Outside of security, data management and analytics likely represents the biggest IT spend of most organizations. The thirst to gather more and more information in a world that is generating new (if not necessarily good or valid) data at an exponential rate is too great to ignore.
But the frantic way in which businesses want to consume, track and trend data is outpacing the ability for most organizations to do it right. By right, I mean basic concepts of figuring out what these new data elements mean to the business. My experience is that they drive inferential rather relational understanding of what is happening.
When data is applied on an inferential basis without proper validation it generally means it is applied according to assumptions: or even worse staged and used to back up previously implied business challenges. These are likely wrong.
In his book The Signal and the Noise Nate Silver goes into detail about why many predictions fail. He maintains it is because our context is at fault. And Harvard Business Review article on why analytics projects fail, blames IT to some extent because our risk adverse mentality means that we need to define a business case and outcome in order to set out requirements. It argues that data analytics are really successful based on answering questions you haven’t even thought of asking – rather than the ones you already have the context to answer.
My point is data is becoming the problem and not the answer. People cite information as justification to spend money, change business strategy and justify rash actions without understanding the data and context by which it should be evaluated. IT is unfortunately complicit in this huge organizational failure.
We layer bad systems and processes on top of ill-defined data models and non-validated data structures, and easily forget about the inherent falsities in our historic data relationships which now shape as gospel the new platform from which the organization assesses and initiates critical business strategy.
Data is the most mishandled and misunderstood IT resource, and it is compounding the issue of IT credibility because we tell our business they can trust it and they cannot. We need to be more assertive about what our data is and what it isn’t.
We need to determine the gaps in our approach and change the way we manage, structure and process data to make it more meaningful and more pliable. We also need to be the arbitrator that tells the organization where context is sound and more importantly where it isn’t.
Let’s be honest with ourselves. Our data is not nearly as clean and meaningful as we all testify to, and it is time to change that by both owning up to it and ensuring everyone knows the limitations and risk that this presents to our understanding of our business.
The Naked CIO is an anonymous technology executive.
More Naked CIO