Unwrap the Promise of Your Data (939 words)
Strategies for building a quality database
By Don Hinman
As a data consultant, I spend much of my time answering questions about customer information and how it is used to improve revenue and business operations. But with all the advanced data products available today, the question I hear most often is: "How do I build a quality database?"
This question illustrates data's current place in the corporate world. For many corporations, customer data are partially unwrapped presents under the Christmas tree—so promising, so big and impressive. So intimidating. Yet many companies can't take advantage of the data products on the market because they don't have the framework they need: a quality database structure, thoughtfully prepared by professionals.
Here, then, is a guide to establishing a useful, comprehensive database that will grow with your company.
Where Did this Stuff Come From?
The first step in developing a new database is to ask about the purpose of the database and the sources that feed it. What lines of business are represented in the database? Is it used for prospecting, customer retention, or both? Is it used for more than marketing purposes? What sources are used to update it, and where are those sources located? Do I expect it to represent a single customer view of all activities from my customers?
Over the years, I've worked with companies in the insurance, retail, media, publishing, telecommunications, automotive, catalog and financial services businesses. But regardless of the industry, the first issue we run into is generally the quality of the input data. They can be extremely "dirty," and you must have the processes and procedures in place to identify and repair those mistakes.
The next problem is parsing. Data elements such as name and address get jammed together in a string, and it's difficult to separate them correctly. The high-tech industry uses a good deal of data submitted through Web sites, and they are the worst data available because not only are they hand-keyed by consumers, they also are full of made-up names and addresses.