Cloud based intelligence could be key to retail's analytical future

Blue Yonder touts power of cloud-based data for improved decision-making

Blue Yonder, the business intelligence cloud computing company, recently revealed that it has partnered with Microsoft Azure as its primary platform.

The predictive analytics company, which has its origins at the CERN facility in Geneva, uses the cloud to calculate variables at speed, primarily aimed at allowing retailers to make decisions on what to buy, what to stock and what to make in order to ensure that when you look for something in store or online, it'll be there.

"The fast-growing amount of data from customer transactions and replenishment processes makes it imperative that retailers switch to automated decisions based on data," says Jan Karstens, CTO of Blue Yonder.

"The high expectations of consumers in the face of contemporary cross-channel concepts like Click & Collect , where items are ordered online and picked up in store, can only be met using machine learning."

By crunching data ranging from the weather in previous years, to customer footfall, perhaps due to known factors like transport links to a particular branch, the company has been able to perfect a technique that uses machine learning to constantly improve stock availability and decrease waste.

"Microsoft Azure enables retailers to use artificial intelligence not only to optimise existing processes, but also to pioneer new business models that will drive the transformation in their industry," said Sabine Bendiek, Area Vice President Microsoft Germany.

Brands including Morrisons, Otto, Kautland, Eat and Natsu are using the Blue Yonder platform, powered by the Microsoft Cloud, leading to a string of awards from across the sector.

Last year, V3's sister title The INQUIRER looked at Eat, the sandwich chain with over 100 stores across London and beyond, which was able to demonstrate a 14 per cent reduction in waste during the initial trials of the service.

EAT CFO Strahan Wilson told us, "The skills for predictive forecasting are very specific and very technical. We lacked the skills in-house to build the model and, even if we paid someone to build it, we don't have the skills to maintain it,"

And that's part of the point. What Blue Yonder is doing is so far removed from anything that 99.9 per cent of companies could do in house, it isn't worth even trying.

"I'm desperately trying to keep a lid on the excitement that this is building in the business. Everyone can see the possibilities and I'm just begging people to let me get to July when we go full rollout before we start to look at just how much it can actually do," said Wilson at the time.

More recently, German retailer Otto has shown great success in controlling stock using Blue Yonder services to measure the price that the public is likely to be willing to pay in the current climate (figuratively and literally speaking).

It is currently looking at ways to predict stock levels to optimise delivery to the customer in the fastest ways possible, by ensuring that the right products are at the most accessible places at the right time.

Cloud based intelligence could be key to retail's analytical future

Blue Yonder touts power of cloud-based data for improved decision-making

Professor Dr Michael Feindt, founder and CEO of Blue Yonder told us, "Machine learning can be used to make effective predictions in any business where there is enough data. It doesn't give you a crystal ball into the future - we do not live in a deterministic world and tomorrow is always uncertain.

"But modern algorithms can make predictions in form of probabilities for each possible future, including the mean trend but also individual uncertainty. Clever algorithms then can make optimal decisions on the basis of these probability distributions."

"No matter how advanced AI may get, it cannot predict the exact outcome of the future. But machines can draw decisions under uncertainty, quantify it and determine a range of probabilities. Only if businesses have this information they can make the best decisions today for tomorrow."

In order to achieve this, Blue Yonder needs access to the most reliable, robust, scalable and elastic cloud services to cope with the billions of variable it is analysing at any given moment. Feindt continues,

"In retail, Blue Yonder makes over 600 million daily predictions to help retailers make decisions around pricing and replenishment. How many tomatoes should I have in stock tomorrow at a particular store? What is the optimal price I can sell this to my customers today, taking into account predicted demand, competitor pricing and price elasticity? These are the kinds of accurate decisions that predictive analytics can enable on a daily basis.

"This technology already is successful in retail, but in principle it can be much broader. Having its roots and successful applications in fundamental physics research, machine learning has found applications in replenishment, pricing, customer targeting, finance, insurance and medicine. "

More and more, we will find in the coming years that decisions are made based on the "best guesses" of computer systems housed in the cloud.

"We chose one of the leading providers of cloud technologies," said Karstens. "Microsoft Azure's global availability helps us secure our international growth and make our machine learning solutions permanently available. Another key factor in this decision is Microsoft's reputation as a globally reliable and respected partner that values data security. This is incredibly important for retailers.

"Blue Yonder is in the process of migrating customers on heritage systems to the Microsoft Azure . The cooperation with Microsoft gives us - above all - the possibility to expand our cloud and machine learning acquisition efforts to an international level," he concluded.