Sep 13, 2017, (Research Matters):
“It is fine to celebrate success, but it is more important to heed the lessons of failure”, said Bill Gates, the founder of Microsoft and the richest man on the planet. Now, an international team of researchers has studied the failure of two different social systems - marriages and large firms – and has pointed out what lessons can one heed from other. Their analyses showed that marriages and firms are alike when it comes to failure, and this finding could help us find the relationship between these two social systems.
A social system is a set of relationships between individuals and larger groups to form a single, whole, coherent system. The members of any social system share a common purpose and interact within limited means. Some examples of a social system would be religions, kingdoms, families and firms. In short, any large group of people, with some level of organisation, may be termed a social system.
While a marriage involves two individuals, large firms may have hundreds or even thousands of employees working for them. But what is common between families and firms? One of the researchers involved in the study, Prof. Sidhartha S. Padhi, an Associate Professor at the Indian Institute of Management - Kozhikode, Kerala, explains why -- "We tried to understand the failure rate of different firms, religions, kingdoms, and crises in different countries. In the process, we found a common factor --individuals are involved in all the above social systems. Thus, we tried to come up with extreme sampling to compare the smallest system with the largest system operating, with some available data to compare."
Analyzing online open-source data from the USA, the UK, and Germany, the team constructed detailed models. Unfortunately, similar data for India was not available. The models revealed similarities between different social systems, like a marriage or a firm, especially in the way they failed. When the reasons for failures were plotted on a log-log graph, by taking logarithmic value of both the axes, the curves look alarmingly similar.
Two kinds of curves were observed in the study - fat tailed distribution and stretched exponentials. Before we understand both these curves, let’s first look at the normal distribution curve or the “bell” curve. As an example consider the performance of a class of 100 students in a test. The performance of the students, when plotted against the number of students, usually takes the shape of a bell, giving it the unusual name. Notice how the curve on both ends is very close to the bottom axis? These are the regions of the moderately extreme outcomes where a student scores a zero or a complete hundred percent on the test.
In a fat tail distribution curve, a large group of students fall somewhat below the mean passing marks. Roughly 10-15% of the students are far above the mean passing marks, and a small group is far below the mean passing mark. As such, the concept of ‘mean’ becomes meaningless. Unlike the normal distribution or ‘bell’ curve, moderately extreme outcomes are more likely than expected. This means that the probability of student scoring a zero or a hundred is higher than in a bell curve. Or in the case of this study, probabilities of failure of a social system due to unlikely outcome were found to be much higher.
Prof. Padhi clarifies, "We propose that fat-tailed distributions of failure result from individuals' least efforts to maintain a social system whether it is a small system of two individuals, like a marriage, or a set of individuals working in a firm. This can be explained using the Zipf's law.”
Zipf law can be understood in the context of the number of times a word occurs in normal conversations. If the most frequently used word is uttered 1000 times on a given day, then the second most frequent word is used around 250 times, the third and fourth most frequent word around 50 and 10 times, and so on. This, when plotted, gives us a power law curve and much steep-curve than an exponential curve. Social systems show a similar scaling when different reasons for failure are considered.
Prof. Padhi adds, "It is difficult to get the actual data or proxy for the Indian context. However, these conclusions still hold for any country. Moreover, we learnt that in India, the failure of marriage is 10 to 20%, which is an aggregate measure. But the main reasons are the same."
The study provides clues as to how different social systems behave similarly, irrespective of their size. Whether a marriage of two or a firm of thousands of individuals, there are lessons that one can learn from the other, either to keep their marriage intact or to keep their firms functioning.