Whether you call it inventory cost, total inventory cost, total cost of inventory ownership, or some other accounting term, minimizing it is hard work.

The problem is that service parts planners have no way of knowing exactly when a specific part is going to fail. Even advanced service parts forecasting algorithms are limited by the random and intermittent nature of demand.

It’s easy to dismiss any discussion about the Internet of Things as buzz, but I don’t believe that’s the case when the conversation revolves around service parts management. Using IoT data to predict when a part is likely to fail has huge implications for your inventory cost savings.

A value proposition for a medical devices manufacturer

We commonly see end user-centric IoT and predictive analytics use cases geared toward reducing machine downtime or supporting connected field service programs. I recently thought about how such use cases could benefit OEMs. I analyzed a medical device OEM’s service supply chain data to estimate the value that IoT and predictive analytics can bring to them.

I took a look at worldwide inventory distribution of x-ray tubes. While expensive, they are a critical component of CT scanners and are therefore stocked all over the world, even at locations where you may get demand only once a year!

I put myself in the shoes of a parts planner. Carrying inventory across all locations to meet infrequent demand is very expensive. It makes sense to centralize the stock at a handful of key hubs and move it to the right locations when demand is likely to occur.

A quick modeling exercise showed that centralizing inventory at five locations instead of dozens would reduce the overall inventory by 20% - which translates to well over a million dollars for each such part, all while significantly improving service levels.

Applying predictive algorithms to IoT data

The idea of predicting x-ray tube failures is decades old, and the advent of IoT has brought the math and technology to a broad spectrum of machines. Much of the focus is on accurately predicting when the failure would occur in order to reduce downtime.

My analysis shows that using predictive analytics helps reduce service parts inventory cost as you don’t need to hold stock at every location throughout the year. The savings are from the fact that you can now centralize the inventory and move it tactically based on the outcomes of predictive models. Even if the predictive models are coarse rather than well refined, you would still get the benefits by stocking the part for 2-3 months rather than all year long.

The strategy answers the longstanding challenge of stocking expensive service parts on a global scale. It shows how a company can use connected service parts management in combination with IoT data and predictive analytics to make central planning possible. A business can reduce inventory, significantly improve service levels, and offer the equipment as a service.

If you want an overview of what connected service parts management is, explore our definitive guide on the topic below: