Bad Data In, Bad Data Out

It's funny -- no, scratch that; it's pathetic -- how easily media types like me can get dragged into the latest big news trend. Data integration has been a hot topic in the business intelligence realm of late, mostly because of developments on the vendor side. But maybe the topic's a little hotter than it should be. Other, more fundamental business intelligence questions remain more important to the real BI practitioners out there.

I'm talking about the BI problem that won't go away: Data quality. Inaccurate, badly formatted, inappropriate, unclean, redundant or irrelevant data continues to thwart the business intelligence process in organizations of all sizes. This seemingly intractable problem is reflected in my conversations and correspondence with BI pros of every stripe.

In our current Business Intelligence Pipeline poll, we're asking whether you expect data integration or data quality to be the bigger challenge this year. Despite all the chirping coming from vendors about data integration, and the media's repetition of said chirping, Pipeline readers have so far cited data quality as their larger concern by a margin of two to one.

We're leaving the poll up on the site for another week for two reasons: 1) It's getting a big response, and 2) Its results have fluctuated more wildly over the last week than most polls do. So we're taking the time to hammer out a real consensus. Take a second to cast your vote.

Last year I asked readers to share tips, advice and wisdom with me about how they tackle the data quality issue at their organization. We got solid feedback from readers in the not-for-profit, private and public sectors. You can read a little of it here.

Now, I'm asking for more of the same. BI is becoming more central to business operations, intelligence demands from decision-makers are increasing, and turnaround times on analysis and reporting are tightening. You've gotten better at data warehousing, data integration, on-the-fly analysis and reporting. Have you gotten better at ensuring data quality? If so, how? What have you learned?

Send me an e-mail. Let me know if you'd prefer that your identity be kept anonymous, and I'll do that. Otherwise, indicate your name, position and the organization you work for.

I'm going to return to data quality issues in coming columns. Give me something to share. Is data quality an IT issue or an organizational problem? What are the fixes? Is it even possible for IT to fix the problem on its own? By pooling what we've learned, maybe everybody can benefit.