As a worldwide leader in the manufacturing and marketing of network equipment, Bay Networks, Inc., sells frame and ATM switches, routers, shared media, remote and Internet access solutions, IP services and networking management applications. Their customers draw from corporate enterprises, service providers and the telecommunications industry. The company's 250,000-record customer data warehouse consists of customer names in two portals: a SAP application and a networked EDI system that feeds point-of-sale information into the Oracle database via resellers.

Until very recently, these 250,000 customer records--costly to procure and intrinsically valuable to nearly every department within Bay Networks--were inaccurate and, therefore, not very usable. When various departments, such as finance, sales and marketing, attempted to extrapolate data to generate reports, they had to manually look up the data, which was costly and inefficient.

Patrick Firouzian, architect of the customer database warehouse within Bay Networks' information services department, and his staff struggled with five issues: inconsistent customer data across several databases, naming standards, spelling variations, lack of knowledge about customers, and the inability to generate accurate reports for finance, sales and marketing.

One of the most costly problems was the high percentage of duplication and erroneous records that were entered into the system. "About 500 new customers are entered each month," Firouzian says. "Of those, 15 percent should not be there." Not only was the customer database inaccurate, it was a hindrance to learning more about Bay Networks' customers. "Without a clean database," Firouzian explains, "we couldn't utilize bureau tools such as Dun & Bradstreet's database information." Furthermore, the inability to clean up customer records as they were entered into the system meant that data mining and electronic commerce were harder to implement.

After struggling with in-house customized scrubbing and cleansing routines, Bay Networks decided to shop for data cleansing vendors. "We thought it would be a big savings to let an outside vendor maintain the code instead of doing the code ourselves," he explains. "Creating an application that is already written isn't cost effective."

Firouzian and his team of six developers tried a number of solutions, including Firstlogic, Inc.'s i.d. Centric product suite. With the help of i.d.Centric's sales and technical team, they set up a prototype and ran the data cleansing program for a two-month trial period starting in October of 1997. In December they signed the contract.

Bay Networks identified i.d.Centric's technological advantages over other tools. "The suite has libraries we could integrate into our processing streams to cleanse customer information as it comes into the system. We enter customers into our system via our SAP application at multiple entry points--some on-line, some batch oriented--and i.d.Centric enables us to cleanse in both situations."

Particularly progressive, Firouzian says, is i.d.Centric's pick, or selection, list feature. "If you are looking for a customer but don't know exactly what they are called or what the address is, this pick feature gives you a list of customers that are similar to the one you entered and you choose from that."

The Implementation Period: Full of Surprises

Firouzian's team implemented the entire i.d.Centric software suite of data cleansing tools: Address Correction & Encoding (ACE), ACE Library, ACE Canada Library, DataRight, TrueName Library, Match/Consolidation and Match/Consolidation Library. The implementation took five months. "This was not because of the tools," Firouzian explains, "but because it was a lot of work to integrate everything together. All the different systems in the company had to be ready to cleanse their customer database. The commissions system, the channel capture system, SAP--all had to adapt to a new architecture to use the services."

The biggest hurdle, Firouzian notes, was developing communication processes among all the departments. "It is a lot of interfacing--finding out what each department wants, then making them understand how they can use our cleansing services. We also had to test the system and then educate and train the people."

Despite the adjustments required of each department, Firouzian's team was greeted with a surprisingly positive reception. "As soon as we announced we essentially had a clean database for all the customers of Bay Networks, the different departments wanted to immediately utilize it for cleansing their customers and pooling data for their systems." Firouzian admits that he had not expected such a positive reaction.

The i.d.Centric tools helped revamp the database by consolidating duplicate records and correcting name and address spelling variations and inaccuracies. For example, many of Bay Networks' customers within the database had fallen victim to inconsistent entries. AT&T also might be found as ATT. US West could be duplicated under USWEST. i.d.Centric also enabled Bay Networks to group all its customers who had many divisions and addresses under the proper parent company, ending what had been a nightmare of confusion for the sales and finance departments.

Firouzian estimated that the cost of the i.d.Centric product suite and that of reconfiguring the system architecture of each department would pay for itself within one year through savings realized in the following areas:

FINANCE: "We'll do a much better job of paying commissions to the right people."

SALES: "We'll have better reporting, using features such as vertical segmentation of customers, parent/ child reporting and headquarters identification reporting."

MARKETING: "Our department can now do very focused marketing campaigns by using standard industry codes. We're now aware of the true customer base, rather than wasting mail on companies that are not in business anymore."

IT DEPARTMENT: "Everyone will be using the database directly, so there will be fewer requests from users within the company regarding data quality. They can focus on things that will be adding true value."

DATA WAREHOUSE: Being able to clean up the duplicate and erroneous data entries, Firouzian estimates, "is a huge gain as far as the quality of reporting in the data warehouse. It's a long-term savings, difficult to quantify. This impacts sales awards, commissions, performance analysis and marketing."

DATA MINING: "It's difficult to say how much return this will bring in, but it's definitely easier to do than before."

ELECTRONIC COMMERCE: "Uniquely identifying customers and accurately representing them will help the electronic commerce at Bay Networks."

As for better knowledge of the customers, Firouzian says, "Dun and Bradstreet's product can now be implemented corporate wide. Every department can learn more about the customer. Sales people, in particular, will better target their customers."

The i.d.Centric implementation was designed for scalable growth. "We have 250,000 true Bay Networks customers," Firouzian says. "But we also have 800,000 prospective customers. Thanks to i.d.Centric, all this is clean data, too."

The importance of the i.d.Centric cleansing operation, Firouzian says, in some ways is immeasurable. "Customers are what allow a business to grow. If you have a coherent, consistent customer database, everyone is going to use it, depend on it and make decisions based on it. Now that we have raised expectations company wide, we could never go back to the way we were before."

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access