ICP Group

ICP Group, founded in 2015, grew rapidly through acquisitions into one of North America's largest manufacturers of coatings, adhesives, paints, and sealants. By the time leadership initiated this project, the company operated across multiple production facilities, distribution centers, hundreds of customer locations, and thousands of SKUs — a network complexity that had outpaced its internal analytical capabilities. The energy and chemicals supply chain sector faces particular margin pressure on logistics cost, compounded by the persistent disruptions that followed COVID-19. Without in-house network optimization tools, ICP could not model total cost to serve, evaluate footprint decisions, or stress-test distribution scenarios — leaving significant cost savings unidentified and supply chain resilience unmeasured.

ICP partnered with SimWell to design and implement a supply chain digital twin using anyLogistix as the core simulation and optimization engine. The architecture was structured in three deliberate layers: Alteryx handled ETL data consolidation from ICP's disparate operational systems; anyLogistix served as the modeling and scenario planning platform; and Power BI delivered results visualization for stakeholders. A critical design requirement was sustainability — the system was built to be owned and operated by ICP's internal team, explicitly avoiding the black-box consultant dependency that leadership had rejected. SimWell applied an agile development methodology, with continuous validation cycles and customer feedback loops at each iteration. The digital twin addressed three core challenge areas — supply chain footprint, asset utilization, and transportation and distribution — enabling both greenfield analysis (GFA) and full network optimization (NO) scenarios.

Network optimization runs identified an 18% reduction in supply chain costs versus ICP's historical baseline configuration, and a 7% savings against an adjusted model reflecting recently discontinued facilities. Greenfield analysis determined optimal distribution center locations across North America, quantifying how many DCs were required to serve customers within target service distances. Key outcomes:

• 18% cost savings vs. historical baseline network
• 7% cost savings vs. post-closure adjusted model
• Optimal DC footprint identified for the North American market
• ICP established repeatable, in-house scenario planning capability, reducing ongoing reliance on external consultants

The implementation was presented at the anyLogistix 2023 user conference, reflecting broader industry recognition of the approach.

• Architecture is a strategic decision: a layered ETL → simulation engine → visualization stack keeps models maintainable long after initial delivery.
• Knowledge transfer must be a contractual priority, not an afterthought — ICP structured the engagement specifically so internal teams could own and operate the model independently.
• Agile delivery cycles outperform waterfall for digital twin projects, where business priorities and network conditions shift during implementation.
• A validated baseline model is the prerequisite for any optimization scenario; without it, savings estimates lack the credibility needed for executive decision-making.
• Greenfield analysis and network optimization answer different questions — run both to separate location decisions from configuration decisions.