Ziad Nejmeldeen has a history of crossing borders. He arrived in this country at six, fleeing war-torn Lebanon with his mother. His father, who’d risen from unlettered lawyer’s assistant to the minister of defense, remained in Beirut. Nejmeldeen didn’t speak English when he got to school, but his teachers quickly learned he was already many grades ahead in math.
School was an opportunity to try new options. In seventh grade it was finance. In high school he was both the editor of his high school newspaper and the valedictorian.
He followed with a triple major at Berkeley — in economics, mathematics, and statistics. And then — for a different, East Coast, perspective — a PhD in econometrics at MIT. But Nejmeldeen wasn’t content to be a pure theoretician. So instead of an academic career (“I absolutely loved teaching”) he chose the private sector. He wanted to see how economics could be applied to the outside world, and ended up working for technology companies such as Oracle.
When Nejmeldeen was recruited by Infor’s president to build a new data science lab for the business software provider, he knew there’d be additional borders to cross based on his recent experience working at Oracle.
At Oracle, he saw an organization struggling with a disconnect between research and product development. He recognized one group, programmers and quantitative experts, focused on perfecting its software and algorithms. And he saw another, the integrators and consultants, applying technology with customers to solve their business problems.
Today, as head of Infor’s Dynamic Science Labs, or DSL, Dr. Nejmeldeen’s idea is to eliminate that split: to create a borderless, integrated organization, where customer insights are the drivers of development. “If you want to create software customers will use,” he says, “you can’t go off on your own for a year and then come to market and hope it works.”
Infor’s Dynamic Science Lab is still young, but it is starting to live up to Nejmeldeen’s hopes and solve a range of business problems by applying data science in close collaboration with customers in healthcare, retail, financial services and other industries. In the competitive market for business software, the capability gives Infor a leg up over many of its scrappy rivals, and helps level the playing field with the industry’s giants.
“If you want to create software customers will use you can’t go off on your own for a year and then come to market and hope it works.”
Nejmeldeen and his colleagues are based out of an office near Kendall Square in Cambridge, Massachusetts, where they can tap into top engineers coming out of Harvard and MIT. Currently, the lab hosts about 20 researchers and more are on the way. Many of them work out of a large, light-filled room where connected desks make it easy to exchange ideas. A framed poster on top of a bookshelf says “Smart Data is Sexy.”
Recent projects show where the potential lies. One produced a better, more efficient way to determine hospital inventory quantities. For decades, U.S. healthcare has operated under a model of carrying inventory in excess because it was easier to do that than go lean while still meeting the demand for everything from IV solutions to beds to needles. The answer, from DSL, was a study, followed by a test, expressed as a software solution for inventory optimization. Data is taken from existing Infor solutions and run through an optimization engine to arrive at inventory targets that, on average, take 15% of inventory out of the store room and cut out-of-stock rates in half.
Another program worked out better (and more cost efficient) ways of getting fresher, tastier foods into consumer’s shopping carts. It grew out of Infor’s unique partnership with Whole Foods. The upscale grocery chain was looking for ways to increase customer loyalty and attract new shoppers with selective price cuts that wouldn’t ruin its bottom line.
The answer was writing code that gave the merchant better analytics. One outgrowth was an “attachment score” measuring how much extra shoppers spent in total when they bought discounted products. Mushrooms were a surprising winner. Price cuts were directed at food with high attachment scores, which offset the reduction in margins. “You don’t usually buy them on their own,” said Nejmeldeen. “You buy them when you’re making a bigger meal.”
Then there are the breakthroughs DSL is bringing to the science of supply chains. Data analytics isn’t new here. What is different is that DSL is analyzing the entire chain — from the start, when companies source components, to the very end when goods hit warehouses and store shelves. Shipping times are one example. The variables are huge: different routes, a range of carriers, no central source of data, plus the weather and the unexpected vagaries of international politics.
Despite these challenges, DSL is beginning to project reliable ETA’s. They’re working with GT Nexus, an Infor acquisition whose technology DSL is building on. One ingenious part is pooling information from a range of different, ordinarily competitive, suppliers. The algorithm lets them share their shipping data while protecting their proprietary systems. The result is a broader data base and more reliable time estimates. Down the road, the program will not only chart shipping delays, it will map out alternate suppliers, the best routes and the carriers that will best hit promised delivery deadlines.
It’s rare that anyone gets the chance to put his ideas into practice. But that is what Nejmeldeen has done at Dynamic Science Labs. For most of his career, he watched as data scientists were divided into theorists and practitioners. He saw them struggle with unnecessary borders. Now he’s breaking down those barriers to help businesses.
