Highly-complex, globally-dispersed medical device supply chains continue to be under extreme pressure due to the COVID-19 pandemic and supply chain crisis. We worked with a world-renowned medical device company that was facing surplus inventory, surging logistics costs and struggling to meet customer demand.
We used Faculty Frontier to help improve the accuracy of their demand forecasting, enabling better decision-making and facilitating critical trade-offs based on leading market indicators.
It operates in over 150 countries
A Fortune 500 medical device company
It manufactures well over 40,000 products
Off the shelf demand forecasting tools were unable to cope with the complexity of their business model and the volatility of the markets they operate in, creating inventory surpluses and driving up costs.
Supply chains for regulated products like medical devices are extraordinarily complicated. Assembly involves sourcing parts from multiple countries, manufacturing in specialist locations, and shipping to distribution centres across the globe. Even without any shocks, these are complicated operations. But the rise in demand for medical devices sparked by COVID-19, coupled with high air freight costs, exposed vulnerabilities. Off the shelf demand forecasting only told demand planners what happened yesterday.
Our customer produced tens of thousands of products with unique demand patterns in each region. The team was using patchy sales data from a range of markets for their forecasts and manually adjusting them in Microsoft Excel – inviting human error and wasting valuable time. Their inaccurate demand forecasts and error-riddled predictions led to reactive decision-making, surplus inventory and SKU fragmentation. Faculty was brought in to help deliver more trustworthy demand forecasts in both mature and emerging markets to serve their customers better.
Faculty Frontier delivered accurate and actionable forecasts across the company’s rapidly-changing, multi-tiered supply chains.
We built a picture of the company’s strategic priorities and current processes, and seamlessly embedded Frontier into their existing technology and data infrastructure.
Our forecasts went down to the individual SKU-level
Faculty Frontier was able to generate granular forecasts at a SKU-level so demand planners received up-to-date, data-driven insights. We could make predictions even with limited data thanks to our highly-advanced machine learning.
We integrated external data sources
We further enhanced prediction accuracy with additional external sources such as distributor data for indirect markets and customer usage data. We also tested other leading indicators of demand, such as weather, but they didn’t produce sufficient signal – we only added datasets that could generate better results.
We improved confidence in forecasts
We prioritised providing hyper-accurate demand forecasts and building the team’s confidence in using statistical forecasts. Taken together, this would enable them to drive better decisions across their supply chain operations, correcting challenges in a globally diversified supply chain.
We delivered these forecasts through interactive dashboards which could be adjusted to model different scenarios, with each forecast clearly explaining the underlying drivers of demand. Adjustments to the forecasts were approved by a financial planner to create a better audit trail, so forecasts were updated in a more controlled way. As users were able to plan for scenarios, compare forecasts with actuals and understand why predictions were made, Frontier fostered more confidence in decision making.
Our forecasting performance went beyond the capabilities of their existing software stack, securing stakeholder confidence in demand forecasting and driving millions in projected cost savings and revenue growth potential.
By improving the accuracy of demand forecasting, our customer could save millions across inventory and logistics in just one region. Across the markets and product portfolio tested, forecasts exceeded the performance of existing software by an average of 20%. Most importantly, this new way of forecasting had significant potential to support the business to improve product availability to key customer accounts, thus driving top line growth.
improvement in forecast accuracy (Weighted
mean absolute percentage error)
Case study: Supply Chain
Creating hyper-accurate demand forecasts for one of the world’s largest medical device companies
Get in touch