Supply-chain orchestration for a plastics processor

An internationally active plastics-extrusion group suffered from irregular order fulfilment because the production planning of its four large European plants ran deeply isolated, side by side. Central procurement could reliably predict neither granulate consumption nor machine availability — leading to expensive material hoarding and frequent rush surcharges in shipping.

Each plant manager worked with highly complex Excel workbooks grown over fifteen years and full of nested macros. These tools captured local machine quirks, humidity fluctuations during curing and regional driver availability with precision. There was firm resistance to the corporate mandate to introduce a rigid, standardized global ERP planning module — central algorithms would ignore decisive local factors.

The task was to connect the sites into a robust group-wide view without destroying local autonomy and the experiential knowledge stored within it.

We took on the role of central enterprise architects and decided to turn the local Excel workbooks into decentralized input sources rather than abolish them. We designed a cloud-based ingestion pipeline in which plant managers kept working in their familiar spreadsheets: the system read the native files, parsed the variables automatically and aggregated everything into a unified global forecasting engine.

As a three-person team, we defined the overarching data schemas, API specifications and governance rules. We steered the implementation of the validation and ingestion microservices together with an offshore development partner. Python-based parsers extracted the calculations directly from the uploaded spreadsheets and fed a central cloud data warehouse that delivered near-real-time dashboards to executives.

The plant managers came fully on board because their historical autonomy and their specialized formulas were preserved — acceptance across all four sites was complete. Four isolated planning islands became one shared, robust data foundation.

Material hoarding fell noticeably and group-wide predictability of lead times rose markedly, because procurement could plan across all plants for the first time — without a single site having to give up its proven way of working.