Those Data Quality Dimensions

The quality of our data was in doubt,

We hired consultants to sort it out.

A major issue that struck them as key?

The absence of a data strategy.

They taught us six quality dimensions,

They were sure that they’d ease data tensions.

The six are: valid, consistent, unique,

And also accurate, timely, complete.

We asked “Where is it that we should begin?”

They said “‘Valid’, don’t let bad data in.

All inputs and updates – please validate!

To keep your data in a shipshape state”.

“Another quality aspect you need

If your data strategy is to succeed:

Is that your data should be consistent

Even records that reside somewhere distant”.

“All records stored must be truly unique”

We agree, but we don't know the technique.

We will review identifiers and keys,

But we don’t yet have the correct expertise.

“Complete accuracy must be pursued”.

Impossible. (Then we said something rude!).

In the real world data’s always in flux.

“There are good tools but their cost is big bucks”.

Timeliness was much trickier to explain,

And a few unanswered questions remain.

Something like ‘data at a point in time’?

But I’m sure it will be something sublime.

Another measure we don’t think we’ll meet

Is for our data to all be complete.

What’s the best way to find all of the gaps?

“Try the latest data profiling apps”.

They told us “Don’t be put off by the cost,

Consider opportunities you’ve lost

By having data that’s full of defects,

Because you would not do your data checks”.

“The problems that impact business the worst,

Are those you should attempt to fix first.

To our proposals, we hope you’ll agree,

But please remember to send our fat fee”.

©️ Ray Cohen, May 2023