About the client
Our client is a global supplier of micro-electronic semiconductor solutions for engineering, headquartered in Belgium. With more than 20 offices worldwide and around $1B in annual turnover, the company runs a complex, multi-echelon supply chain that combines internal and external production in a portfolio where volatility is simply part of the business. In this project, we worked closely with the client’s Business Process Owner for Inventory, within the Supply Chain Management department.
The challenge
Inside that multi-echelon network, safety stock buffers were duplicated at every level of the supply chain. While planning policies existed on paper, they were routinely overridden to avoid escalations rather than followed. The result was a familiar imbalance: too much stock tied up in intermediates, too little available in finished goods, making it harder and harder to hit service targets.
That imbalance was amplified by a data problem. Before any real inventory analysis could happen, the team had to spend considerable manual effort just to get the data into shape.
The client wanted to move away from gut feeling and human bias, toward a structured, defensible framework for inventory management, grounded in its own data and proven industry practice. The recent rollout of Databricks as an analytics platform gave them the opening they needed: a modern foundation to finally build an inventory-optimization pipeline inside their own environment.
The project
We started with a proof-of-concept, built in Honeycomb – our in-house data science platform powered by Databricks and shaped by more than 25 years of supply chain intelligence. Our team challenged both the decoupling point that dictated the client’s planning policy and the size of the safety stock itself, designing and testing eight different scenarios that accounted for the full complexity of the client’s multi-tier bill of materials.
What began as a proof-of-concept quickly grew into something bigger. It sparked internal discussions across the client’s supply chain organization: about service level differentiation, about data management, and about whether agreed processes were actually being followed. To keep that momentum going, the client’s own team needed to recalculate safety stocks, run sensitivity analysis, break down stock, segment the portfolio, and design new scenarios themselves. In-house, without waiting for external support each time.
Building that capability from scratch would have cost the client over a year, plus significant capacity from both their supply chain and data teams. Instead, they chose Honeycomb API: a modular add-on to their existing Databricks environment. It gave them direct access to our proven Honeycomb intelligence, including portfolio segmentation, outlier correction, demand and supply statistics, stock breakdown, and more without having to rebuild any of it themselves.
Six weeks later, the inventory pipeline was integrated and live inside the client’s own Databricks production environment. We combined the custom code built during the proof-of-concept with secure API connections into Honeycomb, adapting everything to fit the client’s in-house standards. Our team led the integration alongside the client’s Inventory BPO, a data analyst, and IT representatives, and closed the project with a detailed, hands-on handover.
Results
The proof-of-concept alone pointed to savings potential of up to €8.1M (a 46% reduction) while ensuring that target service levels are met. It also surfaced clear trade-offs between centralizing and decentralizing stock, and our recommendations on process refinement gave the client’s internal discussions real direction to build on.
The integration that followed delivered more than a working pipeline. It made the client’s team step up and handed the client true ownership of it. The solution is robust enough to rely on for repeated analysis, and flexible enough for their team to adapt as the business changes – without ever needing to rebuild it from scratch. It marked a meaningful step toward inventory management that is data-driven, scalable, and increasingly automated, and it positioned Databricks as a genuine layer of innovation on top of the client’s ERP.
With an ongoing Honeycomb API subscription, the client also benefits from our continued technical support, maintenance, and improvement of the underlying intelligence modules. Taking that burden off their own IT team’s plate for good.
About Honeycomb API
This project shows what’s possible when proven supply chain intelligence meets a client’s own systems, instead of replacing them. By embedding 25+ years of inventory expertise directly into the client’s Databricks environment through Honeycomb API, our team helped them gain enterprise-grade capability without an enterprise-grade rebuild. In six weeks, not a year.
This is part of a broader shift we’re seeing: more organizations want forecasting, inventory optimization, and scenario planning to run on Databricks. Not to replace their planning systems, but to extend them. As a Databricks Consulting & SI partner, we help companies bring supply chain solutions like this into production on Databricks, combining supply chain expertise with data science and AI.
If your organization is sitting on a modern data platform but rebuilding inventory logic from scratch every time, Honeycomb API can help you tap into proven supply chain intelligence instead. Learn more here or get in touch with our team to see what it could do for your environment.