Blog-hero.jpg

The Innovator's Solution

     

Industrial Supply Chain Planning is Different. Here’s How.

Posted by Jeff Bodenstab on Mar 16, 2017 5:45:00 PM

Benchmark.gif

Two recent benchmarking surveys, one by Gartner and another by ToolsGroup, identified specific issues that make industrial supply chain planning different. Here’s what they discovered.

Gartner surveyed 112 industrial supply chains across North America, Western Europe and Asia/Pacific. They found a median demand forecast error of 26%, indicating industrial manufacturers “are struggling with variability that can translate up the supply chain and impact cost and service.” The benchmark data showed a correlation between this forecast error and orders on-time-in-full (OTIF). Hence, more than half of industrial manufacturers delivered less than 91% of their OTIF, “highlighting the difficulty aligning supply capacity and inventory to demand across often complex product offerings.”

In a multi-industry survey of more than 300 companies in its customer base, ToolsGroup found that industrial companies had forecast accuracy in the high 60s, well below other verticals such as consumer products and retailers. Unique challenges contributing to the shortfall include large bills of material and item variations, lengthy lead times, multitier supply chains and slow-moving aftermarket spare parts. However service levels were around 95%, indicating that the customers were better able to cope with these challenges. ToolsGroup also found that demand planners in the industrial sector have a lot on their plate, typically managing 18,000 SKUs. For more on this survey, see our recent blog post located here.

A benchmark in surveying (see photo above) is a point of reference defined by position and elevation. A benchmark in business helps establish a point of reference to help you understand and improve your operations. Here is what these surveys identified.

First, for more accurate demand forecasting, Gartner says cross-functional teams should make trade-offs that benefit the entire organization. Otherwise, sales groups locked in on revenue may overlook inventory costs when submitting an overly optimistic forecast. Or logistics, wanting to optimize costs with full truckloads, may downplay lower service levels, making customers wait for orders. Gartner says that industrial manufacturers can lower forecast errors by strengthening this collaboration across business areas with improved demand visibility and demand shaping. Demand visibility can be enhanced by point-of-sale data, IoT signals and channel information. Demand can be shaped through things like promotional pricing to “steer” customer orders.

Second, they identify the need for advanced demand planning software that can automatically generate the baseline forecasts. Planners at industrial companies with large portfolios and complex distribution networks are spread too thin to do a good job across all SKUs. But systems with powerful statistical engines (and even machine learning) can crunch huge quantities of incoming demand data, “learn” from past successes and errors to improve the approach, and analyze variability and patterns to better predict upcoming needs. Planners and managers can then fine-tune the plan with their knowledge of the business.

Third, the report says that it’s important to get the forecast in granular detail, down to the SKU and item/location level, as “this drives the supply chain decisions and actions before the actual customer demand is known.” Analytics can help in determining the causes of demand.

The Gartner report offers Lennox Residential, a leading provider of home heating and cooling systems, as an example of an industrial company who reduced forecast error despite a challenging supply chain. They faced a 250% increase in the size of their aftermarket parts distribution network while dealing with highly seasonal products and frequent new product introductions superseding existing product lines.

Lennox exploited analytics in its demand planning software that includes an embedded machine-learning algorithm that Gartner says “clusters products based on attributes, like locations and climate zones, to generate an accurate demand forecast.” Analytics also set safety stock levels that optimize service levels and revenue growth while keeping inventory to a minimum.   Lennox reduced inventory by 20% and improved service levels by 20%. It lowered distribution costs as a percentage of sales by more than 15% and increased the number of orders it can deliver the next morning from 35% to 98%.


 

Click below for more on the Lennox story:

Lennox Case Study

Topics: Forecasting Demand and Analytics