CRM across the enterprise: integrating the channels.

Even the best CRM implementations will fail to meet expectations if the underlying data that feed the CRM system have inaccurate, duplicate data streaming in from all levels of the enterprise.

A centralized data repository is a critical foundation for any CRM initiative. It requires customer data integrating from a disparate set of databases located in various parts of the enterprise, the data then collected from varying sources and channels, as well as from a multitude of differing architectures and platforms.

However, the mere assembly of that data in one place is not enough. Effective cross-selling and upselling by use of an enterprise CRM system require not only the channels be integrated, but also the data cleansed, standardized and consolidated. Defective data in the form of bad addresses, non-standard titles, misspelled firm names and informal first names will result in duplicate account records and customer records.

Cross-selling demands that account records be correctly consolidated. Consolidation cannot occur if the data seemingly represent multiple customers because of erroneous addresses and other content. The cross-selling challenge does not end with proper cleansing; the consolidation process, if not performed correctly, can leave duplicate records, split customer records and lost accounts.

In attempting to deliver a centralized customer information repository, many enterprises severely underestimate the challenges of data extraction, transformation and cleansing. The resources that are required to perform the cleansing and consolidation of customer data depend on the size and complexity of the environment (i.e., the number and types of data sources, data volumes and prevalence of data quality issues).

"Although creating an integrated customer data repository is, in itself, a challenging task, merely assembling the data in one place is not enough," said Walter Janowski, research director, Gartner, Inc. "Clean data are a requisite for CRM success. Enterprises that fail to address data quality issues risk missed opportunities and operational inefficiencies in their CRM system."

Incoming Data

There are at least five distinct channels that commonly feed data into a CRM system, including:

* Purchased data or third-party lists;

* Legacy data migration;

* Call center data from customer service, and customer support;

* Paper forms and reports captured by account managers; and

* E-commerce sales transactions and activity.

If the data are inaccurate or out-of-date in any of these channels, an organization is at a greater risk to lose customers or sales opportunities. Also, with increased government regulations on corporate accountability and with privacy mandates such as Sarbanes-Oxley, the US Patriot Act, HIPPA or the "do not call" legislation, an organization risks heavy fines if it is not in strict compliance.

Within operational constraints, what are the opportunities to cleanse and consolidate data? In general, these opportunities fall during any of the following categories: transactional updates; operational feeds; purchased data; legacy migration; and during regular maintenance.

Transactional Data

This approach suggests organizations want to take a proactive approach to cleansing data. Organizations can identify the entry points of information into the organization--in this case, during transactions--as well as where flawed data exposure may be occurring. When a transaction is processed, organizations have a great opportunity to validate the data prior to saving that same data to an operational system. Transactional updating also affords the chance to validate data as they are created or as they arrive, rich with contextual information, in the information packet. This contextual setting is lost as soon as the data are sent down stream.

Operational Feeds

The second opportunity to cleanse and consolidate data is during operational feeds. These are regular monthly, weekly or nightly updates supplied from distributed sites to a central data store. A weekly upload from a subsidiary's CRM system to the corporate data warehouse is merely one example. Regular operational feeds allow an organization to implement batch-oriented data quality functions in the path of the data stream and around the entire operation. By their very nature, transaction updates force organizations to handle individual information packets as they become available, which implies slower processing and wider distribution of implementation. However, transaction-oriented cleansing is often implemented in conjunction with operational feeds. Transaction cleansing validates data going into the operational system, and operational cleansing validates data going into the next larger system, typically a data warehouse.

Purchased Data

The third opportunity to take advantage of data quality processing is when data are purchased from a third party. Many organizations erroneously consider the data to be clean when purchased. Not so. If organizations don't check the purchased list, they're essentially abdicating their data quality standards to those of the vendor. Validating purchased data extends to matching the purchased data to your current data set. The merging of two clean data sets is not the equivalent of two clean rivers converging into one; rather, it is like pouring a gallon of red paint into blue. In the case of a merge, 1 + 1 does not equal 2; oftentimes, it is 1.5 with "leftovers" because of duplication. What was separately clean is together dirty. The merged data sets must be matched and consolidated as one new, entirely different set to ensure continuity. A hidden danger with purchased data is it occurs as an ad-hoc event, which means no regular process exists to incorporate the data into an existing system. The lack of regularly occurring processes enhances the chance that data quality will be overlooked.

Legacy Migration

The fourth opportunity to improve the quality of data is during legacy migration. Any time data from an existing system are exported to a new system, the data must be robustly checked and validated. Frequently, organizations will discover the need for new data fields to cure field-over-use issues. For example, a manufacturing company during a data quality assessment discovered it (the company) had three types of addresses (site location, billing address and corporate headquarters), but only one address record per account. In order to capture all three addresses, it was duplicating account records. What this company needed to do was extend its data model to hold three separate address records for each account, which impacted the data model of the new system being built. It would not have known this had it not prior assessed (a critical quality function) its data.

Regular Maintenance

The fifth opportunity to improve data quality is during regular maintenance. Even if an organization starts with perfect data today, that data will tomorrow be flawed. Data ages, and it does so more quickly than most expect. For example, 17 percent of U.S. households move each year, and in some years as many as 60 percent of phone records change in some way. Moreover, every day people get married, get divorced, have children, have birthdays, get new jobs, get promoted and change titles. And if that isn't enough, the companies we work for start up, go bankrupt, merge, acquire, rename and spin-off. Organizations must implement regular data cleansing and consolidation processes, be they nightly, weekly or monthly. The longer the interval between regular data quality activities, the lower the overall value of your data.

As a relationship with a customer evolves, it becomes more and more difficult for a customer to establish the intimacy level with the competition that he or she enjoys with your organization. The result is an increase in revenue by enhancing customer satisfaction and thus retaining and developing the most valuable customers.

Integrating the channels in your CRM system isn't the only thing that ensures success. Organizations need to integrate "clean data" into the enterprise to more intelligently integrate the channels.

By Frank Dravis, Firstlogic, Inc.

Frank Dravis is vice president of Information Quality and Practices for Firstlogic, Inc., headquartered in La Crosse, WI. He has 16 years of experience in information technology and software development, with the majority of his experience centered in solutions design and implementation. Frank can be reached at 608-782-5000, or [e-mail]