Balancing service and storage
A leading food manufacturer came to us for clarity on a fundamental question: what service levels could they achieve if they no longer used costly inventory overflow off-site storage?
The goal was twofold. First, identify the achievable service levels within the existing 45,000-pallet warehouse. Second, quantify how many additional pallets would be required to reach higher service thresholds, giving the business the data to make informed trade-offs between service levels and warehousing costs.
This wasn’t our first project with the client. We had undertaken similar work three years earlier, which gave us a strong foundation to build on, along with the opportunity to assess how conditions had changed.
Starting with what had shifted
We began by gathering data from across the supply chain network, combining our initial assessment with information extracted from the client’s ERP system. Through collaborative workshops, we shared early findings with the team which helped surface gaps, particularly around items whose data sat outside the core ERP.
A key part of the analysis was understanding what had changed since our previous engagement. Our review revealed that both forecast accuracy and production yield had deteriorated over the three-year period. This signaled that inventory buffers would likely need to increase to maintain the same service levels.
We also identified specific SKUs that had become increasingly difficult to forecast, or where production plan adherence had slipped. These items were driving much of the volatility in the planning process. By flagging them early, we could monitor for excessive inventory build-up in the final recommendations and give the business a focused list of areas to address.
Modelling service level scenarios
With the data in place, we applied our bespoke Stochastic Inventory Tool to model inventory requirements across a range of service level targets – from 90% through to 99.9%.
The tool uses historical demand variability and supply performance to generate statistically grounded inventory targets for each SKU. This approach moves the business away from rules-of-thumb or manually-tweaked safety stocks toward targets that reflect actual operating conditions.
We also segmented the analysis by ABC-XYZ product categories, showing the business how different service level choices would affect inventory requirements across high-value versus high-variability items. This gave them the flexibility to apply differentiated strategies rather than a one-size-fits-all approach.
The outputs were reviewed and sense-checked with key stakeholders before being finalised, ensuring the recommendations were grounded in both data and operational reality.
What we delivered
Following the analysis, the client had:
- Inventory targets for every SKU, calibrated against a range of service level scenarios (90%–99.9%)
- Clear visibility of the trade-offs between service level improvements and the associated increase in pallet requirements
- Segmented recommendations by ABC-XYZ category, enabling a tailored inventory strategy
- Bulk export and uploaded targets directly into their inventory management system, ready for operational use
The project also highlighted areas for ongoing improvement. By identifying the SKUs with the poorest forecast accuracy and plan adherence, we gave the business a roadmap to reduce inventory requirements further, without compromising customer service.
Our client started this project with inventory targets that had drifted over time through manual adjustments and changing conditions. They finished with targets rooted in data, a clear framework for decision-making, and the analytical tools to keep those targets current as conditions continue to evolve.
