How Böllhoff connected 20 million parts per day into one AI data layer

Germany

Fastener Manufacturing

3,000+ employees

700-800M EUR revenue

14 production sites worldwide

14 production sites, 20 million parts per day, and no shared layer connecting machine data, energy readings, and production orders. Read how UMH changed that in under a month.

"We produce almost 20 million fasteners per day. That generates enormous amounts of data. Machine data, energy data, production orders. But it was all in separate systems. We had the data. We just couldn't use it."

Head of Processes, Digitalisation and Applications

1 month

First use case live

Established

AI data foundation

Energy cost reduction

-85%

Time to collect data

The challenge

Siloed Data Blocking Data-Driven Operations

Separate systems per plant, no shared layer, every new site rebuilt from scratch.

Data existed, but in silos, with no shared version across departments

Böllhoff generates nearly 20 million parts per day across its plants. Machine parameters, energy meters, production orders. All of it existed, but in separate systems with no shared layer. OT data stayed in OT. IT data stayed in IT. Management received spreadsheets, weeks after the fact. When departments compared figures, the discussion was about whose data was right, not about what to do next.

Every question was a project, with no self-service for the shop floor

Answering a basic operational question, say the energy consumption of a specific line or a performance comparison between two machines, required exporting data from one system, pulling it from another, and reconciling in a spreadsheet. The cost of asking a question was too high for anyone to ask often. Line managers and operators who needed answers had to request them from central IT. By the time a report was built and delivered, the question was often no longer relevant.

Every plant was a restart, with no way to replicate what worked

Each Böllhoff site had different machines, different local systems, and different levels of OT infrastructure. A configuration that worked at one plant could not be transferred to the next. Connection logic, dashboard setups, and data models had to be rebuilt from scratch every time. For a company with 14 production sites worldwide, this made any centralized data initiative impractical. The effort multiplied with every site instead of decreasing.

THE SOLUTION

One platform, replicated across plants

One platform, managed by three people, with templates that replicate across sites.

Fast deployment with a standard architecture

UMH runs as a single open-source platform covering data collection, storage, visualization, and analysis. Böllhoff retains full control over architecture and data, with no vendor lock-in. After the decision was made, the first use case was live in less than a month. Multiple European plants were connected using a template-based rollout: same architecture, adapted per machine type.

Unified Namespace as the shared data model

All data, machine parameters, energy readings, and production orders, flows into a single Unified Namespace. The Management Console provides one topic tree view across all connected systems. The data that used to live in separate OT, IT, and MES systems now shares one structure.

Self-service for shop floor teams

A team of three from IT and OT set up the platform themselves, without external consultants, and manages it across all sites. They build initial configurations and first dashboards. From there, shop floor staff are trained to build their own views without filing tickets or waiting on central IT. A question that used to require three exports and a spreadsheet now takes a click.

Architecture