Intuition and Business Intelligence

In my first blog, The Recruitment Fingerprint (click here if you missed it), I claimed that business intelligence should not replace intuition, but should complement it.  Here’s a bit more background into how I reached this conclusion. 

How does intuition work?

According to Dr Daniel Kahneman’s bestselling book “Thinking, Fast and Slow”, our brain has two connected systems at work, which he calls System 1, which is our subconscious mind, and System 2, which is our consciousness. Whenever we need to make a choice, System 1 draws on past experience, or knowledge of the experiences of others, and from the decisions made in those experiences and their consequences. This is the intuitive answer. System 1 then presents this answer to System 2, our conscious mind, which then accepts or rejects the recommendation.

So it’s based on facts then?

Not always. It is based on memories that the decision maker believes are facts. Our brain needs a lot of energy when it’s working hard, and conscious reasoning is too slow to respond to situations where an immediate decision is needed, for example when driving a car or, for our ancestors, when facing a hungry leopard. So system 1 is designed to be energy and time efficient. It is amazingly good at what it does, but it has compromises in some areas significant in business.

What you see is all there is

One compromise is an inability of System 1 to recognise that facts critical to the decision may be missing. Dr Kahneman has a name for this: What you see is all there is. So, for example, a world class chess player with many years of experience playing chess can scan a board, intuitively decide on a range of choices on what move to make next, and all of these will be strong choices. The chess player will consistently make good intuitive decisions about chess. However, if the chess player has no experience of the stock exchange and decides to invest their earnings on the stock market, they are more likely to make bad decisions while they build up experience and discover factors that influence their success but that they had not considered before. 

Image: www.funnyjunksite.com

System 2 can recognise easily that a lack of knowledge will lead to bad decisions, but System 1 cannot. For System 1, what you see is all there is. It makes a decision, and leaves it up to System 2 to decide whether it’s a good one or a bad one. System 2, however, does not know what experience or memories System 1 drew its conclusion from, which is why intuition sometimes presents a choice which you feel is the right one, but you don’t know why. One of those gut feeling moments.

To mitigate against this, business intelligence arms the decision maker with the knowledge relevant to the situation, and access to the detail behind the headline KPIs. The ability to drill down into a metric, to break it down into relevant groups and categories, allows the decision maker to see precisely which measures are influencing the KPI. In addition, charts and timelines make it possible to see trending that may not have been obvious from simply looking at a headline number.

Substitution

When making decisions on a subject that we have no experience of, System 1 still tries. Instead of just drawing a blank, which would be entirely useless when that leopard is bearing down on you for the first time, System 1 searches – quickly – for memories that appear to be similar to the current situation. It substitutes a question it can’t answer for a question that it can. Perhaps you saw another tribesman being chased by a lion. This could lead to a good decision, for example you may have noticed that a lion runs faster than a man, therefore a leopard probably can too. I can’t outrun a lion, and so I shouldn’t try outrunning this leopard. It could also lead to a bad decision. The tribesman may have escaped the lion’s attack by climbing up a tree, so you hoof it up that nearby baobab, only to discover that leopards are better climbers than lions.

For our chess player dabbling on the stock market, the question that they need to answer, but cannot, would be “which company’s stock is undervalued?”. Instead System 1 will make a substitution, and possibly instead ask “which company do I admire and think will do well?” Even if the chess player answers that question correctly, they aren’t going to make a good investment if that company’s stock is already valued appropriately for its future performance.

In business intelligence terms, we can provide the chess player with more market knowledge to allow them to decide whether a company is currently undervalued or not. Performance of competitors may indicate an upturn in the market. Historical data may show that the company’s stock value traditionally rises after the release of a new product, and perhaps a new release is due. Big Data may point to growing public interest in a particular type of product, indicating a potential upturn in fortunes for companies supplying that product. 

Risk Aversion and Narrow Framing

Studies have shown that we feel the impact of financial loss roughly twice as strongly as we feel the impact of financial gain. The pain of losing £100 is twice as strong as the pleasure of gaining £100. This makes us risk averse when faced with a potential loss that could have financial gain, and conversely it makes us risk seeking when choosing between taking an inevitable loss, or a gamble that could either result in a greater loss, or no loss at all. When you’re £5000 down at the casino and faced with a crippling debt, you might feel more compelled to put the last £100 in your pocket on a roulette number and hope you get that one lucky break that drags you out of the red.

Narrow framing means that we tend to consider each risk as it comes, independently, rather than as part of a portfolio of risk. If you were presented with a gamble which has a 2 in 3 chance of winning £100 but a 1 in 3 chance of losing £100, would you take it? What if it was £1000, or £10,000? Probably not, as the potential gain does not outweigh the potential loss even though the chance of gaining is greater.  Now consider your position were you given the opportunity to take the same gamble three times? Or thirty times? It looks more attractive now, because, as long as you can afford it, the more times you take the risk the more likely you are to come out on top. Do it enough times and you will earn an average of £66 per go. Now, what if you were presented with gambles with these odds many times, but many years apart?

Put this in the context of a large company with, say twenty divisions. The managers of each division are each faced with a risky option where they have a 50 – 50 chance of losing a large amount of their capital, or earning double that amount. The managers, individually, are likely to not want to take that risk. The CEO, however, would want all of them to take the risk. 

Business intelligence helps here by removing the narrow frame, and showing which risks have good odds and which risks have bad ones. It can show which risks, if taken enough times, will eventually pay off, and which ones will eventually not (The lottery, for example. Even a jackpot winner, if they could live long enough, would lose overall if they continued betting on the lottery. They’d have to live an awfully long time though). Business intelligence allows us to remove the narrow frame that compels us to consider each risk independently, and instead see it as a component of an overall risk portfolio.

What do you think? Should Business Intelligence complement intuition rather than replace it?