Thank you to the author of this article...
http://www.hcamag.com/business-news/big-data-is-still-only-a-little-helpful-217026.aspx
Barely a day goes by without someone in the tech world writing about the benefits of Predictive Analytics and Big Data. Don't get me wrong, I am all for data and information which supports decision-making, and am associated with several companies that play in this field. However I am also starting to ponder whether predictive is distracting us from proactive. I will explain.
The case for 'predictive data'
There is so much information being made available (IBM suggests that 90% of the world's information was generated in the last 2 years!!), it is not only logical but downright necessary to try and make the best use of it. And as we learn more, the number of people who are able to make use of it, combined with the computing power that is becoming available, it is becoming a realistic option for many companies.
The lure of being able to make better decisions is a strong one. Imagine if we could look after customers better, or provide employees with a better experience. Imagine if we could pick up on a better way of doing something, or shave off costs to boost profitability. Imagine if we could predict the future? "Yes please, give me more. Show me a way, because it will really improve my business!".
Correlation or Causation?
For me, herein lies the confusion. If something is predictable, it is because the underlying cause is known. Changes in the actual cause of something can therefore produce relatively predictable outcomes. However, how much of what we think are causes are actually correlations? Let's use a simple example. Say we are Orange Growers, and we want to be able to improve our Orange Growing operation. Over the last 5 years, we have been collecting information about every aspect of the business, from the average environmental conditions for each hour of the day, to the picking rates of the pickers (cross-referenced to their demographic data), the age of the trees, the composition of the soil below each tree, transport routes and customer satisfaction ratings (amongst a bunch of other things). Awesome. The analytics might tell us any number of things, from the best time of day to pick, to the best type of people to hire, to the best brand of trucks to transport the oranges. Individually, all of these things might make a difference. However are they the causes of a better Orange Growing Operation, or are they simply correlated based on the evidence presented. What is the impact of the data that hasn't been considered, that might influence the result (like the CEO, or the composition of the board, or the demand for oranges) ?.
The point being that many companies, whilst they have access to very good data, don't have the completeness of data that could warrant an absolute pinpointing of 'Cause'. And the issue is, it is really hard to tell. Cause and Effect seem easy to link, but actually they can be completely disassociated. And by investing in correlations, it potentially increases business risk.
The case for 'proactive action'.
Let's now, for argument's sake, go the other way. Let's say that rather than collecting all the data, the Leadership Team executed its duties in that it set the Vision of the Business, agreed on the underlying Values and Policies that were important, and made a pact to being open, transparent, and accountable. Rather than relying on data, they then spent the time putting in place the things that would most matter to the business in line with its strategy. Hiring the right people, constantly communicating with the workforce about what is happening, and proactively checking in at appropriate intervals. It proactively up-skilled people (because it was a policy to hire people who wanted to be up-skilled), and put in place a number of measures that helped to understand effectiveness. It went through a process of risk assessment, and put the contingencies in place.
Which is the smarter bet?
Who really knows, right? There are people who will argue strongly for either side, and as long as the company achieved its objectives, does it really matter? However the point I wanted to make is, predictability is really difficult, and fraught with danger. Wayne Gretzky was famously quoted that he always played according to where the puck was going to be. However the prediction was not made from some level of data analysis, it was made from experience and playing the current circumstances, and that is a skill that will always separate the best Leaders. Often, instant decisions are called for that won't wait for the analysis.
By the way, if you are someone who is heavily invested already in the big-data phenomenon, please understand that I mean you no harm. But with so many claims being made, I hope that it doesn't become the next 'magic pill' that leads companies further into a path of pain and suffering. Let's not forget about the amazing Big Data machine called the human being. Until we are ruled by Cyborgs, developing these skills will always be necessary.
Which bet will you make?