AutoTrader.com: Bringing e-intelligence to the World's Largest Web-Based Automobile "Super Site"
Traditional brick and mortar companies building e-commerce systems generally start out with a solid base of customer knowledge from previous operations. But e-intelligence offers companies with little or no previous customer connections, such as companies founded on Web models, a chance to quickly gather tremendous knowledge about their customers.
Joe Whitehurst, President of Whitehurst Associates, is helping AutoTrader.com do more with its Web transaction information. "The value of e-commerce is not merely being able to transact at lower cost," observes Aaron Goldberg, Vice President and Principal Analyst at Ziff-Davis. "As time goes on, the value clearly is in the ability to customize transactions to the individual in real time, ultimately boosting revenues and profit." It is just this sort of customer knowledge that drove Atlanta-based AutoTrader.com, a result of the merger of AutoConnect.com and AutoTrader Online, to explore and exploit online buying patterns of its customers. Following the merger, the company today is the world's largest Web-based auto super site. With some 1.25 million used-car listings, AutoTrader.com is reinventing the way individuals shop for used vehicles, while becoming a part of a "community of interest" about auto-related information. In addition to viewing the largest database of used vehicles anywhere, customers browsing the AutoTrader.com site find information on leasing, financing, insurance, vehicle performance reviews, special offers, and a wealth of other facts and figures they'd never find at a typical used-car lot. With this kind of valuable information just a click away, and with scores of partners referring customers to AutoTrader.com, attracting visitors to the site has not been an issue. The challenge clearly was how best to harvest the information from those visits, and then to analyze it to extract knowledge from the data to make better business decisions. This challenge fell to Whitehurst Associates, an Atlanta-based consultancy that has been using SAS software to design applications and perform analyses for more than 20 years.
According to Joe Whitehurst, President of Whitehurst Associates, AutoTrader.com's management sought to view customer site visits from a number of perspectives and levels. They wanted to know who is visiting their site, what were the preferred paths the customers took to get to certain parts of the site, at what levels customers entered the site, how often the same customers visited the sites and what sorts of buying actions repeat visitors took, who referred the customer to the site, and how the buying habits of customers referred to the site by AutoTrader.com partners compared to non-referred customers.
Regarding actual purchase decisions made by customers, AutoTrader.com sought to harvest information on specific makes and models in greatest or least demand -- and on whether buying habits for particular styles varied from one area to another. "We approached this not so much as a Web-site-specific issue, but more as a data warehousing type of issue," notes Jay Stevens, Principal Consultant at Whitehurst. "So our plan was to use the AutoTrader.com Web data in addition to other data sources, clean it all, and then begin building business models to extract business information and intelligence based on actual customer activity." With data being crunched on a Sun Microsystems 4000 four-processor machine, Whitehurst employed a series of SAS software products -- including integrated capabilities for interactive applications development, information and presentation graphics, Web enablement, and SAS Institute's Scalable Performance and MultiDimensional Data Servers. "We could see growth by a factor of 10 in the next year on traffic to this site," Whitehurst says. "We get great scalability with the Scalable Performance Data Server, so we aren't too worried about that kind of growth."
One of the most user-friendly applications is a "management dashboard." As Whitehurst explains, the application jump-starts automatically around 2 a.m., searching for logs of site activity, unpacking and parsing them. Via an internal intranet built with SAS software, and leveraging the software's multidimensional viewing capabilities, the application generates internal Web pages with appropriate statistics and graphs for early-morning management review. Managers can easily drill down in the reports to view data in tabular format or as 3-D charts. What do they see? "They look at analyses of traffic generated by particular partners or the results of specific advertising campaigns and the impact they have on individual buyers," Whitehurst says. The information gathered also helps AutoTrader.com with Web page design, as management can analyze browsing patterns and see how buyers respond to specific pages. That is, AutoTrader.com is moving toward tailoring pages to specific, individual buyers -- a sort of holy grail of personalization and the ultimate promise of e-commerce. "Rather than trying to scan through reams of paper reports, sales people and managers can go right into the system interactively and extract business intelligence directly from the application," notes Stevens.
Submitted by Joe Whitehurst
Joe Whitehurst, President of Whitehurst Associates, is helping AutoTrader.com do more with its Web transaction information. "The value of e-commerce is not merely being able to transact at lower cost," observes Aaron Goldberg, Vice President and Principal Analyst at Ziff-Davis. "As time goes on, the value clearly is in the ability to customize transactions to the individual in real time, ultimately boosting revenues and profit." It is just this sort of customer knowledge that drove Atlanta-based AutoTrader.com, a result of the merger of AutoConnect.com and AutoTrader Online, to explore and exploit online buying patterns of its customers. Following the merger, the company today is the world's largest Web-based auto super site. With some 1.25 million used-car listings, AutoTrader.com is reinventing the way individuals shop for used vehicles, while becoming a part of a "community of interest" about auto-related information. In addition to viewing the largest database of used vehicles anywhere, customers browsing the AutoTrader.com site find information on leasing, financing, insurance, vehicle performance reviews, special offers, and a wealth of other facts and figures they'd never find at a typical used-car lot. With this kind of valuable information just a click away, and with scores of partners referring customers to AutoTrader.com, attracting visitors to the site has not been an issue. The challenge clearly was how best to harvest the information from those visits, and then to analyze it to extract knowledge from the data to make better business decisions. This challenge fell to Whitehurst Associates, an Atlanta-based consultancy that has been using SAS software to design applications and perform analyses for more than 20 years.
According to Joe Whitehurst, President of Whitehurst Associates, AutoTrader.com's management sought to view customer site visits from a number of perspectives and levels. They wanted to know who is visiting their site, what were the preferred paths the customers took to get to certain parts of the site, at what levels customers entered the site, how often the same customers visited the sites and what sorts of buying actions repeat visitors took, who referred the customer to the site, and how the buying habits of customers referred to the site by AutoTrader.com partners compared to non-referred customers.
Regarding actual purchase decisions made by customers, AutoTrader.com sought to harvest information on specific makes and models in greatest or least demand -- and on whether buying habits for particular styles varied from one area to another. "We approached this not so much as a Web-site-specific issue, but more as a data warehousing type of issue," notes Jay Stevens, Principal Consultant at Whitehurst. "So our plan was to use the AutoTrader.com Web data in addition to other data sources, clean it all, and then begin building business models to extract business information and intelligence based on actual customer activity." With data being crunched on a Sun Microsystems 4000 four-processor machine, Whitehurst employed a series of SAS software products -- including integrated capabilities for interactive applications development, information and presentation graphics, Web enablement, and SAS Institute's Scalable Performance and MultiDimensional Data Servers. "We could see growth by a factor of 10 in the next year on traffic to this site," Whitehurst says. "We get great scalability with the Scalable Performance Data Server, so we aren't too worried about that kind of growth."
One of the most user-friendly applications is a "management dashboard." As Whitehurst explains, the application jump-starts automatically around 2 a.m., searching for logs of site activity, unpacking and parsing them. Via an internal intranet built with SAS software, and leveraging the software's multidimensional viewing capabilities, the application generates internal Web pages with appropriate statistics and graphs for early-morning management review. Managers can easily drill down in the reports to view data in tabular format or as 3-D charts. What do they see? "They look at analyses of traffic generated by particular partners or the results of specific advertising campaigns and the impact they have on individual buyers," Whitehurst says. The information gathered also helps AutoTrader.com with Web page design, as management can analyze browsing patterns and see how buyers respond to specific pages. That is, AutoTrader.com is moving toward tailoring pages to specific, individual buyers -- a sort of holy grail of personalization and the ultimate promise of e-commerce. "Rather than trying to scan through reams of paper reports, sales people and managers can go right into the system interactively and extract business intelligence directly from the application," notes Stevens.
Submitted by Joe Whitehurst
Tags: sas
Advertisement
Featured
Analytic Jobs
Events
© 2009 Created by Vincent Granville
