skip to Main Content

A New Take on “Data Quality?”

Post by Rob Collie

image

Bad Data DOES Lead to Bad Results.  But Good Data Can STILL Lead to Bad Results.

Garbage in, Garbage out.  We’re all familiar with this.  If you’re being given junky source data, it’s going to be hard to perform ANY meaningful analysis or reporting on said data until the quality of the inputs is addressed.

The term “Data Quality” has come to mean precisely that – the quality of your inputs.

But at my recent PASS BA presentation on the Bottom Line, I was talking at length about how we often generate poor outputs – our reports and dashboards often leave much to be desired, because we ourselves, the producers of the work, need to be better.

It’s one of my most-emphasized themes:  we’ve been given this amazing new toolset (Power Pivot and the rest of the Power BI stack).  We shouldn’t just use it to produce the same stuff we produced for decades (even though we can do so much faster and more efficiently than before).

We should strive for more meaningful metrics for instance – metrics that remove noise and produce a clearer picture than the age-old default of “raw dollars.” 

A product may be generating more dollars than last year for instance, but that could be misleading.  Is it generating more profit (it may also be more expensive for us to acquire this year)?  Is it generating more profit per store (we may have increased the number of stores that sell it)?  Per day (maybe it was introduced in May of last year, but this year it’s been available since Jan 1)? 

Read the Rest