Algorithm in the future of banking

This blogpost is based on an article in Harvard Business Review January-February. This again is based on the book Algorithm Need Managers too, by Michael Luca, John Kleinberg and Sendhi Mullainathan.

Algorithm and robots are on the agenda for most banks and financial institutions today. Big Data analytic can aggregate large number of data, both from internal sources in the bank, from the market (stock exchanges, Forex markets, interest market and other markets) and other external sources. It is the way we use the date and how we design our algorithm that makes the difference.  And that we do understand what algorithm is not good at.

Netflix had a million dollar competition some years ago to develop an algorithm that could identify which movies a given user would like, teams of data sciences joined forces and produced a winner. But it was one that applied to DVD – and as Netflix user’s transitioned to streaming movies, their preference shifted in ways that didn’t match the algorithm prediction. It is a difference between movie one would like to buy and the movie one would like to stream now.

Within marketing there are algorithms that estimate what information users are likely to click on. Google have such algorithm. The target of clicks for the marketing company is to sell more, generate revenue and profit. Most of these clickable algorithm does not estimate efficiency of sale and profit, “only” what generates most clicks. It is important to know what these kinds of algorithms is good at and what they are not so good at.  

In the latest Avengers movie, Tony Stark (also known as Iron Man) created Ultron, an artificial intelligence defense system tasked with protecting earth. But Ultron interprets the task literally, concluding that the best way to save the earth is to destroy all humans.

This is the core of an algorithm – is does exactly what it is designed to do. We human might have soft goal. For example it might be best to have a short time loss, for a long time profit. We might strive for equality – even if it causes organizational pain in the short term.

Often one algorithm is used in different situation. There are a lot of example where for example a marketing algorithm works perfect in one country, but not at all in another country. So - design you algorithm for each situation is of extreme importance.

Also remember that correlation still doesn’t mean causation. Suppose that an algorithm predicts that short tweets will get re tweet more often than longer ones. This does not in any way suggest that you should shorten you tweets. This is a prediction, not advice. It works as a prediction because there are many other factors that correlate with short tweets than make them effective. This is also why it fails as advice: Shortening your tweets will not necessarily change those other factors.  

New technology like Blockchain enables banks to easier automate and leverage on technologies like robots and algorithm.