Elliott Cordo, Equinox Fitness' manager of enterprise data, succinctly described the kind of situation that put...
his company on the path to master data management (MDM): "Someone at the front desk would call up a customer name and get five different versions."
Customer-facing applications like marketing, sales and spa services generated millions of customer records, many of them redundant and inconsistent. Duplicate records made it almost impossible "to analyze one customer's behavior as a whole," which is crucial for cross- and upselling, as well as retaining and targeting members, Cordo said.
For a health club chain with a competitive differentiator of personalized service, this was unacceptable. "We need to know not just how many contracts we have, but how many active members who are fully engaged in gym activities," Cordo said. "Do they just use the gym, or do they use the spa or ancillary services?"
The company is now in the middle of a multiyear MDM initiative. Recently completed, the eight-month first phase involved deduplication, cleansing and standardization of customer profiles, the elimination of inconsistencies and identification of "all obvious duplicates," Cordo said. Equinox used SAS Institute Inc. subsidiary DataFlux's dfPower Studio, which Cordo described as "a toolkit for data quality" that automates much of the grunt work.
Sales and marketing teams held off on launching important campaigns until the work was complete, Cordo said. "With better data, we could perform segmenting and calculations on a member level," he said. "It made our targets more accurate; that's the most important part."
Equinox now has "a very simple customer registry" that extends across sales and membership, with spa and retail data to be added shortly, Cordo said. Data duplication has been reduced by 5% to 10%.
Equinox's situation is hardly unique. Consider a recent report from Renton, Wash.-based The Data Warehousing Institute that found 83% of organizations suffer from bad data for reasons that have nothing to do with technology. Among the causes of poor-quality data were inaccurate reports, internal disagreements over which data is appropriate and incorrect definitions that render data unusable.
For years, companies have struggled to clean up and integrate "business entities, like customer data and all related attributes and relationships" with the goal of creating a single, consistent data set that various applications can tap into, said Claudia Imhoff, president of consultancy Intelligent Solutions Inc. in Boulder, Colo.
"Cleaning up all the junk across the various systems gives you much higher-quality data" for business intelligence analysis and especially for customer-facing applications like sales, service and marketing, Imhoff said. It also helps companies satisfy the data demands of government regulators.
Many firms have attacked the problem with data warehousing and extract transform and load (ETL) processing, only to be stymied by the complexity and scope of the task of cleansing and integrating vast quantities of data, Imhoff noted. In contrast, master data management "unburdens the data warehouse team by requiring them to load not the data itself, but all pertinent information about the data into a single trusted master information file," she explained. Rules are then applied to ensure quality and consistency across all entries.
"Global Data Synchronization at Work in the Real World: Illustrating the Business Benefits," a report by GCI/CapGemini, summarizes some of the benefits companies have realized through MDM:
- Dutch retailer Albert Heijn B.V. improved productivity in its data management department by 30%.
- Wal-Mart Stores Inc. decreased item maintenance from 15 to 30 days to only one day.
- Johnson & Johnson decreased out-of-stocks by 2.5% by virtually eliminating data integrity issues.
- Unilever Columbia significantly reduced data inconsistencies and improved new-item time to market by two to three weeks.
- U.S. retailer Wegmans Food Markets Inc. increased store sales by reducing speed to market on new items by two weeks.
Cleaning up all the junk across the various systems gives you much higher-quality data.
Claudia Imhoff, president, Intelligent Solutions
While MDM has historically been enterprise turf, that's changing, industry sources agree. A lot of midmarket companies suffer from the same data redundancy and inconsistency problems as Fortune 500 firms, albeit on a smaller scale, noted Aaron Zornes, chief research officer at The MDM Institute in Burlingame, Calif. But midmarket CIOs have balked at the million-dollar price tag of most MDM systems.
Recently, however, companies like Siperian Inc., DataFlux, Initiate Systems Inc. and Purisma Inc. have introduced smaller-scale MDM platforms that cost $100,000 or less, Zornes said. Oracle recently introduced an "MDM-lite" version of its technology. And Microsoft will shortly unveil its own midmarket MDM solution, which will be closely integrated with popular offerings like SharePoint.
Then, too, midmarket firms are increasingly taking advantage of the data cleansing capabilities and other MDM features that ERP and business intelligence vendors are incorporating into their software. "We've seen smaller companies realize the discipline and benefit of having a single view of the customer and of material" through their ERP systems, said Robin Kearon, CEO of the North American division of global systems integrator Business & Decision SA.
Step by step
Not that implementing MDM is exactly a cakewalk: Just ask Cordo, who is about to embark on the next phase of Equinox's rollout. Scheduled for completion by the fourth quarter, phase two involves introducing data governance rules and automated procedures for standardizing data as it is entered. This will eliminate the need for administrators to do batch data cleansing on a nightly basis, Cordo said.
Creating consistent customer profiles and data definitions shouldn't be too hard, Cordo said. More difficult -- and politically sensitive -- will be working with business groups to devise data governance rules, "Like how to identify duplicates or what to do when someone has a long-standing account balance or if quality is trending down in a particular club or region, who handles it?"
Cordo said sometime next year he plans to build "a consistent lookup and search screen, with a unified view so that when customers come in, operators can identify them across our source systems in all business lines."
Cordo said he has no doubt that the rewards are worth the effort. The chief benefits, he noted, are unquantifiable: "supporting our culture and the way we interact with members and improving club operations." The bottom line: "MDM was the logical thing to do."
Elisabeth Horwitt is a contributing writer based in Waban, Mass. Write to her at firstname.lastname@example.org.