I get paid to make important decisions, often impacting many people or involving large sums of money, and I like to think I’m fairly good at it. The basis of economic theory is that people operate in a rational and logical manner to maximise their potential outcome. Indeed, I think most people like me would say they largely behave rationally and make good decisions. Well here’s a newsflash to us all, people are terrible decision makers and have poor intuitive judgement.
In his ground breaking book ‘Thinking Fast and Slow’ Nobel Prize winning Daniel Kahneman demonstrates just how illogical and inconsistent we are. The book makes for uncomfortable reading for experts or company executives. Our brain relies on two systems; a fast thinking, efficient and subconscious system determining the majority of our decisions; and a conscious but cumbersome lazy system that intervenes when more rigorous thought is required. The challenge is the first system works fast and takes shortcuts so the second system is often not triggered. These shortcuts manifest themselves in us all being worryingly erratic in our judgement.
Kahneman demonstrates through countless experiments that people make different decisions based on altering how a question is framed, what they’ve seen that day or even whether they are hungry. You probably think you are too experienced or intelligent to get caught out by such simple tricks, but serious misjudgement happens all around us. For example a study in Sweden found 93% of Americans believe they are an above average driver; Kahneman references research showing High Court Judges taking different decisions before and after lunch; and Philip Tetlock in his book ‘Expert Political Judgement’ alarmingly found that ‘experts’ commentating on political and economic trends made worse predictions than people with NO knowledge at all.
Fortunately, help is at hand, the robots have landed.
The Robots have Landed
Technology like Big Data, AI, Machine Learning, Deep Learning and IoT are fast becoming commonplace and the combination of them enables machines to think and behave like humans. Or to be more accurate, not think like humans. The technology can understand natural language, analyse information, reason like people, take decisions and importantly learn. What sets machines apart is that they analyse zettabytes of data in a split second, have none of our inbuilt bias’, and make perfectly rational and consistent decisions.
As Viktor Mayer-Schonberger and Keneth Cukier discuss in their insightful book, Big Data, simple algorithms utilising huge messy data sets will almost always outperform humans in making decisions. Organisations now rely on machines to make decisions previously made by people. Companies are using IBM Watson to calculate bonuses and pay increases for employees with little human intervention. UPS use analytics to plan and optimise routes and maintenance schedules saving millions of dollars and tonnes of carbon dioxide emissions in the process. Amazon’s patented anticipatory shipping model determines what people will buy and when in order to send products to their distribution centres. Even in industries where the opinion of the ‘expert’ is king, such as wine tasting, machines have been shown to make better judgements as demonstrated by Orley Ashenfelter’s algorithm to predict the price of fine wines.
The predictive ability of computers and the extent to which they outperform people will only increase. So, is this the end of human decision making in organisations. A future where computers decide and humans enact?
The End of Human Decision Making?
This vision probably evoked an emotional reaction and images of computers ruling the world, but I personally don’t believe that will ever happen. If we look past the emotional for a moment, I’d like to propose some augments for leaders retaining some level of interaction and control in the decision making process.
Firstly, if we allow machines to run our organisations on our behalves we are setting off down a dangerous path. This could lead to a future where there is no human accountability for the actions of companies and leaders hide behind the algorithms that are ‘thinking’ for them. Image a world where machines could take decisions leading to another Enron disaster or Global Financial Crisis. With no individual accountability or repercussions these events will become more frequent and catastrophic.
Secondly, the more we turn people into data the more we dehumanise organisations which can be extremely dangerous as discussed by Simon Sinek. The further removed leaders are from the people impacted by decisions the easier it is to take drastic actions such as enacting layoffs. This could have extreme consequences on the health, wellbeing and productivity of society.
Lastly, we will stop driving society forwards. In a world of sophisticated technology and easily accessible ‘perfect’ information everyone will make good logical but similar decisions and strategies. Companies will be hugely more productive and efficient, but infinitely less innovative. No algorithm would decide to make the iPad, hire Steve Jobs or Bill Gates, develop Facebook or come up with an idea like Airbnb. Over time organisations with an overreliance on analytics will steam ahead very efficiently but uncreatively to their demise.
So my message is not to rely on analytics blindly and retain some control? Well, not quite. We need to work in symbiosis with machines and while this is not a new idea we need to consider what it means. We’ll need new business models and ways of working and not simply replace human tasks with automated solutions.
Re-Thinking the Future
If your objective is to simply use the technology to redistribute current activities between people and robots you’re missing the point. Computers are infinitely better at some things than people and visa versa, so with this in mind lets reimagine the organisation.
- Empower the Robots
Just like you do with members of your team give the robots the tasks they’re good at and let them do it without interference. This includes large parts of finance, HR (especially recruitment), marketing, supply chain etc. Virtually anything operational, and especially tasks where people make comments like ‘it needs expert judgement’ or ‘gut instinct’.
- Transparency on Steroids
Computer generated decisions can’t be treated like a black box. In a world of Glassdoor, social media and comparison platforms people demand more transparency. The criteria used for the decision and the high-level analytics (as appropriate) to support the decision should be available and easy to find.
If I missed out on a role I’d rather know it was because I hadn’t worked in X industry, only had Y years experience and the analytics show people with my profile didn’t perform as well as others for this role. This may seem uncomfortable in today’s world but it’s what goes on in recruiter’s heads today. The difference being these rules are now applied consistently, fairly and most importantly transparently.
- Increase Leader Accountability
We have to get used to machines making decisions for us and accepting that in many situations they’ll do a better job. However, freeing up leaders from onerous decision making means they’ll have more time to understand what decisions are made and why. Leaders will have greater visibility of more of the decisions across the organisation and importantly visibility of the criteria on which the decisions were made with confidence they were fairly applied.
Organisations should set up new internal groups to regularly and independently audit algorithms and provide assurance. Algorithms should also be regularly reviewed and updated to make sure they are still fit for purpose.
- Re-humanise the Organisation
The role of the leader needs to be redefined. With time freed up from the day to day mundane decision making process leaders should spend more time leading. This means spending time face to face with people; employees, customers or suppliers. Technology used as an enabler to connect leaders and people not the reverse, as seems to be the case today. Meeting people and listening needs to become a part of a leader’s role once more.
- Innovate, Innovate, Innovate
Innovation and speed is the new currency of competitive advantage. Organisations should set up new business units outside of their governance processes with the objective to innovate. Recruitment should be people that do NOT fit the algorithm and they should have free rein to experiment and allow ‘good’ failures as Atlassian do. These business units will engage the community outside the walls of the organisation, gamify ideas to select the best options and bid for seed funding or crowd source investment. This will allow the identification of the next disruption or creative genius that no computer programme can ever find.
I don’t subscribe to the ‘robots will take our jobs’ movement for a second. For every role automated infinite opportunities, new roles and industries are created. But this will never happen while we just look to simply replace human tasks with computer-generated outputs. We need to redesign our organisations.