I'm thinking of a visual I can use to begin to persuade folks at my company that customer data quality is a pervasive and costly problem.  the hard number metrics are really hard to come by, because the actual impact to each person is tiny - but the cumulative impact is large. In other words, numbers just aren't compelling.  Here's my idea:

Take a count of the contacts in our enterprise.  Figure out how many are customers and spend money with  us, and how many of the prospects are at least responding to marketing and/or are in the sales funnel.  The customers and quality prospects are pennies.  Everyone else is a grain of rice (duplicate contacts, ancient, non-contactable out of date records, etc.)  Mix it all up in a large glass container, and ask folks how easy it is to find the value - the pennies - in our customer database.  I'm expecting the volumes to be such that the coins are easily overwhelmed by the rice.

Sales can't sell without manually filtering out the rice.  Marketing has to manually filter out the rice.  Customer service can't be efficient b/c they're seeing nothing but rice.  Even IT is impacted by the rice because we're supporting systems and incuring overhead that equates to lugging this huge glass container around, mostly full of... Rice!

Data quality programs would remove the rice.  Getting people to care about the data will persuade them to STOP adding more rice every day.

Progress could be shown as the program progresses, by scooping the rice out into a side-container labeled "waste".  Preventative measures could be a barrier on the top of the container, preventing more rice from coming in.

I want to put it in the lobby.  :)

E75's picture

I like your idea. I have a suggestion for you, however.

Assumption: looking at a single customer data item, it is not immediately obvious whether the piece is a good or not.

Separating pennies from rice is a simple exercise of using a sift. If my assumption is correct, your analogy breaks here.

I think you should consider using two classes of items that differ perhaps only visually. Red and blue marbles?

Simon Flowers's picture

I like your idea of finding the valuable items.  Maybe you should also do the old style panning for gold - washing away the dirt, sludge, and "fool's gold" to find the valuable nuggets.

Two other good visuals in data I have seen.  One company wanted to establish a single central repository of consumer goods.  They took one of the company's products and got a photo with the name and local material number equivalent in each European country.  Put it together on a map showing the same product described in so many ways showed how ridiculous the situation was.

The other is for data conversion and the need to clean data from legacy systems before conversion.  Each division had a data owner.  The leader of data conversion had a fondue set in its original box.  He explained it was a wedding gift to him 18 years ago, had stayed with him through 4 houses moves, and never been used and just taken up space.  Finally he was making the effort to get rid of it .... by donating it as a "prize" for the whichever data owner deleted the highest % of data before the conversion !  A good visual and a good laugh.