Analysis & Commentary:

MAKING CRM WORK THROUGH CUSTOMER DATA INTEGRATION
by C. Alex Dietz, Acxiom Corporation

Practically everybody in the data technology business is talking about Customer Relationship Management (CRM). It's not just important, it is the single most important strategy for any enterprise. If CRM is not at the top of your list, I have to ask: What is?

Q. What generates revenues and profits for your company?

1. The customer.
2. What is the most important asset of any enterprise?
3. The customer.
4. How do you manage those critical assets?
5. CRM.

So, everybody is doing it -- or everybody should be doing it -- but very few companies are doing it well, and I can tell you why. Let's start looking for the problem by examining the dimensions of every CRM strategy. Each dimension represents a critical component of the CRM equation, but technological advances have improved some dimensions more than others.

One dimension is the database dimension:

This is where customer data is loaded into a central database repository. The design, creation and ongoing maintenance of the CRM database is a complex task that involves designing data models and populating the database with the appropriate information to drive marketing strategy. The final repository, in fact, may not be a single database, but may pull information from multiple databases. Today's database technology is highly evolved to accommodate the needs of CRM, so the problem with CRM isn't found here.

Another dimension is the application dimension:

This is what the end-user-the customer-facing organization-uses to execute CRM. This is typically not one single application, but rather a menu of software systems that delivers different CRM functionality. There may be analytical and data mining tools to help analyze customer segments; count and query tools to size market segments; business intelligence tools to drill down into the financial performance and profitability of CRM initiatives; and campaign management tools to manage complex marketing campaigns. In short, these software systems deliver the information that drives business strategy for the enterprise.

The level of sophistication in the CRM applications area is state-of-the-art. Virtually all of the hardware and software technology advances of the past 15 years have been incorporated into this component of the CRM process. The software is intuitive and easy to use, yet delivers incredible functionality. If there is a weak link in the CRM process, it isn't here.

Could CDI Be the Problem?

Which brings us to the basic, but too often ignored, dimension of CRM-Customer Data Integration (CDI). This is where all of the data sources that will be used to create a single view of the customer are brought together. The core technology layer of a CDI solution includes the hygiene, linking and grouping of customer data as well as the recognition processes necessary for presenting integrated customer data where your business needs it.

But CDI technology hasn't evolved much in the last 15 years. Yes, there are faster computers, faster networks and better data storage devices, but for the most part companies are still performing CDI processing the way they did in the 1970's. CDI still relies on traditional merge/purge software that has not changed dramatically in 20 years. Traditional string matching software is still at the heart of CDI solutions, and that's just not very efficient by today's standards.

For the most part, CDI is still a relatively expensive, large-scale batch process. It is difficult for small- and medium-sized companies to implement an effective CDI solution. There are really only two options for most companies: Develop all the necessary systems and expertise in-house (something that even the largest companies have had difficulty doing) or work with a CDI expert like Acxiom Corporation. Most large companies have tried the first option and, after much frustration, have now turned to the second option.

Data Quality: One Key to Success

Data quality is the overall quality of the data content upon which a company bases its marketing and customer relationship decisions. There are four components to data quality:

Data completeness: This is the percentage of all possible data sources a company has integrated into its decision-support processes. If there is data about customers in 12 different operational data stores, then all of those sources should be integrated into the CRM system. However, this also refers to sources of outside data that may be available.

Data accuracy: The overall accuracy of the data content (contact information and predictive information). Once a company has all of the sources identified, then how accurate is that data content? Is the contact information accurate? In most companies this answer will vary considerably with each individual source.

Grouping accuracy: The accuracy with which a company can consolidate data from disparate sources for decision-support purposes. Once a company has all of the data sources identified and has performed the necessary hygiene to ensure that they are accurate, how well can the company identify and group multiple occurrences of the same customer in order to provide a comprehensive customer portrait?

Data access: The speed with which a company can integrate its data into its decision-support processes. Because if all of the activities described in the first three components take months or weeks to complete, a company still won't be driving its CRM strategy with the latest information. Ideally, the access to the data should be "real-time."

Theoretically, a company can represent where it is in a "data quality continuum" by mapping each dimension into a four dimensional cube, the point being to constantly improve the company's position in the data quality continuum. Each improvement will directly influence the effectiveness of a company's CRM strategy.

In fact, I believe the way a company manages the data quality in its enterprise is by using a set of processes called Customer Data Integration. This means the creation of a "comprehensive customer portrait" with these CDI processes is the most complex and, in many ways, the most important component of overall CRM strategy.

Creating the Customer Portrait

Customer Data Integration is necessary for your CRM strategy because it helps determine two things:

• First, your entire relationship with your customer. You need to know all of the purchasing behavior from all of your product lines in order to complete an accurate picture of your customer. However, you also need to know other information about marketing campaigns and customer service interactions in order to understand the ROI for each customer.
• Second, you have to be able to assemble other demographic and predictive characteristics about your customer. This kind of information will typically come from external data sources. Information such as demographic and lifestyle information help paint a portrait of your typical customer, your best customer and your worst customer.

The internal data about your customer is roughly 75 percent of the view. But it's not a complete view until you include external marketing information. How do you do this? You can either develop all of the processes internally to do it, or you can work with a CDI solutions provider like Acxiom. Solutions providers can help you move quickly up the continuum of data quality by the use of CDI products and expertise.

Test Your Data Quality First

CDI solutions providers should start their work with you with data diagnostics, like Acxiom's OpticxSM, which can quickly give you a snapshot of how good your customer data is.

For instance, in one Opticx test on a client's data, we found that almost 64% of the records were good -- but there were over 36% of the records with data anomalies. Obviously, the client needed a data cleansing or data hygiene operation. In this case, the data hygiene improved about 16% of the undeliverable addresses. There were still about 20% with anomalies. But we found that many were people the company didn't have addresses for. So, the company could apply Acxiom's InfoBase BestAddress and update the records.

The most important component in a CDI solution is "linking technology." Linking technology provides dramatic improvements in grouping accuracy and enables integrated customer information to be presented at the point of contact. Acxiom CDI software that incorporates linking technology is AbiliTec.

The Advantages of Linking Technology

First of all, linking technology significantly outperforms old CDI technologies. In the past when we were tweaking our traditional de-duplication technologies, we would be thrilled if our enhancements created a 1/2 percent better de-duplication rate or a 2% overall performance improvement. Linking technology creates a minimum de-duplication improvement of over 5%, and in some cases, we have seen improvements of 20% over conventional technologies. Perhaps more importantly, the efficiency of the CDI process has improved an order of magnitude -- that's a ten-fold increase in performance.

Linking technology also enables customer information from a variety of internal and external databases to be pulled together in real time. This instantaneous 360-degree view of the customer can then be used with appropriate business rules to "create" the proper experience for the customer when they suddenly pop up at one of a company's many touch points.

Linking technology can be implemented with minimal changes to existing legacy systems. A cross-reference database can be implemented to tie customer instances from different internal and external data stores together with no changes to the existing databases.

How It Works

So there is no doubt that linking technology is a real break-though, but how does it work? At the heart of the AbiliTec is Acxiom's knowledge base. This is a large repository of both consumer and business names and addresses that is created from many sources and continuously maintained and updated.

The fundamental idea behind the knowledge base concept is that historical occupancies and entity representations are maintained and tracked. In many ways it is a temporal database that keeps track of how consumers and businesses change over time. The knowledge base also contains associative information as well as business rules that help link occupancies together. At Acxiom we have developed a number of proprietary data mining strategies to help us identify linked occupancies.

The knowledge base houses the information used to create and maintain consistent links for each name and occupancy in the knowledge base. As new information and sources are added to the knowledge base, all changes affecting representations of names and addresses are incorporated into the business rule structure that leads to the creation of AbiliTec Links.

Creating and maintaining the knowledge base and building AbiliTec Links from it is a highly specialized activity. It is unlikely that most enterprises will have the resources or the inclination to develop this capability internally. No one we have discussed these concepts with is tempted to undertake this process. It makes sense to work with a Customer Data Integration solutions provider like Acxiom to take full advantage of the expertise we use to take enormous amounts of information and distill it to a tool that can link together all of your disparate silos of customer data without having to redesign your infrastructure.

One important point to note about CDI using linking technology: No new information can be discovered from the AbiliTec Link. That means that when a name or address is submitted to Acxiom for AbiliTec processing, links are returned. If you don't have an occupancy, then you cannot retrieve links for that occupancy. Also you cannot retrieve any historical information or data from AbiliTec. It is strictly used for maintaining the links necessary to perform CDI.

Characteristics of Linking Technology

Here are things you should be thinking about as you prepare to implement CDI using linking technology. First and foremost, the technology used to perform customer linking, matching and grouping must be built from a knowledge base. Without a knowledge base driven CDI solution, CDI results will not be satisfactory because they are at best derived from an inexact algorithm system for matching, not historical data linked with current information.

Your CDI solution must be able to perform linking for customer information on both consumers and businesses. At Acxiom, we have delivered billions of links since AbiliTec went live two years ago, and only a very small percentage of the files contain only consumer links or only business links. Most files are a blend of consumer and business records.

The CDI solution must be global in scope. Acxiom is constantly expanding its capability to perform CDI for its customers on a global basis, all built from the AbiliTec platform. The point is that the Internet has created a global economy and almost everyone has a need for a global CRM strategy.

Isolating the Problem

Earlier, I asked the question: Why aren't companies doing a better job with CRM? It's obvious to me that the problem is found in not beginning the CRM process by assessing data quality and the need for Customer Data Integration. And it's just as obvious to me that there is now a state-of-the-art technology to solve the CDI problem, AbiliTec.

CRM is the most essential business process in any enterprise. Yet the ability to execute an effective CRM strategy has been elusive because of the inadequacies of traditional CDI technologies. In fact, all aspects of CRM have benefited greatly from the technology boom of the last 20 years with the exception of CDI.

Now, however, the emergence of linking technology such as Acxiom's AbiliTec has created the potential for a quantum technological leap forward in CDI. In fact the demands of real-time CRM requires a CDI linking technology to work.

There really is no effective way to do real-time CRM without employing a CDI linking technology. And, in today's changing marketplace, real-time CRM is a requirement for success.

As Leader of the Products Division for Acxiom Corporation, C. Alex Dietz was directly involved in the development of AbiliTec, Acxiom's CDI software powered by patented linking technology. AbiliTec has been adopted by several major companies to help them create a single view of their customers across the enterprise. Alex has served in a number of positions within Acxiom, including president of the National Marketing Services Division, manager of Acxiom's Advanced Systems Division, and chief information officer. He graduated from Tulane University in 1965 with a bachelor of science degree in electrical engineering. Before joining Acxiom as a vice president in 1970, Alex worked as a systems engineer for IBM.