Remember the movie “Minority Report” where Tom Cruise walks into the mall and billboards ping his iris, recognize him and make targeted offers? While that film was set 50 years hence, its reality may be closer than you think.
“We know what you are buying and what channel you use,” said John Strain, CIO of Williams Sonoma. “We can tell if you are a full-price, a discount or clearance shopper.”
At a recent conference hosted by data analytics provider SAS, Williams Sonoma was one of several retail companies that shared how it uses Big Data analytics to get closer to its customers.
Williams Sonoma: On Target with Big Data
Using tools such as the SAS Scoring Accelerator and a Teradata data warehouse, Williams Sonoma aims to know everything about its buyers. It creates targeted marketing campaigns which are all about relevancy. In addition, Strain said, analytics is not just about long-term historical buying habits. Companies like his also want to detect an ongoing buying cycle like the purchase of a couch.
“If you have been browsing couches, we have to contact you fast or we could lose the sale and then the data on your couch browsing history is useless,” he said. “We are learning to target customers based on days-old behavior cues.”
Williams Sonoma captures vast quantities of data online related to purchases, clicks, click-throughs, demographics and general Web browsing history, then mines it and retains it. The information is synthesized into predictive models built for each customer and each product. The end result is a score against each of five product categories per customer.
Based on that score, the company sends email marketing and other offers that it hopes will resonate with a customer. Rather than a barrage of emails across the range, it wants to present the right message at the right time.
“If someone abandoned something in the cart, for example, that’s a big driver for targeting,” said Strain. “In addition, cookies track where you have been and we feed ads based on that.”
If an item goes on sale, emails go out to anyone who has bought it before, as well as anyone who browsed for the item. The company's evaluation of data has also highlighted the importance of the local connection. Some users respond more readily to calls or emails from their own local store, rather than a corporate message. So Williams Sonoma sends lists to its stores so they can reach out to locals using the latest analytics.
The next stage for Williams Sonoma is real-time analytics based on what the customer is buying today – using the information to up-sell or make additional offers at the point of sale.
“We have to apply a lot more analytics to be relevant at the point of sale in real time,” Strain said. “Also, we are working on purchasing habits around larger items. We’ve noticed that people tend to look at them a few times before committing. So we are figuring out how to get their nearest store to interact appropriately for these items.”
Major League Soccer: Centralizing CRM Efforts
Major League Soccer (MLS), which oversees a network of about 30 teams around the U.S., is moving toward a centralized approach to business intelligence and CRM. The plan is to develop a centralized data warehouse to consolidate many existing data silos held by individual clubs.
“We are developing the infrastructure right now starting with merchandizing and digital subscriptions,” said Charlie Shin, CRM and Fan Engagement director for MLS. “We are shifting our focus from sales to being more customer driven.”
Because the clubs compete against each other, the database is being set up so that they can only view their own data and not see the others. They might be shown averages to see how they are doing on certain metrics, Shin noted.
“Teams can now tap into leads the league has received, while the league can tap into teams’ core fan base for mutual benefit,” he said.
The league is allowing each team to select its own CRM tools to harvest their own databases. When it comes to building analytics capabilities, though, MLS is moving forward at a league level. That will enable MLS to better consult the clubs on business intelligence matters.
This centralized BI approach is driven by the fact that some teams are far wealthier than others. By developing analytics at a league level using SAS BI tools, less affluent teams will be able to access the data relevant to their databases, Shin explained.
Expedia: Bunch of Business Intelligence Tools
Expedia.com is all about analytics. Joe Megibow, vice president and general manager of Expedia.com, said the high volume, transaction-focused world he lives in has been involved in Big Data for a long time.
“Due to the commoditization of the travel market and the challenge of choice, we are putting a lot of emphasis on creating affinity models, recommendation engines and analyzing behavioral data,” Megibow said. “We have to do a better job of assembling the data that is most relevant to a user.”
Expedia has accumulated half a petabyte of Hadoop data. It feeds data into Hadoop first and from there into its data warehouse. To analyze all of this information, it has accumulated a blend of business intelligence and analytics tools from SAS, Tableau, BusinessObjects and Google Analytics.
“No one system can analyze everything,” Megibow said.
As well as using lots of BI and analytics tools, Expedia employs an ever-expanding team of statisticians and data scientists to help decide what items to show to customers, how to provide the data and how to fit content to user types. The idea is for the system to take on more of the role that travel agents used to perform in tailoring a vacation package to user needs and preferences.
“With every click, we are building a signature of their choices and building affinity models based on similar users,” Megibow said.
Macys.com: IT in a Supporting Role
Kerem Tomak, vice president of Marketing Analytics, Macys.com, is a big fan of BI tools. But he said there is no escaping the human element.
“There is a human aspect to analytics that cannot be commoditized,” he said. “Even though we need to automate a lot of analytics to cope with Big Data, you still need a person to revisit it to see that it continues to be valid over time.”
He recently introduced Hadoop into analytics equation at Macys.com. He characterized it as a locomotive train, while a database management system is more like a sports car. By that he means Hadoop can be utilized to move large amounts of data and feed it into BI applications.
“Hadoop is a sandbox for our analytics teams,” he said. “They can use it to play with all the data you want.”
His analytics team needed a dedicated system for experimentation and Hadoop fits the bill. It was the cheapest way forward, he added, and speeded the Macys.com adoption of Big Data analytics by many years.
The retail giant has its analytics team operate independently of IT. The statisticians and data scientists on the team have skill sets that encompass coding and analytics – not an easy combination to find, according to Tomak. Rapid prototyping is done using the company's Hadoop and SAS-based analytics infrastructure to find insights and build working concepts. Once found, these are handed over to the IT team, which puts the concepts into production on the Macys.com website.
“Marketing will become more and more analytics driven in the coming years,” Tomak said. “The marketing teams will no longer need to go to IT for insights.”
Drew Robb is a freelance writer specializing in technology and engineering. Currently living in California, he is originally from Scotland, where he received a degree in geology and geography from the University of Strathclyde. He is the author of Server Disk Management in a Windows Environment (CRC Press).