A Customer Need Not Have So Many Faces
by Dr. Fern Halper, Partner
What is a customer? The answer may seem obvious, but in reality, identifying a customer from company data can be very difficult, a nightmare, in fact, especially for companies in the B2B space. A common customer view is critical for customer service, billing, and marketing, yet it is often elusive.
The Many Faces of Customers
Why is a customer so elusive? Consider the following scenario: A telephone company is selling a number of services to their business customers including: wireless, wireline, VOIP, and Internet services. In addition to providing these customers with a single bill and the right level of service, they also want to be able to analyze the data related to these customers in order to sell to, market to, and service them more effectively. But, guess what? It isn?t as easy as it seems. Why? There are a number of reasons:
- Information about a company is often stored differently in different systems and the relationships can be complex. Let?s use the case of a company called Computer Services Limited:
- In the telephone company?s CRM system the customer may be referred to as Computer Services Limited. In the company?s sales system the company is referred to as Computer Svc. Ltd. And, in the marketing database the company is referred to as CSL.
- Computer Services Limited is actually a subsidiary of Computer USA. How does the telephone company want to deal with this relationship?
- Contact names and addresses may be different, across systems. Trying to understand who belongs where in a corporate hierarchy may be difficult. Additionally, each system may view the customer differently. For example, a CRM system may view a customer as a personal contact. A billing system may view the customer as an address or a phone number.
- In addition to representing information differently across systems, most systems will also have different customer identifiers for the same customer account.
- External data, such as demographic data, may be hard to map to in-house data.
In fact, a rough estimate is that there can be a 30-40% mismatch simply trying to match names across just a few systems. This is often due to the fact that some systems, for example ERP systems, can?t deal with matching slightly different names, so whole new accounts are created for customers that already exist in the system. Typically, companies find that they are faced with issues of one to many, many to one, and many to many relationships when they try to sort through their system data and identify a customer. Often, companies create custom code to deal with this problem, but still find the problem unmanageable and costly.
The Purisma Solution
Purisma has developed a patented software solution to unravel and manage complex customer identities. The solution utilizes fuzzy logic and rules to enable a company to determine a customer identity. How is this done? Purisma has developed a registry, called the Purisma Customer Registry to do this. The registry consists of three parts:
- The Correlation Engine. Purisma utilizes web services to connect to disparate systems and pull customer data from these systems. Advanced pattern matching algorithms (fuzzy logic) are applied to the data along with an iterative inference engine that fires off rules and recognizes and determines relationships. When matches are found, the customer information is put in the Customer Registry. The Registry creates a unified ?master? for that customer. This is a cross index that links together all of the information about that customer from all of the systems where this information is stored. This is an iterative process, where relationships are recognized and determined, rules are tweaked and more relationships are recognized. The Purisma solution can also link information across corporate hierarchies. So, using the example above, the software can link together information about Computer Services Limited and Computer USA.
- The Data Stewardship Module. This module is used to visually display customer identity information. It can be used in several ways. In cases where it is difficult to recognize and establish a relationship (for example, in a many to many situation), the data can be viewed and manually corrected. Or, if companies simply want to review their data and make modifications to what the automated correlation engine did, they can do this here.
- The Integration Module. This module enables companies to integrate to third party Dun & Bradstreet data services. It also utilizes web services to do things like notify an ERP system not to make a duplicate record if one is already found in the Customer Registry, or integrate into ETL and EII or EAI technologies.
The software is Unicode compliant and supports languages from 50 different countries.
The Value of Unraveling Complex Customer Identities
Hurwitz & Associates believes the value to companies of being able to deal with complex customer identities is high. Benefits include:
- Improved Customer Service. A typical problem that companies have is not being able to service their high value customers effectively because their systems can?t distinguish who these customers are because there is no mapping of customer information between systems. The Purisma solution will help to solve this problem.
- Lower Costs. The Purisma solution helps to decrease costs associated with custom coding and at the same time reduces errors ? another costly proposition.
- Increased Revenue. Better customer data can also help companies understand the customer as a whole, which means that companies can market and sell to them more effectively
- Better Data for Analysis and Decision Making. The master data created by the Purisma solution helps companies to get better quality data for their analysis. Better analysis leads to better decision-making and can have a significant top-line impact.
- Better Data for Compliance Purposes. A single view of the customer will help companies meet their compliance obligations.
Determining customer identities is a problem that has been around for years and it?s good that vendors are addressing the issue – at last. A customer need not have so many faces, or if they do, at least the faces can now be tied to one identity.