Up…

Dec 20, 2010 by Rich Murnane in Data Quality

Like kids on Christmas morning, almost immediately after receiving and installing our favorite data quality tool at our shop we went wild and started cleansing data like crazy. After the indulgence of all this data cleansing (a data cleansing hangover?) we decided we’d take a breather and spend some cycles building metrics about our data. Yes, we should have done this first, but we needed to get some data cleaned up and we needed to do it fast. Once our framework for capturing data about our clients’ “data quality” was designed and implemented, we decided we’d represent these metrics as numeric values called “data quality scores.”

After some deep thought and significant effort, we were finally ready to communicate these “data quality scores” for clients to our senior management. We put together some reports and graphics, met with our senior managers and sure enough we were completely surprised when the fruits of our labor were met with disdain.

What did we do wrong? Something so simple it’s almost embarrassing. We didn’t know our audience. Our audience was comprised of senior executives and the senior executives at our shop like to hear good news and don’t like to hear bad news. The delivery of our message and associated “data quality scores” was completely backwards because we focused only on the negative.

We walked into the meeting with reports saying things like, “9% of client X’s data has issues” while we should have been saying “91% of client X’s data is error-free.” Our graphs should have indicated that our data quality scores went up when things were better, but they didn’t, they went down. Although mathematically our graphs were correct, the executives at our company like to think positive and graphically speaking they think “UP is good” – so our message (and our graphs) were all going the wrong direction.

Now, I’m not saying that we couldn’t explain our way out of this, nor am I saying that all “metrics” or “measurements” should always go up, but as data quality professionals we tend to focus on the negative way too often. Had we changed our communication strategy a little bit our hard work and cool graphs would have been much better received. In your data quality efforts, I’d strongly recommend focusing on the positive and when in doubt – remember – “UP is good.”

Happy holidays to all…Rich

What other data challenges do consultants face with clients?
Read more posts on the topic.

Tags: consultants, Data Quality Score

9 Responses to “Up…”

Jim Harris

Dec 20, 2010

He blogs, he scores . . . with a post about the power of positive thinking regarding data quality scores, you have once again proven that you are a very UP person, Rich

Thanks and Happy Holidays,

Jim

Reply to this comment
Henrik Liliendahl Sørensen

Dec 20, 2010

I’m with you Rich. We must strive to make our databases happy (and our sponsors too).

Merry Christmas and a Happy New Year.

Reply to this comment
Rich Murnane

Dec 21, 2010

Hey Jim, thanks for the comment. We always need to look UP to find out where we are going.

Henrik, I think perhaps you are the “database happy” focused person out there!

Best to both of you this Christmas!

Reply to this comment
Phil Simon

Dec 21, 2010

This isn’t the point of the post, but this got my goat:

Our audience was comprised of senior executives and the senior executives at our shop like to hear good news and don’t like to hear bad news.

Argh! To quote Nancy Kerrigan, “Why? Why? Why?”

Reply to this comment
Rich Murnane

Dec 21, 2010

Hey Phil,

The execs at our shop understand that bad news needs to be heard and perhaps I didn’t state it clear enough, but the moral of the story is that:

91% of Client X’s data is good

lands “much softer” than

9% of Client X’s data is bad

Just tweaking the delivery really does make feel a bit better, even though it’s mathematically equivalent. “Leading” the conversation with the negative seems to discount all the efforts put in place already, guess it’s plain old psychology.

Best…Rich

Reply to this comment
Sean Sinclair

Dec 22, 2010

Rich,
It’a a really good point and unfortunately I’m guilty as accused..! Have done exact same thing recently (although haven’t had any grief). Glad I read your blog – am going to change it for next month….
Thanks,
Sean

Reply to this comment
Victor Hudy

Dec 30, 2010

Rich,

Thank you for your message – executives always want to see positive results. The positive message also demonstrates that our data quality and governance efforts are working and providing positive ROI. Do executuves like to provide financial support for “negative” efforts – I don’t think so. The positive results help support our funding to continue our efforts. Keep looking up!

Reply to this comment
Mark Horseman

Oct 17, 2011

I’ve been selling “lower” as a “Golf Score” Using terms like, “Eagle, Birdie, Par, Bogie, and Other”, traction is not ubiquitous on this idea.

Reply to this comment
Mark Horseman

Oct 17, 2011

What business concept do you use as a denominator such that a percentage can be calculated? What is the scope of data that could be clean? Is it based on an individual unit test (or Data Quality “Screen”)?

Reply to this comment

Up…

9 Responses to “Up…”

Leave a Reply