Successful CRM Hinges On Data Quality

With increased competition in online and offline marketplaces, organizations are turning to CRM to attract and retain customers. According to a recent study by Gartner Group, worldwide spending on CRM licenses and services is expected to reach $76.3 billion in 2005, more than tripling the $23.3 billion spent in 2000. In spite of this major investment in CRM, organizations continue to miss sales opportunities and to lose valued customers due to corrupt data.

A comprehensive customer data quality (CDQ) management program is the key to ensuring accurate customer information throughout the enterprise. The components of a CDQ program include data conversion and cleansing, continuous data quality maintenance and the sharing of updates throughout the enterprise.

Why Do CRM Projects Fail?

CRM initiatives can fail for a variety of reasons. However, the silent killer of many CRM projects is the organization's failure to address the most essential component of the initiative: accurate customer information. According to a recent Gartner Group report, bad data is the number one reason that CRM projects fail. The consequences can be costly: poor data quality compromises the organization's total CRM investment and it undermines the organization's relationships with customers.

Ensuring accurate customer information is no easy task. Customers are dynamic, as is their information. They move, marry, divorce and have changing relationships with other customers at both the individual and organizational level. Customers also have changing product and service needs throughout their lifetimes. And in the midst of all this change, they expect their providers to deliver personalized service and offerings.

To anticipate and manage these varying customer needs and expectations, CRM applications need to present a unified, real-time view of the customer across the enterprise. To achieve this objective, organizations must develop a comprehensive customer data quality (CDQ) management program designed to ensure that not only is basic contact information up to date but that more complex and, even, non-intuitive customer relationships are identified and represented in the CRM application.

Establishing a Customer Data Quality Management Program

Best practitioners of CDQ combine vision, technology, culture and business rules to cultivate more intimate and relevant relationships with their customers.

Establishing a successful CDQ program requires more than just applying the right data quality software to the problem. It is imperative that organizations first identify and resolve any underlying business or cultural issues impeding customer information management. Prioritizing objectives to be achieved with the CDQ program is the next order of business.

When those fundamental matters have been adequately addressed, an organization can confidently embark upon the implementation of a CDQ program, which is comprised of the following four components:

• Discovery & Analysis
• Data Conversion & Cleansing
• Data Quality Maintenance
• CDQ in Enterprise-wide CRM

Discovery & Analysis

To implement a CRM application, data migration must take place. This involves consolidating data from various enterprise source systems and databases and mapping it to the target CRM system.

Too often, organizations undertaking data migration rely on inaccurate metadata and out-of-date documentation, resulting in design specifications based on erroneous assumptions about the source data. Data migration based on flawed mapping specs is akin to building a house on a shaky foundation. Each successive phase of data migration - including extraction, cleansing, matching and householding - that builds on the faulty mapping specs is inherently flawed.

Unfortunately, the resulting data problems don't manifest themselves until the testing phase, when it's too late. Data analysts are then forced to go back to the discovery and design phase, adjust the assumptions, and repeat the entire process. As a result, data migration projects can spiral out of control, wasting precious time and resources.

Clearly, early detection and correction of problems in the source data can drastically reduce the risks associated with data migration. The CRM project can then be planned and executed with accuracy and confidence . As well, project costs can be managed more effectively: an error detected during the project-testing phase can cost up to 100 times more to correct than the same error found during the design phase .

To handle source data issues upfront, organizations need to employ a powerful, automated discovery and analysis tool that provides a detailed description of the data's true content, structure, relationships and quality. This kind of tool can reduce the time required for data analysis by as much as 90 percent over manual methods.

Data discovery and analysis allows organizations to understand whether their source systems are account-based or customer-based and whether customers have been properly identified as individuals or organizations. It is also important to know if more than one record exists per customer, and if so, how to identify the best record. Or better yet, how to combine the best elements of each record into a new, consolidated record.

Having determined the true nature of the underlying source data, the organization is ready to create the necessary data quality transformation rules for the data migration to its CRM system.

Data Conversion & Cleansing

At this stage, the organization is ready to identify and correct formatting errors in customer and product information. This information can range from basic contact information to tax ID numbers and product numbers to any other information the user may wish to correct. Misspellings, transpositions and other anomalies can be amended, as well.

Data quality also includes the identification and elimination of duplicate records for individuals, corporations and households. But more than reducing duplicate records, this kind of data matching or householding is a powerful way to establish highly sophisticated customer networks. By linking customers who share a name, address, account number or other user-specified commonality, organizations can begin to ascertain a customer's total sphere of influence and to formulate a complete view of the customer's wealth and potential for wealth. This type of householding can include linking individuals to other individuals and individuals to corporations.

For example, a financial services institution implementing a marketing and service campaign targeting its premier clients would want to understand the customer's entire relationship with the bank, including personal accounts, business holdings and any custodial or trust accounts associated with the customer. This total view of the customer's network provides the institution with a much richer understanding of how best to market to and service this premier client. Without understanding the extent of this individual's relationship, the institution would likely not target this client as part of its premier campaign, losing the opportunity to realize the benefits of its CRM strategy.

Building more sophisticated networks also provides manufacturing organizations with a better view of their business partners, as well as their customers. Corporate householding allows the manufacturer to build hierarchical relationships with the different business entities of a single supplier. For example, Acme Manufacturing may make purchases from Division A, Division B and Division C of parent company Global Supplier. Yet, each division, as well as the parent company, may be based in different locations and do business under different names. If the manufacturer can link the divisions to the parent, it can begin to understand the total amount of business it transacts with this single corporate entity. This puts the manufacturer in a much better position to negotiate volume discounts with the parent company. These discounts can then be applied to the various division-level purchases.

In the case of automotive manufacturers, establishing the customer's identity can be a complex proposition. Customers can be the dealers who sell the cars or the individuals who buy the cars. However, there is a hierarchical relationship in these customer networks that the automotive company cannot afford to overlook: the end-client "belongs" to the company, as well as to the dealer. To establish effective marketing and service strategies for either the dealer or the end-client, the automotive company must first understand the relationship that exists between the dealer and the end-client, as well as the relationship that the end-client and the dealer each share with the automotive company. These are sophisticated relationship networks that go beyond linking on common name and address.

Implementing the necessary conversion rules for correcting errors and establishing customer networks requires more than software. It also requires an understanding of the organization's business rules. Only with that knowledge is the company ready to use software that includes built-in business rules for data quality processing. These pre-configured solutions can save an organization substantial time and money when compared to the more laborious and error-prone method of manual cleansing or when compared to software that requires the organization to build all of the necessary business rules for conversion from scratch. By taking advantage of intelligent solutions that provide comprehensive sets of universal data quality transformation rules, organizations can focus on building only the rules that are truly unique to their business. This approach allows organizations to accelerate the implementation of their CDQ programs.

Data Quality Maintenance

Best practitioners of CDQ won't stop once they've completed the initial data analysis, data conversion and data cleansing. Ongoing data quality processing is necessary to maintaining the integrity of any CRM system. Changes are constantly being made and new data is always being introduced into the system from various sales channels, including and especially the Web.

The Web presents a special challenge to data quality maintenance because the responsibility for data input lies more with the e-customer and less with the organization. According to the U.S. Department of Commerce, in 2000, more than 58 million U.S. consumers engaged in online transactions, generating $28.5 billion in sales. As e-commerce continues to grow, organizations committed to CRM increasingly will need to ensure that data is clean as it comes into the organization from the Web.

Organizations that value their customer information will place a data quality filter at all customer interaction touch points, including the Web. This filter is the organization's defense against customer data corruption. After the CRM system is populated with cleansed and linked data, organizations will want to focus their preventative data quality measures on the front-line, rather than on more costly and time-intensive back-office clean-up. Just as business rules were critical to the conversion, data quality filters must be flexible and robust enough to support the organization's established business rules. This consistency will ensure the organization's data quality conversion efforts are maintained going forward.

CDQ in Practicing Enterprise-wide CRM

It is important to understand that an organization's CDQ program is not just a subset of the CRM effort. CDQ is the foundation of the organization's CRM strategy, which must go beyond the CRM application system and span the entire enterprise.

Enterprise CRM strategies are successful only when organizations understand that CDQ principles must be applied to a customer data repository that is designed to feed -- and be fed by -- the enterprise. This customer data repository is aligned with the CRM application system, and bi-directional updates between the two systems are synchronized to ensure that the representatives at the front-line have the benefit of the latest, enterprise view of the customer. Likewise, the enterprise is now positioned to leverage the valuable information about its customers that is captured and stored in its CRM application. Without synchronizing the CRM application with other enterprise systems via the customer data repository, organizations run the risk of creating the very problem they were attempting to solve: silos of customer information distributed throughout the organization, undermining the organization's ability to understand the total relationship it shares with its customers.

A singular customer focus, supported by a unified view of the customer across the enterprise through a CRM system, was supposed to enable organizations to communicate personally and cost-effectively with their customers and prospects. This would allow companies to build customer loyalty, expand market share and increase sales. Huge investments in CRM applications were justified on the basis of what they would contribute to higher revenues and profitability. Six years and hundreds of millions of dollars later, many organizations are still waiting for measurable results.

The most critical factor to the success of CRM initiatives is accurate customer data. Data entering the CRM application from every customer channel must be cleansed, initially and on an ongoing basis. For the entire organization to benefit from CRM, enterprise-wide CRM must be practiced by synchronizing the CRM application with the rest of the enterprise systems. Additionally, there should be as much focus on the "R" of CRM as there is on the "C." Understanding the relationship between the customer and the organization is fundamental. Understanding the relationships shared between the organization's customers, both individuals and corporations, gets to the heart of CRM's true promise. By implementing a strong customer data quality management program, organizations can begin to realize this vision of CRM.

R. Jeffrey Canter is executive vice president of operations at Innovative Systems, Inc.