Do we really see the world around us with our eyes or do we see it using our mental models. Behavioral scientists are concluding that when it comes to perceiving the world around us, our prejudices shaped by past experiences play a major role. This means that, we as human beings can have different perceptions from the same observation. The problem aggravates when it comes to management of social organizations. On one side, we need collaborative approach from multiple minds to achieve anything meaningful in organizations but on the other hand, perceptual differences about the same observation can lead to conflicts. It is not uncommon to find managers, having widely varying conclusions from the same set of data or two managers with exactly different view-points, having “hard” data to back up their claims. It seems we see what we believe and not the other way round.
We have a problem. “Hard” data is not helping us differentiate between truth and perception. Data was supposed to solve everything. Remember the famous quote “In God we trust and for everything else, get the data”. Just getting the data is not helping.
Data at the most validates the existence of an entity but the inferences (or deductions) from the data are made in the minds of the managers.Inferences are hypotheses of cause and effect. Most managers would admit that decisions have to be based on root cause and not just observed effects. But as we dive deeper to seek the root cause of a problem, the causality entity becomes abstract and is not directly observable with data. For example, let us take a case of chronic poor deliveries from a specific vendor. Ask for deeper causality (ask why few times) and many managers are likely to attribute it to “attitude” problem of the vendor. Consequently, a decision might be taken to change the vendor/counsel the vendor. In this example data can validate the existence of the entity “poor deliveries”. It is difficult to get any meaningful data on “undesirable attitude”. Decisions in organizations are taken based on such causality models in the minds of managers, which are shaped, by prejudices and past experiences.
For effective decision-making, do we need to have more data or an ability to develop correct mental models? Human beings are supposed to be the most intelligent living beings on the planet. So why are we failing to see reality as it is without biases?
Behavioral scientists are reminding us that we as human beings have serious flaws in rational thinking. This is primarily because, as human beings, we have been designed to take fast intuitive-heuristics based decisions to protect ourselves from any danger (we do not want to analyze too much when faced with an imminent danger) and at the same time, we also have the ability to do slow rigorous analytical evaluation in forced classroom situations.
The flaws in decisions come from using the faster mind (which “jumps to conclusions”) when one needs to use the slower analytical processor. The faster mind draws heavily from past biases and experiences.
We need to use the slow analytical processor in many situations (which are not life threatening). In organization context, the analytical processor has to also think in terms of holistic understanding and not just departments. This is primarily because the performance of any system(like an organization) is guided more by the interaction between parts (departments) rather than the property of the parts themselves. The knowledge of the interaction is more important than the detailed knowledge of the parts. But managers are formally trained only to understand parts (departments or functions) like accounting, sales, marketing, production, logistics etc. Managers are not formally trained to think, analyze and understand the cause and effect flowing through various different functional areas.
We need formal training because we, as human beings, also have poor intuitive understanding of non-linear loops of cause and effect, particularly when they are distanced by space (cause in one department creates an effect in another department) and time (the time lag in effect materializing after the cause).
How do we train our mind from falling into the trap of taking fast decisions from erroneous mental models?
Stop Confirming – Look For Falsification
The best way forward is to take the approach of the scientists. Scientists know that it just takes one contrary observation to prove a hypothesis wrong but at the same time many observations cannot prove a hypothesis right. So the best way to check reality is the test of failure. If a hypothesis has survived many tests of failure, it can be assumed to be right for the time being. As managers, we have to develop the ability to attempt and disprove our hypothesis rather than trying to confirm it.An honest attempt at failure testing can lead to us to develop a hypothesis, which reflects the reality as-is.
Think Holistic – Not Local
Departments are not just entities, which classify and groups managers in organizations, but they also compartmentalize the thought process of managers. This is primarily re-enforced by the culture of blame game and “turf” protection approach of managers. If managers train themselves in avoiding the blame game, they will see reality in a different way. Simulators can also help managers understand the cause-effect loops without falling into the trap of blame game. Computer simulators also have another advantage – they can collapse the time and space dimensions (a day of real life can be modeled as a second in computer time and a manager on a simulator can visualize the effects spanning across departments) to make managers aware of the non-linear loops of cause and effect across departments.
Beware of the Amplified Noise
Amplified noise is a minor regular uncertainty (or variation), which appears as a big problem because it is the “visible” and most “talked about” reason for a mishap.The issue of “amplified noise” prevents managers from differentiating between the real issues and the irrelevant noise. For example, if we have started late from our place to reach the meeting point, every small traffic problem appears as a big problem – the amplified noise, which otherwise we would have ignored if we had started on time. Many production managers suffering from poor on-time delivery performance complain about a plethora of problems like quality problems, absenteeism, machine breakdowns etc. as the list of reasons for poor on-time performance. However in many production shop floors, these problems get amplified when there is less available time to complete an order. The real problem in many shop floors is thehigh waiting times in front of machines in the upstream work centers, which takes away,crucial time to deal with day to day uncertainties in downstream work centers. Managers are also intuitively aware of the problem of high waiting time but this problem is not as amplified as the noise. Signals of amplified noise take away the mind from the real problem of wastage of buffers.
Using the above approaches, one can develop an ability to develop correct mental models. Now the million-dollar question, do we need lot of data for decision-making?
Being part of a system, every manager gets affected by the existence of a problem. They are aware of many data elements intuitively. For example, if there is chronic problem of stock availability issue – they are aware of the problem based on exposure to frequent customer complaints and expediting requests (even though data can showvery high service levels based on some unique definition of the term).
If the “hunch” of plant managers says their on-time performance (based on agreed definition) is around 40-50 per cent, there is no need to check if it is actually 65.2 per cent or 25.2 per cent. From the point of view of system improvement, the level of reliability of deliveries is poor in any of the scenarios and one needs a dramatic improvement in performance to create a meaningful experiential difference to the end customer. (Level of perception of deliveries of end consumer is almost the same for on-time performance of 25 per cent or 40 per cent or even 65 per cent because while placing the order, the end customer is not sure of the date of deliveries in any of the cases). Managers are always exposed to issues even at times when formal data (processed using unique definitions) camouflages the problem. Their intuition is more powerful than the data derived from the system. However their intuition is also cluttered with many non-issues and incorrect causalities. With properly structured thought process, one can get the managers to verbalize the real issues, which also occupy their minds.
A group of cross-functional managers have a wealth of collective intuition about different issues in the organization. If we are able to use the right thinking tools, the collective consciousness of a group of cross-functional managers can be guided to build the right mental model – the one which does not suffer from confirmation bias, is able to differentiate amplified noise from real issues and able to visualize the cause and effect links spanning across departmental boundaries.
It is high time organizations invested in improving the thoughtware, much more than they do for getting more and precise data elements through additional investment in hardware and software.