work for us, so it can then work for our customers. Because making the data available is key. If you have great data, it’s of no use if you keep it locked up.”
- Choice. Kolhatkar calls this “the not-so-secret-sauce.” @walmartlabs combines a plethora of tools from a variety of partners, vendors, open source and in-house developers. They started with an enterprise data warehouse and built an infrastructure for the data. What enables this approach to work is the use of multiple BI and analytic tools. This enables @walmartlabs to hire people with different skill sets. Whether it’s accessing a Teradata warehouse or using NoSQL, SAS and Hadoop technologies, they can bring in people with diverse talent. Using different technologies helps them go from data to prototypes to launching at scale.
Building apps for an improved customer experience
More than 240 million customers visit a brick-and-mortar Walmart store each week. So it’s not just about online shopping. “We wanted to develop mobile apps to give our in-store customers the best experience possible,” says Kolhatkar.
Twenty-two million customers actively use the Walmart app each month and it ranks among the top three retail apps in the Google and Apple app stores. The Walmart app enhances the shopping experience in Walmart stores with features that include checking in to pick up an online order at a Walmart store, refilling pharmacy prescriptions and finding an item’s store location.
An in-store mobile navigation system was developed so that customers can search for items and see exactly where they are located in the store. Another app, e-receipts helps customers eliminate the pain of keeping up with printed receipts. With real-time processing for thousands of transactions, it also helps reduce item-return fraud. Wishlist was introduced in the 2015 holiday season for customers who want to create lists of items they want or find someone else’s list.
Putting all of their customer data together is working. “People are now seeing the value of what is being generated,” says Kolhatkar. Her advice? “Figure out what data to collect and then look at areas that will deliver the most value. That’s how we’re building the next generation of e-commerce for our customers.”
- MBA marketing research students from Northwestern University's Kellogg School of Business provide guidance on how retailers can improve their customer loyalty programs.
- Retailers are among those organizations faced with Internet of Things data. Learn how to manage and analyze IoT data in real time.