Inside Logistics

IoT data can help reduce inventory costs: research

Simulations show a reduction of four to to 10 percent, which could translate into tens of millions of dollars in cost savings


February 3, 2017
by

ASHLAND, Massachusetts—Using Internet of Things (IoT) data to predict machine failures could reduce costly spare parts inventory stock by up to 10 percent.

Joint research by the Massachusetts Institute of Technology’s (MIT) Center for Transportation & Logistics (CTL) and OnProcess Technology also points to a shift from static to more dynamic inventory planning and the favorable impact that systematic machine data collection and monitoring could have on a company’s post-sale profitability.

“Inventory planning has long been a static, ‘review-and-stock’ endeavor. It hasn’t accounted for variabilities resulting from failures of parts in the field. As a result, it’s often inaccurate and leads to overstocking, which is expensive for business, and under-stocking, which hurts customers,” said Dan Gettens, chief analytics officer, OnProcess Technology.

“The new, dynamic, IoT-based inventory model developed as part of the MIT research is incredibly promising, as it provides a way for companies to anticipate and accommodate for failures. We believe it will also require a seismic shift in the way supply chain practitioners view service parts inventory planning.”

Many companies have started taking steps to leverage IoT data. However, data collection, which has primarily been designed to respond to signal failures, is often haphazard and subject to the willingness of buyers to participate.

More systematic collection of machine data provides a sound baseline for analyzing machine performance and predicting failures, and ultimately improves quality of service and profitability.

By improving data collection, businesses can reduce average inventory requirements and increase service levels. Simulations showed a reduction of four to to 10 percent, which could translate into tens of millions of dollars in cost savings.

“We found that even relatively weak signals that are not strong predictors of individual machine failure could provide useful and significant information when aggregated,” added Chris Caplice, executive director of MIT CTL.

“These insights could be used to improve the inventory levels and positioning for service and repair networks.”

To learn more about the OnProcess/MIT research and how it can affect IoT data collection and inventory planning, attend the complimentary webinar “How New IoT-Based Models Could Reduce Service Parts Inventory” on February 22, 2017 at 11:00AM ET. Click here to register.