20.4. Getting Information Quality

Information quality requires measuring the quality of the information stored in company databases, identifying quality issues, and having processes in place to fix them.

20.4.1. Building In Quality

The best way to achieve high-quality customer information is to capture it correctly in the first place. However, those responsible for data capture often are managed by siloed organizations with no data quality performance measures. This can be changed over time, but it's seldom an effective place to start. Start showing the value of the data, illuminate the cost of hygiene and the poorest quality sources, and go after these.

Usually, companies only become concerned with quality of data when it flows into a reference file or a consolidated data warehouse.

20.4.2. Managing Quality for the Life of the Data

We know that there are many factors that impact the accuracy and completeness of our customer information. We use some tools proactively to prevent and/or fix reoccurring problems. We can also use many of the same tools to repair unexpected problems that occur.

Proactive (Preventive) Maintenance

There are several standard data hygiene activities that all companies should use as a matter of course, such as processing return mail. It costs just a little more to have the undelivered post returned to you, and you can save much more than that by not mailing to the same address again and again.

Scheduled database audits that test for quality levels by segment and data tier are also very helpful. Some scheduled data quality efforts are costly and time consuming, so we want to do them only when they are needed. Sometimes, the results of one of these audits will drive the launch of an ad hoc data quality action plan to address the specific issues that were identified. Of course, we can also use any of the ad hoc maintenance plans we've created as part of scheduled maintenance if they turn out to be useful.

Some, but not all, of these data hygiene activities can be done by computer using special hardware and software. Examples are the several software products that are available for address standardization and matching. Also, many data services firms have proprietary processes that accomplish the same thing and that can insert additional information into each of the company's customer records. Of course, some of the work may be done directly on the database itself. But don't give everyone access to directly update your customer reference file, only qualified and trained data administrators should be given this responsibility.

Ad Hoc Maintenance

Ad hoc maintenance is used to meet the criteria of a specific project effort or when a data quality problem is discovered unexpectedly. If the problem appears to recur, these ad hoc cleaning projects should be made part of regularly scheduled maintenance.