There is a new report recently published by the IBM Institute, “Analytics:A blueprint for value”, well worth reading. The image above – shows 66% leaders “trust” their data.
First, let me clarify the context of that chart, which is that leaders believe in using data in daily decision making. They trust the “quality” of the data. They trust the capabilities of their analytics and technology teams.
But selecting data based on these criteria raises an important problem. You can be confident in adopting a data-driven management approach, but a good manager must always be skeptical of the data.
Be skeptical of the data, don't be skeptical of its use
Data is supposed to settle arguments, helping us get out of situations where people just argue. (according to HiPPO).
However, ironically, it can have the opposite effect:
- Everyone has a belief.
- They found data to support that belief.
- They use that data to bolster their belief in irrefutable truth.
The two central issues of this dilemma are:
- There is an almost unlimited amount of data to choose from.
- There is an almost unlimited number of stories one can build around that data.
Once we realized that, we realized that data doesn't necessarily solve the above problems. Data can tell us “what,” but it cannot inherently tell us “why.” To find out “why,” we need a very specific type of data: data from validated experiments. Because the “why” is often more important than the “what” in determining future action, testing is a more powerful management tool than data analysis.
Of course, not all the data we process comes from validated experiments. For that matter, it's very difficult to have a perfectly controlled experiment in the real world. We have to deal with imperfect data to help us answer “why.” But imperfect data can be much better than no data at all.
You can be confident in the use of data while not fully trusting the data. Not that you doubt its accuracy. But you have doubts about the context, relevance and completeness of the data in the conclusions drawn from it.
I believe that the best way to work with data in a strategic way is to adopt a “Bayesian” worldview, advocated by two strategic Marketing thinkers, Gord Hotchkiss and Greg Satell.
In the Bayesian strategy, it is not the data that is used with certainty but the data that is used probabilistically to continuously update the market situation. It's not a certainty, it's a relative thing - with an open mind to data, you can proactively update data changes.
“Fact-based decision making” is an increasingly popular term. I still caution that “truth” is a very nebulous thing, especially when you go from “what” to “why.” All too easily opinions can be disguised with data to look like facts.
A decision based on actual data is better if the data has been confirmed.
Original post: Why data-driven marketers shouldn't trust the data fully – Chief Marketing Technologist (chiefmartec.com)
Translator: Hoang Bach
