You think you make smart decisions – Think again!
Being “data-driven” is hot these days. The main challenge is not technical nor is it analytical, but cognitive. The way our brains are wired leads us astray. Failing to address this obstacle, and we will not succeed in our quest for fact-based decisions.
For most organizations, becoming data driven is an important initiative and priority, and for good reasons. The amount of information that is available to organizations these days is growing exponentially. Data from within the company, customers, value-chain, third-party databases etc. Add to this the explosive growth in data from connected devices – everything is measuring and collecting data. With these huge amounts of data, we need to have tools to turn this data into real-time insights. When we can do this successfully we are able to make the right decisions and improve profit margin, reduce costs even gain competitive advantages.
Technology plays an important role in turning huge amounts of data into information with amazing speed and generating insights that has never been possible before. With self-service BI tools, almost any business user can visualize data into meaningful graphs and dashboards in no time. One would think that with all these data and technical possibilities making the right decisions would be a piece of cake – well not quite. The amount of data and the right tools does not guarantee in anyway high quality decision making – We (humans) give ourselves too much credit.
We assume that we make rational and smart decisions, but we do not. There is an increasing gap between organizations capacity to produce analysis and its ability to apply them effectively to business issues[i]
The greatest enemy of knowledge is not ignorance; it is the illusion of knowledge! Stephen Hawking
Our cognition are full of flaws
Look at the two lines. Which one is longest?
This is a well-known optical illusion, and as you might already have guessed, the two lines are of equal length. The important thing to acknowledge is that, even though you know that is the case, they still appear to be of different length. This indicates an important conclusion: Our automatic and unconscious cognition are prone to errors.
Look at the two tables, which table is longest and which is widest?
They have exactly the same dimension. If you do not take our words for it, use a ruler. These examples show how our automatic cognition makes wrong judgments, and leads us astray. Now a more business relevant example. Imagine you own two projects. Both projects cost 100 points, and both projects are approved. Project A has a payback time of 4 months, and project B has a payback time of 24 months
You can improve one of the projects payback time. Either improving Project A from 4 to 3 months, or project B from 24 to 12 months. What do you choose?
Most people will automatically select the largest absolute improvement e.g. project B. Which is wrong! Even without calculating with an internal rate, the increased value of Project A is twice as big; using any form of internal rate will increase the gap in favor of option A.
We automatically assume a linear relationship, because that is how our brain works. We therefore fail to see that when payback time approaches zero, the value approaches infinity.
We have a bunch of these biases, and it screws up our thinking constantly. Moreover, most of our mental models are constructed without any form of deep reflecting and thereby a very likely to lead us astray.
So how does this affect decision-making as it is practiced in most organizations? In our experience, most companies practice a HIPPO (Highest paid person’s opinion) decision-culture, which is obviously not the preferred state, so what can be done about this?
…managers generally base decisions on gut instinct. What’s surprising is not just how bad those decisions typically are, but how good managers feel about them[ii]
How do we overcome our cognitive inadequacies?
We use a concept called “Equation of the Firm”, which to our knowledge is a term that originates from the data insight team in Google[iii]. The purpose is to describe what has priority to a firm in an equation. If profit is the goal, the aim is to understand the variables influencing profit including these relationships to each other. The equation is scalable (think DuPont Model), so it can both cover a company wide perspective, and an atomic perspective. In the right perspective, it is therefore a relevant concept for all corporate decision-makers.
By reflecting about what variables that influences your goal, you improve your mental models, which governs your actions and decisions. For “customer retention”, the variables could be quality of products and services, loyalty, competing offerings etc. We use this exercise as an intervention that stimulates a transformation towards a more fact-based culture. It ignites curiosity and a search for better answers, and at the same time, the exercise improves thinking.
After this intervention, the goal is to validate the model of how different variables affects the key metric. We have several different techniques we use, but one of the most effective, easy available and underrated techniques are natural experiments. An experiment is a procedure carried out to validate or refute a hypothesis. Performed by changing something in one area and measuring the results and compare them to a control group. Changes happens all the time. So natural experiments just captures these natural occurring learning opportunities by measuring the results and compare them to a control group.
As an example, we are working with a client, who believes their physical shops are important for sales through other channels. A belief that they have very little factual support for, which results in political decisions instead of fact-based decisions. These shops open and close frequently, so this constitutes for a very good natural experiment. All data are available it just needs to be captured, prior to a shop opening or closing. We believe that these social interventions are the missing link in the transformation towards a more fact-based and intelligent organization.
About the authors
[i] Minding the Analytics Gap, Sam Ransbotham, David Kiron, and Pamela Kirk Prentice, MIT Sloan Management Review 2017.
[ii] A Step-by-Step Guide to Smart Business, Eric T. Anderson and Duncan Simester, Harvard Business Review 2011
[iii] Winning with Data: Transform Your Culture, Empower Your People, and Shape the Future, Tomasz Tunguz and Frank Bien, 2016