Conversations on Analytics

In the much mourned absence of the best communicator of value that I think we've ever seen in the field of data and analytics, and in an attempt to honour his legacy, I hope that this post will help all of us who try to bridge the gap between data and business in order to unlock value, drive growth and solve problems.

Firstly I'd like to offer some advice on how not to do it. Believe it or not the exchange below actually happened, in a board room not too far away from where I'm sitting now, between the CEO of a huge blue-chip and the then Head of Analytics.

Whilst it is of course an extreme example, no matter what we think we are saying, it it what is heard that is important. If when extolling the virtues of data to business executives, we rely on how clever the models, the maths, the stats or the AI is, we risk conveying the same dismissive and destructive message. No-one other than the guy above would actually call executives stupid, not if they're expecting buy-in, or continued employment, but think how you feel if someone uses deep jargon or assumes your prior knowledge of a very specialist subject!

It isn't easy, however, to get across how important analytics are, without talking about the nuts and bolts of what goes on under the covers. One good way to plan such a conversation is to consider the starting point of the listener, and to be able to explain a couple of things that, whilst seemingly simple perhaps, will pave the way to a quicker mutual understanding.

My belief is that anyone interacting with business or government execs would do well to consider how, without resorting to tech talk, they would answer the following two questions. Whilst they may or may not always come up in real life, the approach to answering them certainly will come in useful.

The first question :-

I've lost count of the number of times I've had this in Q+A sessions. In fact I pre-empt it now by bringing it up-front in my talks and workshops. When we talk data, analytics and all the other great stuff that excites us, often newcomers to the field think in terms of information, and they draw analogies with how they access all the other information they need to go about their daily work. Hence the Google question.

The second scenario, equally prevalent and usually the most common exposure newcomers already have to the field of predictive analytics :-

My experience is that if you ask any group what it actually means when the weather forecast predicts an 80% chance of rain, you get as many answers as there are individuals in the group. I certainly don't know for sure what it exactly means, and I've been doing this for 30 years or more! Then even if you get a common understanding of what it does mean, you always get divergent opinions about what to do, what action to take, in the light of that prediction.

And yet we ask for investment on a predictive customer churn model because the new one is going to provide 94.7% accuracy versus the old one's 84.6%? Our audience's "already listening" may or may not be working in the way that will help yield the outcome we want.

Sadly I can't offer a silver bullet, nor am I going to try and offer a cookie-cutter answer to either of the examples. My point is rather to encourage you to view the discussion from the perspective of what the audience already knows about and can comfortably embrace. I believe that knowing how, again without using language that shrouds and obfuscates, you would go about explaining the two scenarios above, and then applying that thinking to your executive engagements, you will get buy-in far more quickly.