Alex Ferguson’s best decision, ever
Decision making sets a good manager apart from a mediocre one. But what is a good decision? How can one distinguish a good decision from a bad one?
Let’s take as a starting point the best decision by arguably the greatest manager ever in Football: Manchester United-manager Alex Ferguson signing the French player Eric Cantona in November 1992.
United were looking for goals. A shortage of goals had costed them the Premier League in the previous season and early in the 1992-1993 season United’s goal scoring was further impaired when their striker got injured. This showed in their 10th position on the ladder, way below the teams competing for the title.
After Ferguson signed Cantona for just £1.2 million, United started to score goals, found their winning ways and even won the league that season for the first time in 26 years. Ferguson was voted manager of the year, while the eccentric Frenchman, despite him speaking little English and a string of incidents overshadowing his talent, became a legend of his own. King Eric ‘gave’ United four Premier League titles by making other players better – and score goals.
It seems almost futile to ask the question: was it a good decision to sign Cantona? If we’re interested in evaluating decision, however, the answer is not ‘duh’ but ‘that depends’. It depends on Ferguson’s intentions.
What if Ferguson’s intention was for Cantona to score goals? All of a sudden it wasn’t a great decision to sign the Frenchman, as the forward barely made it into the top ten of topscorers in the Premier League (in his best season he would score 18 goals, just over half the number that Alan Shearer put away). That United as a team started to score goals, win matches and collect prizes are beneficial side-effects of something Ferguson’s decision had not intended.
Aim of the decision
When reflecting on decisions, it is important to determine what the aim of the decision was. What did Ferguson try to improve exactly with his decision? Setting clear decision intentions, and evaluating a decision based on these intentions is a first, important step in judging whether a decision was good or bad.
But that’s not it. Even when the aim of a decision is aligned with the results – Ferguson’s intention was for Cantona to help other players score goals – this doesn’t necessarily make it a good decision. Judging decisions solely on their outcomes comes with two problems.
First, there is the attribution problem. In real-life results can seldom be attributed to a decision because many other factors play a role, including luck. In the Ferguson example: was his decision worse if Cantona had gotten injured or another team had dominated the league resulting in less wins and prizes for United? If decision outcomes are all there is to judge a decision, we’d have to say it was a bad decision to sign Cantona.
Second, there is the hypothesis problem. One can think of several scenarios that would have been better than the results of Ferguson’s decision to sign Cantona. If signing Shearer had led to even more goals, more Premier League titles, and European cups for United, does that mean that the decision that has led to ‘only’ four Premier League titles was a bad one? Again, if the results are all we have to judge a decision, we’d have to say ‘yes’.
Both problems highlight that when decisions are judged solely on their outcomes, the thinking behind a decision is neglected. For example, what were other reasonable options at the time of the decision, and how well were these options weighted against each other?
Don’t judge decisions solely on their outcomes
Judging decisions based on their outcomes doesn’t seem fair to the decision maker and isn’t very useful for improving decision making. It also doesn’t help us decide – the only thing we can do is make the decision and wait to see if it was a good or a bad one.
When reflecting on a decision, the outcomes of the decision are not the be all and end all. Bad results can be a symptom of bad decision making or bad luck.
To judge whether a decision was good or bad, we need to go back to the time the decision was made and evaluate what the decision process or hygiene was. In their book ‘Noise, a flaw in human judgment’ Daniel Kahneman, Oliver Sibony and Cass R. Sunstein provide practical advice on how to evaluate this decision process. One approach is to observe how the decision process performs when it is applied to a large number of cases.
Another recommendation by the Nobel laureate Kahneman and his colleagues is to evaluate whether the process of the decision conforms to the principles of logic or probability theory. Examples are: assessing whether alternatives are sufficiently explored, if the reasoning that had led to the decision followed logic, whether all available evidence was considered and with limited bias, or how well the problem at hand was faithfully represented.
Part of this decision process can be an evaluation of the probability of decision outcomes. How likely was Cantona’s performance going to contribute to his teammates making more goals? What were the chances at the time of the decision of Cantona not making other players score more goals, for example because his teammates would reject his eccentric, fiery character?
In future articles I will further discuss how decisions can be evaluated. The point here is to not judge decisions solely on their outcomes. To paraphrase a famous quote by Ferguson: to judge a decision on the information you had at your disposal when making the decision, not on the information that you wish you had.