Statistical Data Analysis

From CERN to retail? Using Big Data to make predictions

Putting a particle physicist in retail may seem like a real mismatch but one man doesn’t seem to think so. Particle physicist Professor Feindt developed the NeuroBayes algorithm at CERN, where it was used to filter out the non-useful collisions out of the millions of collisions that occur between particles every second, inside the Large Hadron Collider. Now, Feindt is applying this algorithm in retail for predicting stock control, waste management and workforce management amongst other things.

“The important thing at CERN is that huge data is created for a long time. The software that was developed for very complicated things can also be applied for dozens of completely different problems,” Feindt explains in the lounge area of the Goring Hotel in London.

Still, as Feindt admits, the transition from the world of physics to retail was not an easy one. Feindt says people would pointedly refer to his physics background and compare it to their 30 years of retail experience on how people make orders. But Feindt believed in his algorithm and to prove it he entered a “prediction competition” for predicting how many fashion items should be stocked during the fashion season.

“German retailer Otto was asking large companies to participate because they wanted to become better than their own prediction system. They asked very large software companies [in 2005] to participate and we actually won this competition by a large margin,” says Feindt.

Winning the competition gave Feindt the confidence to later set up Blue Yonder in 2008 with Otto. He says now they have lots of “positive examples” of helping clients get better but he says there still remain challenges.

“Usually you have one person at c-level who is very tech savvy but others who go by making decisions by gut instinct [are more reluctant to embrace tech]. But nowadays the business model has changed and with businesses that have small margins you can make a huge difference [by using this technology].”

But was Feindt allowed to take the NeuroBayes algorithm from CERN?

“I am lucky because I was a German professor at the time and we had something called the ‘Professor’s right’ which was that all the inventions belonged to the person. So it is mine – it does not belong to CERN or the university but to me.”

Feindt tells me that the NeuroBayes algorithm can be used for demand forecast and automated replenishment – so the client can find out what should be in the market, for how much, and when to deliver it.

“Once you have all the data, you can always monitor how good the decisions are, and also see how good the predictions are. For example, food waste is a big problem – if you order too much you have food waste. But you want to have enough so your customers are satisfied. So this is quite difficult and you need to have the best mathematics to make the prediction as good as possible. Often we make a million decisions per day.”

Some of Blue Yonder’s customers include Muller, Vodafone, and in England, EAT. Feindt says there are others in the pipeline but he’s not allowed to say who at the moment.

What does Feindt make of Big Data in the business world? Are businesses utilising it enough?

“The data is very often not fully exploited. Then of course you can use all sorts of external data from the internet like Facebook. Most of the stuff that is written on Twitter is just noise. In my personal view, this is overrated for many businesses.” Feindt explains.

For now, Feindt is keen to build up his customer base in the UK after already establishing a solid base in Germany. He thinks the UK mentality is more “progressive than the Germans” and sees the “chances more than the dangers”.

“I am convinced that in 10 years or so everyone will either [have this technology] or their business will die.”


« C-suite career advice: Zahir Dehnadi, Navabi


Viewpoint: Why we need voice biometrics »
Ayesha Salim

Ayesha Salim is Staff Writer at IDG Connect

  • twt
  • Mail


Do you think your smartphone is making you a workaholic?