The Data Does Not Speak for Itself

Nov 16, 2017

The Buzz About Data

There’s a lot of buzz out there about data: big data, “sexy” data science jobs, billionaires with dystopian dreams of self-aware AI… The lis

t goes on.

And the buzz is understandable, because there are lots of companies doing cool new work with data and the capabilities of systems are improving dramatically. If you read the business press, you get the sense that if we just have enough data and good enough software, the answers to all our questions will pop out of an algorithm. It reminds me of that old saying: “the data speaks for itself”.

But the reality, as any statistician could tell you, is that the data doesn’t speak for itself. It never has and it never will. There’s always a process of capturing it, interpreting it, and using it. No matter how sophisticated the tools we use, ultimately, it’s people who drive that process. Which is why we all need to get better at talking about our data and how we use it.

Dueling Data

The good news is that it is easier to capture and work with data than ever before. If you don’t already, you will soon have tools at your fingertips to marshal all kinds of information in support of your work. But people with different viewpoints from yours will have access to those same tools. How many times have you heard some invoke “Well, the data says…” as a way to claim authority and protect their point of view from questioning? It become less about finding the truth and more about adding weight to our personal agendas. If we’re not careful, we’ll end up simply trading opposing views, supported by increasingly large sets of competing data points. That won’t be much different than what we did before “big data” – just with more charts and graphs.

If we’re going to realize the potential advantages of a data-rich environment, leaders need to get much better at having transparent conversations about their use of data and the thought process that took them from that data to their point of view. To be clear, I’m not talking about peppering our everyday speech with phrases like “confidence interval”, “standard deviation”, or other technical jargon. I am talking about being able to walk each other through our thinking and the information it’s based on, whether that’s complex technical data or our own anecdotal experience.

Why Is That So Hard?

Trouble is, it’s easier said than done. Luckily, this problem isn’t actually new. Way back in the 70’s, when all the computing power most of us had was a calculator, Professor Chris Argyris was studying the behavior of leaders in business. He articulated the concept of the Ladder of Inference to help understand the often-unconscious mental processes that drive our action. A slightly condensed version looks like this:



As a social scientist, Argyris was talking about data in the broad sense of all observable information. The path from that observable information to our conclusions and actions always goes through a process of filtering and interpretation -even if that happens in a millisecond. In his quest for valid and actionable knowledge, Argyris’ primary concern was how invalid filters, biases, and assumptions can lead us astray. That can be due to errors in thinking we make unconsciously on the way up the ladder, or it can take the form of defensive reasoning where we back into data that justifies our established position. These same concerns certainly hold true with the kind of data-based decision making many organizations strive for today.

Walking Back Down the Ladder

Explicitly walking through the steps on the ladder of inference can be a powerful method for addressing both everyday and technical data. When we find ourselves in disagreement with colleagues, rather than simply trading competing conclusions or recommendations, we can open up our thinking and invite them to do the same.

Research suggests that if we unpack our data and our thought processes, we have a much better chance of influencing people. And if we can encourage others to do the same, we stand a much better chance of understanding where they are coming from and perhaps being influenced ourselves.

My team has found these kinds techniques very powerful in helping leaders collaborate and make business decisions.

How are you helping leaders in your organization to better talk about their data?