To order Alpha Four or Alpha Five, click above to go directly to Alpha, or call 800.451.1018 or 781.229.4500
|
|
|
|
|
|
Tuesday, October 1, 2002
|
|
VOLUME 1
ISSUE 8
|
|
|
|
|  |
 |
 |
Data Integrity
by Geoff Hollander
From your first saved operation to writing a complete application in XBasic, if you clock in with Alpha Five, you're turning into a developer. It's a gradual and insidious process that, happily, is inevitable, and often so absorbing that it's easy to forget about its raison d'etre. In English, that would be your data.
Data is unprocessed information either input through data entry, one keystroke at a time, or through a data exchange where existing data is imported, or appended into a table established to receive it, and as such is typically not usable until structured and organized into information by programs such as Alpha Five.
Unlike word processors, where processing principles are virtually the same for most documents regardless of the content or its format, in a database the marriage of the tool to the raw material with which it works is virtually as important as the tool itself.
High quality, clean data is more than just a "pretty face." It affects every aspect of your database applications from performance to accuracy. Poor data quality can not only inhibit your database operations in much the same way using low octane gas will compromise the performance of a race car, but disregarding your data's quality is a sure way to compromise your database's veracity and, worse, those business decisions you will base on what it tells you.
Data quality can be distilled down to three basic criteria:
- Alignment - The same data, consistently representing the same attribute of each record is always stored in its assigned field throughout the table; First Name fields don't contain prefixes; Company fields don't contain Addresses (and visa versa); Phone Numbers either include area codes and extensions or they don't, storing them in other fields.
- Consistency - All data elements in each field are standardized - formatted, written, abbreviated, punctuated, cased - uniformly throughout the database. For example Address fields should conform to the USPS Addressing Standards and CASS coded for accuracy. (This isn't just to save serious dollars in postage, but also helps you identify duplicate records and determine geographic parameters, which can later be used in geocoding latitudes and longitudes, census information and other kinds of data enhancement used in targeted marketing campaigns.)
- Accuracy - Data is complete, accurately entered and is currently true, where the entity it represents is concerned. Again, the most obvious example of this is addresses. A person's address information may be clean, accurately written and has CASS-qualified - establishing that there is such a location; but is it still the correct one…
It may seem nit-pickesh at first take, but if left unchecked, these things - separately or in concert - will cause your queries, reports and mailings to become increasingly flawed and misleading by, in effect, confusing your database.
If your database is also your mailing list, then these problems start costing you money almost immediately. The USPS throws 35% of all bulk mail sent in the US each year right into the trash. (By postal regulation, incorrectly addressed bulk mail it discarded - unless Address Correction is specifically requested, its required specifications met and return postage is paid by you for every piece returned ($.50 and up per piece). Even if the address information is clean, formatted correctly and even passes USPS CASS Certified Coding standards, is the person to whom you are sending this mail piece still there? At this point, from the USPS point of view, it doesn't matter whether the name and address were written incorrectly or the recipient no longer lives or works there; it's now trash. At an average price per piece of about .35 to .40 cents (including postage), how many can you afford to send to the Bulk Mail center's dumpster? And then there is the cumulative factor.
Take a table with 50,000 names and addresses, mailed every month. The first month, 500 (1%) of the list have become undeliverable for one reason or another. Into the trash they go. Month two, not only are you sending the first, undeliverable pieces a second time, but are also sending a new 500 pieces which have become undeliverable since the last mailing (1% per month change rates are conservative - usually they are higher). The following month, you send out the first 500 pieces a third time, the second 500 pieces a second time and a new 500 undeliverable pieces for the first time; for a total of 3,000 pieces, costing you around $1200.00, with no possibility of ROI (Return On Investment). In six months, assuming that this continues uncorrected, your expense will increase geometrically to $10, 500.
Whether you mine your data, depend on current reports to keep you informed about your company's progress and performance or use direct mail to enhance your CRM and increase sales, or just run a lot of queries and prefer them to be accurate; paying attention to actual and potential data problems and their ongoing resolution now will save you hundreds of hours and thousands of dollars in the future.
Northwest Database Services has been making databases easy for business for nearly fifteen years. The company specializes in cleaning and processing clients’ data, creating custom database applications, purchasing and combining data, and preparing direct response mailing lists to USPS specifications. Based in the Pacific Northwest, the company offers a complete range of options from setting up or optimizing a database for the client to host and run to a complete maintenance package including data entry, cleaning, optimization, and report generation.
Northwest Database Services founder Geoffrey Hollander is co-author along with Debbie Mahterian of How To Make Your Mailing List Pay Off. He has written dozens of industry articles and software reviews for magazines like Windows Magazine, Intelligent Enterprise, Home Office Computing, and is currently on the InfoWorld Review Board. Mr. Hollander has helped found and been an active member of local Postal Customer Councils.
Click Here - to visit the data processing area of Geoff's web site (www.nwdatabase.com).
Email Geoff at gch@nwdatabase.com
[PRINTER FRIENDLY VERSION]
|
|
| |