Blue Yonder and Michael Feindt at CERN Alumni Second Collisions
The word “yonder” has its origins in Middle English and means the far distance. It was whilst on holiday on the German island of Sylt, standing on a beautiful beach, gazing out towards the deep blue sea that Michael Feindt was inspired to select the new name for his company, formerly known as Physics Information Technology. He took a breathtaking, panoramic photo, which even showed the curvature of the earth and sent it to his colleagues with the words, ‘Blue Yonder’. The algorithm which Michael had first written in the early 2000s had been commercialized and his company was by that point a world leader in digital supply chain management. Blue Yonder was enabling retailers across the world to make smart and informed decisions, thanks to an algorithm born out of research at CERN.
Michael Feindt began his career at CERN as a fellow in April 1991, on the DELPHI experiment. This was followed by a fixed-term staff position which prolonged his affiliation for a further four years up to 1997. It was a challenge to obtain a permanent position at CERN, and after an unsuccessful application, Michael was awarded with a professorship at Karlsruhe Institute of Technology (KIT). Ironically, after accepting to move back to Germany, he was offered a permanent position at CERN.
“Essentially there are two reasons why I decided to return to Germany and not take the CERN position. Staying at CERN would have meant no physics output for several years, in the interim between LEP finishing and the LHC starting up. At KIT I had the opportunity to collaborate with CDF in FERMIlab and on the BELLE collaboration in Japan. Additionally, at the time, it was not easy for spouses to find work in France or Switzerland, so my wife and I decided to return to Germany, a decision I have never regretted.”
Measuring the energy of the b quark to supply chain and retail.
The technology, initially developed by Michael for CERN physics purposes has its roots in the algorithm ‘NeuroBayes’, which is a neural network that can learn complete probability distributions. Michael was keen to implement this algorithm within the main LHC collaborations, but he realised that simply having great technology is not enough for it to be fully embraced by all,
“One thing I learnt at CERN is that it is not enough to have great technology. One must work hard to sell the idea, and lobby for the idea. This prepared me for what was to come in the business world as in both business and science, for an idea to be accepted, you have to work really hard to sell it.”
Michael concluded early on that the algorithm he had developed initially to measure the energy of the b quark could easily be applied to other sectors,
“Physics has taught us that the world is not deterministic, and we cannot predict exactly what will happen, there will always be uncertainty. But we can learn from observations from the past. We don’t have a formula, but we have a model that can be added to with more knowledge, and parameters which can be measured. I realised that the same mathematics could be used for all complex systems across the world; banks, insurance and retail for instance “. Michael had developed the first neural network able to calculate probability distribution which was then optimized with an additional step to produce the best expected cost given certain conditions; highly valuable everywhere you have to make decisions in uncertain circumstances.
From humble origins to a worldwide vision
In 2002 Michael founded his first company, Physics Information Technology. At the time, the only employees were brilliant physicists, many of whom had attended KIT, so there was a consistency and unity within the company. They programmed in Fortran at first, then C++ and moved on to python before it was adopted in the physics world. His motivation for setting up the company was twofold. Changes in his financial situation due to a financial crisis where he lost 25% of his investments meant that he was keen to recoup those losses, but he also concluded that despite conducting statistical analysis in his professional life, he was not doing so to manage his own personal finances.
“As I had the technology, I decided to test the application in the financial domain; I downloaded some financial data from Yahoo and built a program to calculate the change of equity value. I saw that there was value in the application, I could apply it elsewhere than physics. Additionally, no one else in industry was using this solution! I certainly underestimated the work required to sell this solution to companies.”
Physics Information Technology endured a long, hard slog to build their reputation amongst established companies. They competed in and won, the data mining cup, many years which helped them make a name for themselves. Michael had a vision for the company and saw the need to expand so he appointed Uwe Weiss as CEO who advised the company to also appoint sales and marketing personnel.
“At that point, around 2014, we decided we needed to scale up and took the decision to focus on supply chain so that we could serve any retailer in the world. I had a vision of our technology spreading worldwide.”
In 2011 Physics Information Technology changed its name to Blue Yonder and in 2018, the company took a huge step, when they were acquired by the much larger US company JDA Software. The acquisition went extremely well. Blue Yonder had such a reputation in Machine Learing and Artificial Intelligence worldwide that in 2020 whole JDA was rebranded to Blue Yonder.
Looking yonder, to the future.
In 2020 COVID-19 plunged the entire world into disarray. Blue Yonder’s solutions helped many retailers survive the pandemic, thanks to the statistically significant data it was able to provide even in fast changing, chaotic circumstances. Furthermore, in a recent report, Deloitte confirmed that Blue Yonder’s product is having a direct, positive impact, in line with one of the United Nations Sustainable Development Goals; namely to half food waste by 2030. The recent announcement that Panasonic is to acquire Blue Yonder and the meeting of hardware and software promises a very bright future for the company.
Join Michael Feindt at the CERN Alumni Second Collisions reunion event from 1 – 3 October to find out more about his trajectory as a physicist and founder of one of the world’s most successful digital supply chain companies and how CERN helped shape his career.