Dirty Little Data Secrets
Despite CRM's rapid evolution as a business process standard, some companies remain vexed by its frequently fuzzy ability to identify a customer at the point of interaction. While the days of messy CRM experiences like integration flameouts and legacy system nightmares have receded for the most part, myriad enterprise systems housing variations of duplicated, incorrect, and/or unusable customer data still frustrate enterprises of all sizes. Customer data integration (CDI), however, is succeeding where CRM has failed, and is helping to make good on CRM's promise.
"The dirty little secret about CRM is, it's become another group of legacy systems within the enterprise," says Jill Dyche, a partner in and cofounder of Baseline Consulting. This isn't to say that CRM has failed in its bid to unite sales, marketing, and customer service, but its success has revolved around unifying business processes. Data quality was an afterthought and relegated to point solution implementations, says Evan Levy, Dyche's partner and cofounder at Baseline Consulting. "The old sales pitch that CRM was the single version of the truth implied a unified source of customer data, enterprisewide," Levy says. "That hasn't happened."
With customer data distributed on average across 20 to 40 systems, companies understand the need to consolidate data via a unified hub to work for that single version of the truth. This realization has led to hard dollars, because despite its relative immaturity, the CDI/MDM market is growing at a dramatic pace. By 2008 the market for CDI solutions will exceed $1 billion, up from$680 million in 2006, according to Gartner, while IDC projects the overall MDM market to exceed $10 billion by 2009. This growth, however, is being hindered by uncertainty on the part of enterprises as they look to understand the cultural changes associated with using CDI and comprehend how it fits into master data management (MDM).
The CDI Difference: It's Baked In
CDI differs from previous data quality tools, and learning how CDI can improve customer relationships is key--again, it's important to note that previous data quality solutions haven't failed, but their success has been limited. "They've missed the big picture," says Navin Sharma, director of product management, global data quality, at Pitney Bowes. "Data quality became an after-the-fact term."
Data quality initiatives used to come into play at the application level and were specific to certain sets of functions like address standardization. As a result, the rules defining how to standardize customer data were specific to the needs of that application and/or department. The results were heterogeneous collections of customer definitions and identifying attributes relevant to a particular department, or group of departments, within the enterprise. "It's a matrix of relationships that ultimately boil down to one particular individual," Dyche says. "Even companies with great CRM and data warehousing are struggling with this."
Data warehousing; enterprise information integration (EII); extraction, transformation, and load (ETL) tools--(see this month's "Re:Tooling"
for an ETL tools review); and operational data stores (ODS) address these problems, and for most companies they have become the de facto remedy for the classic definition of CDI. But there are differences.
With a data warehouse or ETL tool, data integrity and validity are optional. The information a data warehouse collects could be good or bad, depending on the environment. Unlike warehouses, CDI contextualizes the data; it understands the concept of address, and is thus capable of identifying where and how transactions and data are related to a particular individual. A CDI hub standardizes any information about the customer by recognizing, comparing, matching, and reconciling customer data across disparate systems according to predefined rules. With CDI, data standardization and correction is "baked in," Levy says. The same can be said of ODS, which lacks the cleansing and reconciliation capabilities inherent to a CDI solution. Like CDI, an ODS pulls data from source systems like ERP and CRM, but unlike CDI, an ODS system is meant to be queried, not updated in real time.
The flip side to the installation of EII, ETL, and data warehouses is that a company generally has a foundation for CDI, says Ray Wang, Forrester Research's principal analyst of enterprise applications. "The beauty of [the CDI hub approach] is that most organizations already have most of the pieces in place," he says. "They just need to find a way to pull it together." CDI does this by consolidating functions into a unified hub that sits between--and works in conjunction with--a company's current data quality systems and enterprise apps, querying and pulling customer data from enterprise systems. "Think of CDI as a data traffic cop, taking the best [of customer information] from all systems and directing it into a central hub," says Tony Fisher, president and CEO of DataFlux.
The enabler of this data-centric concept, and the reason CDI didn't develop earlier, is SOA, which is what gives CDI the ability to synchronize customer data across systems. Whereas previous data quality solutions were limited to sets of functions or systems based on the format of the data, the development of SOA has meant CDI makes data quality available as a set of Web services. "Creating a new customer in a CRM system means going through a set of steps," says Anurag Wadehra, vice president of management and marketing at Siperian. "Those steps can be mapped out into a set of services that a CDI solution can call on to check other systems for customer data."
Taken in the larger context of MDM, CDI is simply a subset; MDM can be considered the overall discipline of managing information domains across an enterprise. While CDI is specific to automating the management and reconciliation of customer data, MDM involves managing the master data (all enterprise data), including product, pricing, inventory, or competitive data. "In essence, CDI is the MDM of customer data," Wang says.
The CDI Benefit: Streets Go Every Way
CDI's value to CRM lies not only in its matching and standardization capabilities, but in its capacity to then propagate updated customer data back out to enterprise systems, transforming aged data quality practices from one-way roads into multidirectional highways. "If I correct it in one place, I make it available to everybody else," Dyche says.
With operational systems on the same page, companies are consolidating account activities across sales and service channels. In the retail industry many businesses are selling through wholesale retailers and e-commerce sites. With customer data stovepiped in departmentalized systems, a customer who calls product support might not even be recognized. This problem also speaks to companies that leverage third-party data providers like Dun & Bradstreet. CDI and MDM hubs automatically format and cleanse the incoming data, removing duplicates before they reaches the operational systems.
CDI is also helping to relieve miscommunication between sales and marketing by improving lead qualification. Baseline's Levy cites one financial services customer that after using CDI, realized it was unwittingly sending home equity lines of credit to customers who also owned small businesses. These small business owners were responding to the home equity line of credit offers instead of the more profitable small business loan offers. The bank used CDI to match the records for its sales and marketing departments, to the tune of a 50 percent profit increase.
In another example, Sealing Devices, a $40 million a year manufacturer of industrial seals and gaskets, faced a similar dilemma: Dirty data was undermining a competitive differentiator. To continue to deliver stellar customer service while managing its growth, the firm needed to clean up its customer data and consolidate systems, which were, according to Patrick Harris, IT director, "a disaster. We would send quotes containing wrong information, because the customer might be on record as having three or four addresses. In addition, our ERP system would allow only one account per customer, so sales analysis was off."
Sealing Devices chose the Oracle Customer Data Hub so clean data could, well, seal the deal. Data Hub resolved data inconsistencies in the company's list of 10,000 organizations, 25,000 individual contacts, and 19,000 address books, saving approximately 20 days a year required for marketing campaigns. It also improved the productivity of Sealing Devices' inside sales team by 20 percent.
Describe, Define, Govern
For as much as CDI is a technology, it's a business process first, and just like the CRM implementations that preceded it, it should mirror a CRM rollout. Simply put, CDI projects are big, thanks largely to the scale and number of enterprise applications a CDI hub touches. The resulting fallout can be quite disruptive, and will touch upon users at nearly every level of the organization. Accordingly, getting executive buy-in is important. Until recently IT has taken the lead, making CDI very much a "feeds and speeds, features and function infrastructure issue," Levy says. "That's going to change."
Following its decision to implement CDI, a company must define what data contributes to an accurate picture of the customer. The definition differs for each department; management must develop a data management strategy in which all parties agree on definitions and labels for categorizing customer data. The resulting challenge is "where things can get ugly," Fisher says. "This is when politics comes into play."
Business and IT must agree on a data management policy--who has access to what customer information, what can they do with it, and how can they change it? The problem is amplified by the multitude of systems and users that touch the customer data. There must exist established governance processes and technology safeguards. "It's about forcing the departments of an organization to come together for the greater good of the company's data quality practice," Fisher says.
Two caveats: First, CDI implementations aren't cheap. Forrester's Wang says that average CDI installations cost about $5 million for licenses and implementation services. That figure can be divided into a 3:1 ratio of services to software, according to Fisher, though he expects that ratio to drop in coming years as industry best practices are established.
Second, these implementations are not quick. For a well-prepared organization average implementation and roll-out times will land somewhere between six and 12 months. An unprepared organization that falls into either stage 1 or 2 might see a deployment run anywhere from 12 to 36 months (see the sidebar "CDI on Stage").
Some vendors are taking a point-solution approach to CDI and MDM. Rather than boiling the ocean with an enterprisewide implementation (like MDM players IBM, Oracle, and SAP), providers Siperian and Purisma are making it easier by targeting specific departments with point solutions, providing "different entry points" into the broader MDM project, Wang says. "An organization doesn't have to go full throttle into a two- or three-year implementation. It can embark on the stepping stones that lead towards an MDM infrastructure-based approach."
For the market 2008 will be the year of ROI for CDI and MDM. "Last year there was a lot of tire kicking with vendors," Dyche says. "Now enterprises are convinced and vendor selection has begun." While the solutions will continue to be refined, according to Dyche, the real advancements will be made as CDI and MDM become business-driven initiatives. "The number of different relationships that an individual has with your company is directly proportional to the value of that individual as a customer," Dyche says. "If we want to drive the real value from CRM, CDI is the tool, and will help get us there."
Contact Assistant Editor Colin Beasty at cbeasty@destinationCRM.com.
Sidebar: CDI on Stage
Any enterprise hoping to excel with its adoption of MDM and CDI will need to step up the data management ladder, according to Jill Dyche, partner in and cofounder of Baseline Consulting. The model here, outlined by Tony Fisher, president and CEO of DataFlux, has four stages.
- At stage 1, the unaware point, an organization has few rules or processes in place regarding data management. Duplicate data exists in multiple databases, each serving different departments. The company has little visibility into its data management performance, and IT has taken the lead. The result is an enterprisewide failure to understand why problems exist or what impact they have. The associated business risks are high (lost customers and/or improper business procedures) while the rewards are low. Approximately 35 percent of companies are at stage 1.
- Stage 2 is when organizations begin to understand the importance of data management, but deal with data quality issues only as they occur. At best the business reacts to problems to mitigate the severity of outcomes. Tactical data management solutions, such as data profiling and ETL tools, have been implemented, but the organization lacks an integrated data management solution. Risks are still high while rewards are limited and mostly anecdotal. Forty-five percent of companies fall into the reactive stage.
- At stage 3 companies can avoid risk and reduce uncertainty when it comes to customer data. Data management plays a more critical role within the business and receives tangible value from more accurate data. Companies look beyond the horizon to understand the impact of data problems on mission-critical information. Data management initiatives resemble CRM implementations, with executive and managerial support and a data governance team guiding the rollout. Enterprise application integration (EAI), EII, and CDI tools are used to abstract and cleanse data automatically. "Data quality becomes an everyday part of life," Fisher says. Fewer than 15 percent of organizations are at this stage.
- At stage 4 companies reach the nirvana of data management. In the predictive stage data quality is ingrained throughout the company. Processes are automated and all business applications are connected to a central MDM/CDI repository that formats, cleanses, and redistributes updated data in real time. To reach stage 4, making the jump from stage 2 to 3 is the biggest hurdle a company will face, which is why only 5 percent of companies reside at stage 4. "This is where departments that have been operating independently have to come together. It's an enterprisewide shift in how customer data is handled." --C.B.
A New Leader Emerges in Gartner Magic Quadrant for MDM for Customer Data
Gartner Magic Quadrant for MDM for Customer Data '09: As vendors continue to invest in master data management initiatives, Initiate Systems joins IBM and Oracle in this year's top segment.
The Data Quality Conundrum
Fewer than half of all organizations have a defined strategy regarding customer data, according to research by Experian QAS.
Siperian Directs Business Users to Master Data
The MDM vendor releases a new application to empower business issues with data management.
Data Quality Best Practices Boost Revenue by 66 Percent
Research by SiriusDecisions also shows that faulty marketing intelligence has a comparably huge negative effect.
IBM's Vision Cited in Gartner Data Integration Magic Quadrant
Data Integration Magic Quadrant '08: Informatica and SAP-Business Objects join Big Blue as the leading integrators.
Still No Masters in Gartner's 'MDM for Product Data' Magic Quadrant
The latest assessment of master data management for product data covers 11 vendors -- and reveals no leaders. Is product data that much harder than customer data?
The Dangers of Out-of-Date Data
A new Forrester report points to more emphasis on real-time data warehousing -- and outlines five ways to go about it.
Forrester Wave Finds Initiate Systems and Siperian at the Center of Customer Hubs
In addition to those two vendors, Dun & Bradstreet's Purisma and Oracle UCM remain near the top of the research firm's latest report.
If MDM's So Great, Why Aren't Companies Mastering It?
Two recent reports throw a bit of cold water on master data management, despite significant benefits and business process gains.
DataFlux Has the Edge in Gartner's Magic Quadrant on Data Quality Tools
Magic Quadrant for Data Quality Tools '08: Business Objects, IBM, Informatica, and Harte-Hanks' Trillium Software join the SAS Institute unit in the top box.