UK Manufacturing's Data Problem Is Not the Same as Europe's

Brexit did not just create customs paperwork. It created a set of data integration requirements that UK manufacturers are still working through four years later. Exports to the EU now require customs declarations, rules of origin documentation, and commodity coding that was invisible inside the single market. For manufacturers with EU customers, this means connecting the ERP's sales order data to a customs management system that did not previously exist in the data architecture.

The rules of origin requirement is particularly data-intensive. UK manufacturers selling into the EU under the Trade and Cooperation Agreement must demonstrate that their products contain sufficient UK or EU content — thresholds vary 45–60% by HS code and tighten over time. The auto sector is the clearest example: the maximum non-originating content was 60% (2021–2023), 55% (2024–2026) and drops to 45% from 2027, with separate, tighter rules for EV batteries. Proving compliance requires traceability data linking finished goods back to raw material sourcing, with country of origin tracked at the component level. Most UK manufacturer ERPs were not configured to capture this systematically before 2021, which means the compliance data trail has gaps that must be addressed.

Beyond customs, the supplier network disruption caused by Brexit created a secondary data challenge: many UK manufacturers have added new domestic or Commonwealth suppliers to reduce EU dependency, and those new suppliers use different data formats, different lead time structures, and different quality certification systems. Integrating them into the existing procurement and planning data model is a data engineering task, not just a supplier relationship task.

The 2021–2023 energy crisis hit UK manufacturers disproportionately hard. UK industrial energy prices rose faster and higher than continental European equivalents, and the exposure was concentrated in energy-intensive sectors: food processing, glass, ceramics, chemicals, and primary metals. For manufacturers in those sectors, energy became the largest variable cost - and most of them had no granular visibility into energy consumption at the process or product level.

Cost per unit calculations that had been stable for years became unreliable because the energy component was not being measured at sufficient granularity. A bakery that knew its average energy cost per loaf was not tracking whether the proving ovens, the baking ovens, or the cooling conveyors were driving cost increases - because the sub-metering data was not connected to the production system. Finance knew the energy bill. Operations knew the production volume. Nobody could see the relationship between the two at a granular enough level to make targeted improvements.

Sub-metering connected to the production data platform is the technical solution. Smart meters on individual production assets - linked via OPC-UA or pulse counting to a data historian - provide the granular energy consumption data that, when joined with production order data from the ERP or MES, produces cost per unit by process step. UK energy monitoring regulations (ESOS) already require large manufacturers to conduct energy audits. Building a live energy monitoring capability replaces the periodic audit with continuous visibility - and turns an energy audit into an energy management tool.

The Made Smarter programme — UKRI and DSIT-funded (originally BEIS, before its February 2023 abolition and split into DSIT/DBT) digital transformation support for UK manufacturers — has been effective at catalysing investment in industrial digitalisation, particularly among SME manufacturers in the North West, Midlands and Yorkshire. The Adoption programme provides match funding up to 50% (typically capped around £20,000 per grant) and has engaged thousands of manufacturers since the 2019 North West pilot. Made Smarter Innovation has invested £112m in grants plus £200m+ in industry co-investment.

The gap the programme has not fully addressed is data infrastructure. Made Smarter funding has primarily flowed to robotics, cobots, and visible shop-floor technology - investments that are tangible, photographable, and easy to communicate to stakeholders. The data platform investments that make those technologies useful - the OT data collection layer, the semantic model, the cross-system integration - are less visible and harder to justify on a standalone funding application.

The consequence is that many Made Smarter beneficiaries have installed new technology without the data infrastructure to measure its impact. A cobot deployment without production data integration cannot demonstrate OEE improvement. A new MES without a connection to the ERP cannot close the loop on production order costing. The hardware is in place. The data foundation that makes it legible to the business is not.

UK manufacturers sit in a specific technology context. Microsoft 365 adoption is high - most manufacturers already use Teams, SharePoint, and often Power BI. The ERP landscape is fragmented: Sage 200 and Sage X3 are common among SMEs, SAP Business One is present in the mid-market, and Dynamics 365 is growing. On the shop floor, a mix of Siemens, Rockwell, and Mitsubishi PLCs is typical, with SCADA systems ranging from modern WinCC and Ignition deployments to legacy Wonderware.

The architecture that fits this context is Microsoft-centric: Power BI as the reporting layer (already familiar), Microsoft Fabric as the analytical platform (native integration with the M365 estate), and Azure IoT Hub or Event Hubs as the OT data ingestion layer. This is not a default Microsoft endorsement - it is an acknowledgement that the integration complexity is lowest when the analytical platform shares a security model, identity system, and licence structure with the tools the manufacturer already pays for.

The OT integration approach depends on the PLC generation. For Siemens S7 and TIA Portal environments, OPC-UA is standard. For older Rockwell or legacy Mitsubishi environments, a Kepware or similar middleware layer translates proprietary protocols to MQTT or OPC-UA before the data reaches the cloud ingestion layer. The edge gateway - a Raspberry Pi 4 or industrial PC running Azure IoT Edge - buffers data locally and manages the OT-to-cloud handoff without touching the production network directly.

The rules-of-origin trail is the hardest constraint, because the data gap is historical and cannot be backfilled. If component-level country-of-origin was never captured before 2021, no platform reconstructs it retrospectively — you can instrument it going forward, but the legacy compliance trail has holes that are a procurement and supplier-data exercise to close, not an analytics one. Promising clean rules-of-origin traceability without first fixing capture at the component level is selling a report the underlying data cannot yet support.

The new-supplier problem is similarly organisational. Onboarding domestic or Commonwealth suppliers to reduce EU dependency means new data formats, lead-time structures, and certification systems — and conforming those into the procurement and planning model is master-data work that someone has to own. The integration layer reads whatever the supplier sends; it does not standardise a supplier who sends inconsistent data. Under-resourced UK IT teams are exactly where that ownership falls through, which is the argument for a Fractional CDO rather than another tool.

And energy sub-metering carries a real capex and install constraint. Smart meters on individual assets, wired via OPC-UA or pulse counting to a historian, are modest but not free — and on an older site the metering retrofit is a scheduled-downtime job, not a software toggle. The analytics turns the data into cost-per-unit-by-process once it flows; it does not generate the sub-metered signal that was never installed. Scope the metering as part of the build, not an assumption.

The starting point for most UK manufacturers is a data audit rather than a technology selection. Before any platform investment, it is worth mapping which operational data currently exists in structured form, which decisions that data should inform but cannot (because it is not connected or not timely), and what the financial impact of those gaps is. That mapping exercise typically takes two to four weeks and produces a prioritised investment case that is fundable through Made Smarter or internally.

The two highest-value starting points for most UK manufacturers are energy cost visibility (sub-metering connected to production data, enabling cost-per-unit calculations by process step) and production performance reporting (OEE calculated from PLC data rather than operator logs, with automatic downtime categorisation). Both are achievable within three to four months on a modest budget, and both produce measurable financial results within 12 months.

The Made Smarter application process for a combined OT data collection and analytics platform investment is strongest when framed around a specific business outcome - reducing energy cost per unit by 8% over 18 months, or recovering 4 OEE percentage points on the bottleneck line. Funders respond to quantified operational targets, not technology descriptions. The data audit provides the baseline numbers that make those targets credible.

The decision is to stop applying the Continental playbook to a different problem. A large German or French manufacturer builds for scale and a deep data team; the UK mid-market manufacturer is running post-Brexit compliance on a lean IT function and a fragmented Sage/SAP Business One/Dynamics 365 estate. That argues for an architecture that is simpler and more governed, not bigger — Microsoft-centric precisely because it shares identity, security, and licensing with the M365 the business already pays for, which is where integration cost actually lives.

It also starts with a data audit, not a platform purchase. Two to four weeks mapping which operational data exists, which decisions it should inform but cannot, and what the gaps cost produces a prioritised, Made Smarter-fundable investment case — and the two highest-value first builds (energy cost-per-unit by process step, and OEE from PLC data rather than operator logs) are each three to four months to a measurable result. First value framed as a quantified target, not a technology description, is also what wins the funding.

Sequence it as unify, then act: connect the OT and ERP data into a governed Microsoft Fabric foundation, prove one outcome — recovered OEE points or 8% lower energy cost per unit — then expand onto the same estate. The UK manufacturers who treat this as a technology project get a technology result; the ones who treat it as an operational-decision problem, scoped to their actual constraints rather than a Continental template, get the competitive advantage.