Buy vs. Build (vs. do nothing)

Do Nothing

About 35% of enterprises ... ignore the [data quality] problem or hope it will go away when new/upgraded systems are installed.

Meta Group, March 2002

Judging by press reports and analyst commentary, it's rare for a company to identify the actual cost of low quality data and tackle it directly as a distinct project. Instead, IT projects tend to focus on applications with higher visibility, e.g. CRM or business intelligence. Relegated to a secondary role, data quality may not even make it into the budget or on to the schedule.

How does a company determine whether data quality should be getting more attention? Try this. For each major functional area (e.g. sales, marketing, customer service), find one or two end users who have previously made useful suggestions for improving systems or procedures to streamline their job and/or otherwise save the company money. Ask them if data quality is a problem. You're likely to get an earful.


Even when companies put data quality on the agenda, many choose to build a custom solution rather than purchase a commercial product. In fact, we've been in this very situation. For developers with the appropriate background, it's not all that difficult to create custom code to fix a wide variety of data problems. (Development takes time, and the code may require ongoing effort to address new problems that arise -- but these are well-known "buy vs. build" issues that aren't specific to data quality.)

However, there is a trap for the unwary: finding the problems is no easy task.

You may spend much more time checking for errors than cleaning up errors. Most of these errors do not jump out at you.

Larry Greenfield
link_outAn (informal) taxonomy of data warehouse data errors

For example, a few lines of SQL suffice to list all unique values in a field -- but interpreting this information by hand is quite time consuming compared to DQ Now's approach. In addition, there are several types of problems that are difficult to fix with an ad hoc solution.


The track record of in-house software projects remains poor: over budget, delivered late, and still fail to meet user needs. Accordingly, most companies lean towards a BUY decision on software. If data quality made it into the budget, the company will probably do some market research and invite several data quality vendors in to demonstrate their solution.

Until now, we think many potential customers will have been very disappointed by what they saw. DQ Now changes that. Want to see for yourself? Send 1000 records to each vendor. Compare both static reports and interative tools, then decide for yourself which is most productive.

(The above quotes have been gathered from public sources. No endorsement or even knowledge of DQ Now is implied. External links are preceded by link_out.)

Next step: see how our services complement our product.