![]() |
|
|
Letter Perfect
Mar 1, 2007 12:00 PM
, By Richard H. Levey
Overstock.com has a tag line — “It's all about the O.” But with help from Teradata, Overstock has added four more letters to its lexicon: E, as in mail, and C, R and M. The backbone of the online discount retailer's new focus is a centralized data warehouse and campaign management program. Innovations allowed by the new data structure have proven so successful that last December a panel of judges selected by the National Center for Database Marketing chose Overstock and Teradata not only as gold award winners for technological applications, but as platinum award winners — the top honor — for overall excellence. E-mail and the Web are Overstock's primary marketing channels, although it does use radio and television spots, and to a much lesser extent direct mail. But e-mail is its forte, and the new system has worked wonders for Overstock's efforts. Before implementing the new system, it had been broadcasting 15 million messages every week. But these campaigns used only three basic groupings, with virtually no personalization or targeting. The new structure supports the company's e-mail promotions, as well as those done through other channels. It now broadcasts 25 million e-mails weekly — and the system was designed to scale up to 100 million per week. These messages are sent to more than 300 segments, and are part of event-driven trigger campaigns, each of which feature e-mails personalized and targeted to individual customers. The project was ambitious from the start. “At the end of the day, we built one of the most sophisticated data warehouse and business intelligence systems I am aware of,” says Paul Longhurst, Overstock's director of data warehousing. Instead of building a simple marketing database, Overstock elected to create an enterprisewide system that would integrate data from all departments. The company refers to the current structure as “a single version of the truth.” It incorporates sales, marketing, clickstream, service and finance data. Prior to setting up the current database, the company worked off multiple platforms. In doing so it encountered data quality issues that led to inaccurate conclusions, and delayed marketing decisions and targeting efforts. Designing a system that could be used by the entire organization called for significant cooperation among departments. The new data operations combine Overstock's marketing requirements and its IT unit's responsibility for integrating technology with business processes. The finance department also needed a structure that would define measurement and tracking procedures for a return-on-investment period. And operations personnel were charged with managing the organizational changes needed to best take advantage of the new technology and processes. If reconciling all the departments' needs wasn't enough, Overstock relies on historical data for behavioral analysis. The sheer volume of information being stored and analyzed meant that Overstock required a robust data warehouse. that would also support campaign management and reporting functions. Overstock started the initial deployment in March 2005, and within three months had completely migrated from its legacy e-mail system into the new structure. In subsequent stages it's included reporting, data mining and mass-marketing analytics functions (including optimal message delivery time, shopping cart abandonment, and propensity-to-buy models); segmentation and target marketing processes; predictive analysis and personalized communication; and event-based trigger efforts. Some benefits were immediate. “We went from inception to going live — sending out our first e-mail — in 60 days,” says Longhurst. In addition:
In addition, the combined customer warehouse has allowed Overstock to cut query times. Previously, gathering data and running numbers for product analysis took several days, and generating customer analysis reports took several weeks. Today, management receives hourly updates on customer retention, new customer acquisition and product sales. Overstock's management considers the ability to turn campaigns around — changing copy, or reallocating marketing dollars by channel — the most significant achievement. Overstock is relying on NPS to evaluate the system's success. It stands for Net Promoter Score, a new measurement proposed by loyalty marketing guru (and Bain & Co. fellow) Frederick F. Reichheld. It's based on a single question: “How likely is it that you would recommend [company X] to a friend or colleague?” Reichheld found that those offering a 9 or 10 ranking showed the strongest correlation of repurchase and referral rates. Overstock relies on a suite of products from Teradata. A warehouse serves as the central repository for data across the company. Teradata CRM handles Overstock's campaign management, segmentation and personalization functions, and NCR's Postmaster product facilitates the e-mailings. Good as Gold
Besides Overstock.com and Teradata, three other entrants received gold awards at the 2006 NCDM conference: GMC/SpeedTV.com and MindComet for multichannel marketing.
Marsh Affinity Group for analytics and modeling.
BellSouth and Harte-Hanks for business-to-business.
W
For more CRM and database marketing material, go to http://directmag.com/disciplines/crm/. |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
||
| May 1, 2007 | April 1, 2008 | March 1, 2008 | February 1, 2008 | January 1, 2008 | December 1, 2007 | November 1, 2007 | ||
|
|
![]() |
![]() |
![]() |
||
| Subscribe | View Sample | Subscribe | View Sample | Subscribe | ||
| © 2008 Penton Media, Inc. | Home | Penton Media Inc. | Contact Us | For Advertisers | For Search Partners | Privacy Policy |