The Ballad of Dirty Data

Upon my arrival into this place

Self-assured stride, grin on my face

Confident in my analytical skill

Certain my insights would inspire and thrill

Reality quickly came into view

And a realization of nothing so new

 Like many before this place did hold

Habits surviving from days of old

For though these leaders yearned to know

What drove their business to a fro

They failed to see just what they needed

Or understand what their data heeded

In some past time they had been told

By vendor, consultant, or seeker of gold

That should they spend some cash today

The insights would bloom day after day

So spend they did and signed away

Their precious funds to seize the day

But quickly realized soon thereafter

These insights did inspire some laughter

Not only were perceptions grossly flawed

But dirty data turned insight to fraud

And the deeper they dug their anger rose

For governance wasn’t the path that they chose

There had been entries of a manual type

That appeared to be entered after smoking a pipe

And other elements that were not quite rare

Were missing entirely; nothing was there!

So what did the insights mean for them now?

With all of these issues hanging around

How could they rely on such inexact tales?

When confidence in data so obviously fails

What too do we say of similar labels?

Not mapped or linked through relative tables

Taxonomy, dictionary, ontology too

None of these things were given their due

Gut and emotion had driven their why

Of this I informed them; I was not shy

For the truth was harsh and it did sting

But it was a first step to fixing this thing

I explained to them rigor was now so required

Some process, and policy and resources hired

To work through the mess and profile the data

To clean it and test it, first now, and then later

You see governance is not a box to check off

It’s something you live by not something you scoff

You use it to build that foundation of yours

In turn it saves you from unneeded chores

So next time you rush to that insight you want

Remember to pause and search and hunt

For any rogue values that’ll send you askew

If you govern your data, insight is true