Thoughts on Data Science, IT Operations Analytics, programming and other random topics


Five whys of predictive analytics

13 Apr 2015

(first published on IBM Service Management 360 site)

In our domain of IT Operations Analytics (ITOA), we commonly have to determine the root cause of problems, and there are many approaches to this determination. I often use the Five Whys approach as part of my explorations. This investigative technique is useful in many situations beyond root cause analysis. I often apply it when I’m trying to make some basic sense of a confusing situation. When grappling with a simple assertion that you should apply predictive analytics to your area, running through the five whys will probably help you understand the situation a little better. For example:

I could imagine such an interaction between an analytics vendor and a potential customer. This is close to the traditional root-cause approach, but with a little twist we could go down a different path. Extending the “Why?” to “Why do I care?” helps me think about potential value, which is of more concern when considering a purchase. Sometimes, more pointedly, instead of “Why do I care?” I ask “So what?” as shown here:

Whether or not they remain clients probably depends on a number of other considerations too. Another round of “why” questions might shed some light on those aspects of your situation that might also require attention.

Let’s try one more pass and as you read through, think about the different questions and paths you could take instead at each stage.

These exercises may seem trivial, but this relentless questioning can be very useful to get at assumptions and premises underlying a discussion. The real point here is to encourage you to work through the hype by applying some critical thought as you assess the potential value and pitfalls of adopting technologies. Also, don’t limit yourself to five questions (obviously!). Sometimes the clarity comes with three questions, others require more. At times it may be interesting to pause at each step and consider the surrounding implications, each of which may drive further investigative steps:

We don’t have full knowledge of all the changes in our environment.

Why?

Because of process and tool deficiencies.

Where are the gaps?

And so forth; you get the idea!

The technologies we work with have huge potential, but both they and the environments in which they are applied can be a bit overwhelming. Sometimes it really does help to pause and think about things from a base level point of view. Try it―you may stumble upon all kinds of interesting insights about your environment and the suitability of new products and tools.

I look forward to hearing about your ITOA challenges and related thoughts. You can contact me here or on Twitter @rmckeown.

comments powered by Disqus