Follow Datanami:
July 8, 2016

Making ERP Better with Big Data

(Wright Studio/Shutterstock)

At first glance, enterprise resource planning (ERP) software and big data analytics don’t appear to have much in common. But it turns out that big old ERP systems have a thing or two to learn from their newer, nimbler brethren.

One of the ERP companies exploring how to incorporate big data analytics into ERP software is Infor, which is the world’s third or fourth largest ERP vendor, behind the likes of Oracle, SAP, and (sometimes) Sage. Two years ago, Infor co-president Duncan Angove set up the Dynamic Science Labs (DSL) with the goal of using data science techniques to solve a certain class of business problems for its customers.

The Cambridge, Massachusetts-based organization employs about 20 data scientists and data analysts who use their math and coding skills to build proof of concept (POCs) for Infor customers running various Infor ERP systems, including Lawson M3, Infor CRM, Infor LN (formerly Baan), Infor SX.e, Infor Enterprise Asset Management (EAM), and A+.

Free Big Data Expertise

One of the unique aspects of the DSL is that it’s free for Infor customers. Infor offers the services of its data scientists, who build the POC and leave it for the customer to use. Infor benefits by receiving feedback from the customer on the data science application, which Infor uses to either improve the existing ERP system or to develop an add-on product that it can sell.

The DSL’s work typically falls into one of two categories, according to DSL Chief Scientist Ziad Nejmeldeen. “Our concentration mostly is where we can either provide a forecast that’s meaningful,” he tells Datanami, “or where can we provide some level of optimization to make a recommendation.”Infor DSL_logo

The first DSL engagements involved Infor ERP systems used by customers in the healthcare and distribution. Despite being entirely different industries, the business problems the DSL solved for its customers are actually quite similar. “This ability to have one team focus across all industries allows us to take the benefits of scale and to have one problem solved one place easily go over to another area entirely,” Nejmeldeen says.

In this case, Infor’s healthcare cust omer wanted to optimize its inventory across many locations. Given just two things–a master SKU (stock keeping unit) file and a time-series of inventory positions for each product and location—the DSL can make recommendations on what the optimal inventory should be.

“If you are under-inventoried, we can identify that,” Nejmeldeen says. “More often, it’s over-inventoried and you can make reductions in the inventory without adding additional risk to the service level. So you can still guarantee 99.9% availability while still reducing inventory by 15%. That’s where there’s value.”

Many companies continue to optimize their inventory levels using manual methods or using primitive tools like Excel. However, these approaches tend to be so slow and expensive that many companies can only afford to optimize their inventory once per year. By building advanced inventory optimization directly into the ERP system—or by making it an Infor add-on—the DSL customers can optimize their inventories much more often than that. In some cases, it’s done on a weekly basis.

Clean ERP Data

Inventory optimization and its close cousin, pricing optimization, are not new to the retail or distribution businesses. In fact, there are dozens, if not hundreds, of third-party companies writing software that does this. But Infor has an advantage in the fact that it doesn’t have to take data out of the ERP system, and that it can largely trust that the data stored in the ERP systems underlying relational database is clean accurate.

“One of the biggest hurdles in dealing with any new customer is all the data integration cleanup work. We can put that behind us because we can start with data that already exists in the ERP,” Nejmeldeen says. “If we can solve a problem for a particular customer…then we can create a product that does exactly the same thing with that same standard data format from that ERP, and you can do that repeatedly.”

One of the first DSL customers was Sanford Health, an integrated healthcare organization that runs 43 hospitals the Dakotas. According to Nejmeldeen, Sanford Health was the customer that led to the creation of Healthcare Inventory Optimization for Infor Lawson. Since then, two other Infor products have come out of Infor DSL, including Distributor pricing, which runs on Infor SX.e, M3, and A+, and Customer Scoring and Prioritization, which runs on Infor CRM.

Infor DSL_current_projecdtsNejmeldeen has a long list of potential projects that he’d like to tackle at DSL. One of them involves using real-time data to generate better estimated time of arrival (ETAs) of goods for companies that use its transportation and commerce ERP package, GT Nexus.

“We’d like to use that today to update the ETA that it uses it to make re-routing decisions,” Nejmeldeen says. “That’s something we’re working with the GT Nexus team to make use of that existing real-time data.”

ERP in the IoT

Other potential use case involve helping retailers become omni-channel sellers via the IoT. “As more brick and mortar retailers want to move into an environment where they can ship from store to compete with online, one of the first things that has to happen is you need accurate inventory counts,” Nejmeldeen says. “One way of doing it is to put RFID [radio frequency identification] tags on items. RFID is making a huge comeback.”RFID

Infor customers in asset intensive industries, such as trucking and equipment rental, have inquired about potential DSL engagements too. “In asset management, as sensors are being put into things that will help us predict asset failure or what kind of spare part inventory you need to have on hand,” he says. “So we’re interested in IoT data wherever it sits as an addition data feed to help us make a better recommendation.”

Nejmeldeen hopes to eventually have a team of about 50 data scientists working at DSL, a goal that is made easier by the proximity to some of the country’s top universities in the Northeast. Most data scientists have a PhD in math, statistics, or econometrics, and expertise in R and SQL, but that’s just for starters.

“The criteria that’s required is, Do you like solving problems? Are you good at it? And are you good at explaining it?” Nejmeldeen says. “This isn’t an R&D wing of development that sits behind the scenes and is given a set of requirements by product manager and builds something completely in a vacuum without customer engagement. It’s a team that’s going out and deciding how the results should be consumed by the end user and how they will be used.”

Related Items:

How Retailers Are Benefiting from Prescriptive Analytics

Trucking Along with Big Data and the IoT

How Deloitte Transforms Clients into Insight-Driven Organizations

 

Datanami