France to win the Euros, says ex-CERN scientist, who applies same algorithm to retail Credit: Image credit: photo_master2000
Statistical Data Analysis

France to win the Euros, says ex-CERN scientist, who applies same algorithm to retail

Everyone is talking about big data and analytics. Machine learning and other forms of Artificial Intelligence are getting used across a wide variety of industries to crunch numbers and spot a range of different patterns. Yet ex-CERN scientist and founder of predictive analytics company Blue Yonder, Professor Michael Feindt, believes he offers something different.  

To prove his point he has used his system to predict the winner of Euro 2016. Only time will tell if this is correct yet, but he analysed all the data from all 35,000 football games ever played since 1870 to calculate the probability of each team winning. The answer apparently is France.

“We use the example of history to predict what is going to happen tomorrow,” explains Feindt when I meet him at the Royal exchange in central London.

“Soccer is a good example because it has been going for so long. [However] in that time period the rules of soccer did not change. And human behaviour did not change” so this is a purely statistical problem. As it is based on big patterns and probability, it is doesn’t incorporate specifics like which individual footballers will play, he adds.

This is a fun illustration which aims to showcases what Blue Yonder does overall. The majority of predictions that are made are based on case-by-case-specifics, explains Feindt, which is subject to error, just like all those incorrect polls around the recent UK Brexit referendum. “I couldn’t have done it better,” clarifies Feindt, because there is not large amounts of historical data around this kind of vote.

Blue Yonder has 170 people and it is headquartered in Karlsruhe, Germany, with offices in Hamburg, London and recently Dallas. As a business it offers “decisions as a service” says Feindt.

The company’s main client base is in retail because there is a constant need across the industry for competitive advantage and “retail sells products every day for years” so has plenty of data. This said, Feindt adds, the company has worked on numerous projects across other areas, like optimising transportation, but at present there is less drive to make it happen in areas outside retail.

The first step the company takes when working with a client is to produce “predictive analytics”. This means, based on internal data, Blue Yonder creates an initial set of probabilities for different outcomes. This could be along the lines of a store having 20% likelihood of needing 20 blue jumpers in stock and 21% likelihood of needing 25 blue jumpers in stock.

The trouble is this has “almost no value to the organisation,” explains Feindt. This makes the second step “prescriptive analytics” which means the company can provide a store manager with a cost analysis attached to different scenarios so they can select the most cost effective outcome.

“That is something that we work at with the client,” says Feindt. Yet it can be tricky because “many companies don’t know their cost function” and this becomes worse in larger organisation. You get different answers for what is important based on who you ask, he explains, for example a sales manager has very different objectives from a purchase manager.

“This forces companies to think on a higher level,” he adds. “It has to go via the CEO.”

Professor Michael Feindt sees this system as part of a wider solution to improve all processes from the top down. Once applied into an organisation, he explains, “data quality quickly improves”. And, in turn, as the data improves you can make a top level strategic decision – such a CEO focus on aggressive growth – and apply it to each small decision within the company and make each one fully automated.

Feindt does not believe we should be afraid of this gradual process of automation. “Some jobs will change but the strategic decisions will stay human,” he says.

“Everyday decisions may be automated but they will be steered and monitored by humans,” he adds.

Feindt is not sure when exactly this will become normal but is adamant “the economy will go this way. It is unavoidable.”

“Those who wait too long will find themselves in trouble,” he concludes.


«Data centres & VR at 2016 Rio Olympics


VR startup sets up content creation studios»

Add Your Comment

Most Recent Comments

Our Case Studies

IDG Connect delivers full creative solutions to meet all your demand generatlon needs. These cover the full scope of options, from customized content and lead delivery through to fully integrated campaigns.


Our Marketing Research

Our in-house analyst and editorial team create a range of insights for the global marketing community. These look at IT buying preferences, the latest soclal media trends and other zeitgeist topics.



Should the government regulate Artificial Intelligence?