How good is your data?
If you ask most data managers you’ll probably get a response of “Ok” or “about 90% spot on” or some other vague response.
To gauge quality, we need a reference point and this is where most data managers and custodians struggle – they have nothing to cross-check the quality of their data against.
We need to understand the quality of our data because without this level of assurance it is dangerous to make business decisions and attempt everyday operations with any kind of certainty.
In our personal lives we are surrounded by quality reference points. For example, when the tread on the tires in our cars get below a certain legal limit they become dangerous and have to be changed. We have a quality benchmark set by government that we can easily validate against. As the tires get closer to the limit their quality degrades to the point we need to take action. The same applies for brakes, battery acid and many other assets across the car.
Yet when it comes to data we ignore these “quality checks”. I once posed this question to a data manager who responsible for over £5,000,000 worth of utilities assets: “How come you’re prepared to spend several hundred pounds a year validating the quality of your cars assets but £0 on your data assets for a multi-million pound business”.
The big problem of course is expertise, most people lack rudimentary data quality skills but anyone can get started.
The first step is rules. Just like the rules we have for legal limits on the assets of our car there are hundreds of rules that bind our data.
“A customer has to have a valid account identifier”
“All products need to have a SKU, Product Description and Accurate Pricing information”
Simply by documenting these “rules of legality” for your data gives you a starting point to create a data quality benchmark. By doing this you’ll start to understand what “Ok data quality” really means and more importantly, where the gaps in your data exist.
If you think that your data “is ok” but you don’t have these kind of rules then you’re perpetuating the myth, one that is fraught with danger and never-ending peril. Create a set of rules for quality data and you will be surprised at how much value this simple exercise can bring to your business.