Gotcha! Creating a Single Customer DatabankBy David F. Carr | Posted 2004-11-01 Email Print
How Real-World Numbers Make the Case for SSDs in the Data Center
What to look out for when combining databases.
A merger virtually guarantees that the combined entity will possess overlapping and conflicting customer databases. Great care must be taken in melding databases so that companies wind up with a single, current set of accurate data on each customerand one clear architecture for dealing with changes and additions. Customers are often frustrated and bewildered, if not infuriated, when the company they've been doing business with for years can't even find a record of their existence, or gets basic information wrong. And it doesn't matter how many different divisions you have. To them, you're "the bank."
Problem: Combining databases can result in duplicate and inconsistent data.
Resolution: Invest plenty of time in producing clean customer records. One way to weed out duplicate or inaccurate information is to match customer records against public databases such as Dun & Bradstreet's (for businesses) and Experian's (for consumers). For example, by determining all of the business customers whose headquarters address differs from the D&B listing in one or more of your databases, you can flag cases where the address needs to be corrected.
Problem: Combining all customer information in a single database is often impractical, because parts or all of some applications that store data would have to be rewritten.
Resolution: Synchronize files through a "master customer database." This is a central repository that keeps clean copies of customer data, and exchanges updates with whatever databases are in use. Customer data integration (CDI) products, such as Oracle Customer Data Hub, can help with this process. CDI tools are programmed to coordinate databases and reconcile differences.
Problem: It's not easy to decide which product to use for integration.
Resolution: Think about what you have and what you need. If you're committed to an enterprise resource planning (ERP) software vendor, a tool from that company may be best to simplify development and improve integration of data.
But you may have some databases you use and need that are from other vendors and don't play well with your core programs. In such cases, you may need general-purpose tools for cleaning up and integrating data.
Siperian, one such provider, says its advantage is offering tools that don't impose a structure that is fixed in place to work with one vendor's ERP system. Generalized tools also can be set up to automatically reconcile most conflicting customer data, while leaving room for human intervention as needed.
Problem: Putting all your information in a warehouse is not a magic bullet.
Resolution: Be careful how you employ data warehouses. These are central repositories that also gather customer data from multiple systems, but they are designed to collect data at specific times, crunch numbers to spot trends and produce reports. That's useful for executives making decisions, but not necessarily up to speed for a live customer service center. Better to use a master customer database to push data in and out to your call center, Web site and back-end billing systems.