Predicting the Present

futurelab default header

An interesting thought as part of the whole agile thing, is about how competitive advantage will increasingly come from not only being able to make an informed prognosis of the future, but an informed prediction of the present. Much market intelligence, and even important indicators such as retail sales data, are published weeks after the events on which they are reporting on have taken place.

If we were able to benchmark current activity quicker, make accurate predictions about the present, and react to that intelligence in a more timely way, it could lead to a not insignificant advantage.

When the Bank Of England released its latest Quarterly Bulletin (PDF) last week for example, it included an article describing how it can use search data as an economic indicator, benchmarking areas such as housing, tax, benefits and unemployment. Monitoring current economic activity closely, it said, is an important aspect of policymaking, but official economic statistics are generally published with a lag. Accepting some limitations, search data is extremely useful: it’s very timely; it’s a by-product of everyday activity as opposed to survey questions after the event; it provides continuous collection of data on an extremely broad number of areas; it has the potential to inform existing indicators as well as answering different sorts of questions; and it can be effective at helping analyse issues that arise unexpectedly.

In Bounce: How Champions Are Made, Matthew Syed makes the case that rather than some god-given talent, champions are created from a long-term commitment (around 10,000 hours worth to be precise) to focused and concentrated training. It is only this, that enables us to hone the specific qualities that distinguish greatness. One such quality, notable in sports, is the ability to perceive and anticipate outcomes faster than others. Many sports are characterised by what psychologists call “combinatorial explosion“, meaning that as a game progresses the vast number of potential moves and outcomes open to players at every stage, and the subtle ways in which those actions are inter-related, results in a rapid escalation in the number of variables. So in team and ball sports like football and tennis for example, the myriad ways in which the team players might be positioned and move forwards, or the endless potential angles or speed at which the ball might come off a boot or a racket results in layers of combinations of possibilities of overwhelming complexity.

Exceptional players learn to group huge amounts of information into a higher order construct, enabling them to identify patterns and anticipate outcomes faster, in a process that psychologists call ‘chunking‘. So brilliant football players, for example, are able to anticipate in a split second where a player is heading into space and position a perfectly weighted and directed pass into their path. This decision-making happens so naturally, and so fast, that it is almost as if they are not making decisions at all, but acting on some sixth sense or God-given intuition (hence the ‘Talent myth’). Yet the process of recognising the myriad subtle cues needed to predict what will happen next is not congenital. It can only come from thousands of hours of concentrated practice. It is, if you like, like a language that remains foreign to most of us, but one in which those who invest the time and effort become fluent. As Syed says: “Good decision making is about compressing the informational load by decoding the meaning of patterns derived from experience.”

For businesses, the rapid digitisation of products and services creates a potential combinatorial explosion of data. Deriving meaning from such ‘big data’ will increasingly about chunking – recognising the patterns and understanding quickly what is important. Thankfully, the tools we have at our disposal are becoming ever more sophisticated, not least in the advancing field of social analytics. In 2008 Google launched Google Flu Trends, an application that used search data to provide timely estimates of flu activity across 28 countries. It was a great utility that revealed the real possibilities behind this idea – it can take weeks for traditional flu surveillance systems to report on their data, but flu trends updates every day, with good estimates of activity. Flu trends also revealed that whilst tools such as Google Insights for Search were good at allowing you to enter a search term, or even a category, and see the search trends, the potential behind reversing that process and inputting data from real world activity and seeing the search terms that match that trend was even greater. So now we have (just launched) Google Correlate which does just that.

I believe that this kind of intelligence, and these kinds of tools, will increasingly create the competitive advantage of the future. So I’ve written a piece on Canvas8 that delves a little deeper into some of these ideas – you can read it here.

Original Post: