Slashdot: Getting more out of your data warehouse

Understanding your customer’s online behavior can be vexing. There is probably no other time in history when we can collect so much customer data and yet know so little about who visits our websites. Over the years large data warehouses have been built to try to address this question, but often seeing what your customers are doing, such as what decisions they are making as they shop your online storefront, isn’t really captured in the warehouse.

Looking at what data do you throw away could also provide insights, as is linking separate data stores together, or tracking persistent data across different visits. All of these are key to increasing insights into our customer’s behavior. These factors were part of several presentations at this week’s Teradata Partners annual user conference in Dallas. Let’ take a look at some specific examples of how to get better at learning about your customers.

Germany’s largest online retailer, the Otto Group, gets about a million daily visitors to its fifty different Web storefronts. They set out last year on a project to better track their customers. Through a combination of tools including Hadoop and a massive Teradata data warehouse connected to their Intershop ecommerce system, they were able to sift through terabytes of website log files. They came up with what they call “Customer DNA” to identify how their customers come and go on their sites.

“Big Data wasn’t invented so that more people can stare at data flow diagrams, but to look at site optimization and close some loopholes,” said Joachim Glaubrecht, a technical product manager with the company that was one of the presenters. “Just looking at log files won’t tell you about product returns that happen several weeks later and how these impact our overall business. We needed to connect 20 different data stores and show what customers were actually doing, even if they made several visits to our sites over time. “As a result, we can find the ten page impressions for a single user over a span of several hours. Now we can also observe how leads enter our sales funnel and improve our conversion rates. We can shorten our typical site navigation paths too,” he said.

Otto Group was able to mix SQL and NoSQL data collections effectively, and this is something that John O’Brien of consultancy Radiant Advisors is seeing more frequently. “You have to know the relative strengths of both kinds of data and be able to leverage them in today’s modern platforms,” he said at another talk at the Dallas conference. “You need to leverage the different kinds of data persistence in both platforms, and understand the semantic context that data has so you can unleash the best data access for as many data users across your company as possible.”

Daniela Rodrigues, who is a BI Manager at Telefonica Brazil, developed another system to track the effectiveness of marketing campaigns for Brazil’s largest cellular phone provider. They were able to track within days whether a particular campaign was working, something that used to take closer to a month to run an analysis. They used a combination of SAS and MicroStrategy OLAP reports for this project.

Part of the problem is in understanding what specific characteristics of your customers you are looking for. PKO, Poland’s largest bank, was looking to roll out a new epayment app for its smartphone users and needed to identify those customers that were Internet-savvy and had the appropriate smartphones and were also comfortable with downloading apps. They used a combination of tools to comb their data warehouse and target the first 37,000 customers that fit their profile. But more importantly, they were able to measure the number of activations of their app by particular marketing campaigns to see which ones brought in the largest number of customers.

Speaking of banking customers, Mark Swenson, the Director of Campaign Management for Teradata, gave some telling examples of how one bank used ATM data to acquire new customers. “They found more than a million of their customers only interact with ATMs and never come into their branches. But these customers were completely invisible before the bank added this data into their Teradata data warehouse, and were able to track who were going to other banks’ ATMs for their cash withdrawals. By adding marketing messages to their ATM screens, they were able to acquire 1300 new accounts over three weeks.”

This brings up an important point. “Every day, huge amounts of data are literally thrown down the drain. This is a major missed opportunity,” said Hani Mahmassani, who runs the Northwestern University Transportation Center. At the Dallas conference, he gave several transportation examples where trucking firms have instrumented their vehicles mostly for compliance reasons but weren’t using the data collected to make supply chain decisions or to optimize their fleet operations.

Customers can also provide tremendous insight into their own behavior, if you spend the time to listen to then. “What your customers tell you about how they use your website tells you more about your customers too,” says James Taylor of Decision Management Solutions. At the Teradata conference, he spoke of a recent incident with his local grocery store, where he is a loyalty card member. “They sent me a postcard with a generic series of offers that weren’t very compelling. There was no attempt to target the messaging, and no way to load the discounts into my card or my smartphone app. I couldn’t take advantage of the discounts at the store because I didn’t have my loyalty card with me.” All of his purchasing information was tracked through the program, yet it didn’t appear that his grocery store cared about his business. “If you treat your customers like a number, they will eventually defect – interaction quality is as important as product quality.”

Speaking of product quality, another session at the conference was from Mohan Namboodiri, who is the VP of Customer Analytics for Williams Sonoma. “Data science is part of our brand building here,” he said. The retailer is very conscious of delivering a high quality online shopping experience along with its heritage of scouring the world for fantastic products. As a result they are in the top 25 sites according to online annual revenues. To accomplish this they needed to have one central view of all of their marketing programs, both online and off.

Like Otto Group, they developed a series of personas to characterize their site visitors using Teradata’s Aster. They also have various triggers that have been programmed to respond to particular customer actions, such as recent browsers of a particular item that is put on sale in its stores. “It could be borderline creepy, but it is a sale and so saving customers money trumps that,” he said.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.