The Hindsight Bias — or Why You Should Distrust Success Stories
How to act strategy-wise
Drew Housten got the idea for Dropbox during a trip in a Greyhound, where he was terribly frustrated by not having all his data available. Jeff Bezos deliberately refrained for years from generating profits for Amazon in favour of a clear growth strategy. Basecamp pursued a contrary path, in which it placed organic growth above dependence on investors. Spotify developed an organizational model that can still scare many traditional companies. All of them managed to achieve outstanding advantages and successes by following their path and many try to follow.
Should I book a bus trip to develop an outstanding product? Should I acquire investments and spend them with full hands, or should I rather bootstrap my business? And which form of organization should I choose to be successful?
Business literature is full of success stories. They all carry a promise: you can create outstanding successes if you do the same. And sometimes there is even a threat in it: if you deviate from the promised path, you will fail. The case studies are often incredibly convincing. However, they have a big catch. They describe how a single company in a unique situation has written great successes, but they hide the fact that hundreds, maybe thousands of companies might have taken the same approach and failed. It is often overlooked that special constellations were given that made an approach successful. The success stories are thus subject to the so-called hindsight bias — a misperception where the effectiveness of an approach in one company looks enormously successful in hindsight, but for others it has only very low predictive quality.
After decades of intensive management research, we still do not have a predictive model that allows us to forecast under which conditions and which strategic approach maximum success can be achieved. Such a model is also completely unlikely since markets and their rules are in constant change.
We have a good knowledge of single patterns and impact factors, but still cannot predict corporate success precisely.
We know, for example, that business models based on a network effect require fast and extensive investments for early network expansion in order to achieve sufficient critical mass. We know that companies benefit if they succeed in creating an absolute unique selling proposition, possibly even a completely new market. We know indicators for when it makes sense to pursue a first-mover strategy and when it is more sensible to wait until the first developments in a new market have taken place. And there are plenty proven strategies for successful crisis management.
There is no general and reliable formula that says: if I pursue goal A and situation B is present, I must choose strategy C, take measures D and F and I will most likely be successful. Of course, management authors earn money reliably by giving us the illusion that their success story and their procedure models guarantee exactly this certainty. As a manager or start-up founder, however, one should remain extremely critical here.
If a success model were reliably replicable, it would wear out within a very short time, since all companies that follow it would be comparable and no one would gain a strategic advantage from it.
What needs to be done? Every founder and every company have to analyse the situation for themselves and weigh up which approaches are appropriate and which ones increase success. The many success stories can provide inspiration, but nothing more. A fundamental entrepreneurial uncertainty still remains. Don’t look for promises of certainty but try to develop your business step by step based on reliable data. How can this be done?
We have learned from Steve Blank, Eric Ries, Michael Schrage and the founders of Design Thinking what fascinating effects business experiments can have in uncertain strategic situations. Through trial and error and rapid feedback from real users, it is possible to develop a product idea in such a way that the solution represents something completely new and inspires customers. Business Model Generation developed by Alexander Osterwalder and Yves Pigneur makes it possible to transfer an experimental design even to strategy development. All these approaches provide tools that are suitable for significantly reducing the risks of both founding a start-up or managing an existing business, and for guiding the company into secure and profitable paths in the long term.
Nevertheless, there is no such thing as complete security, it remains a demanding task and there is no magic formula. It is an art to manage companies strategically well, but it can be learned and mastered. Business experiments and iterative procedures make it much safer than copying a model that has led a completely different company to success under completely different conditions.
