Why shop-floor data still falls short
4 March 2026 
UK manufacturers face growing pressure to turn shop-floor data into measurable value, yet legacy systems, brownfield sites and fragmented architectures continue to hold progress back. Aaron Blutstein spoke with Tony Coghlan of Turck Banner and Jonathan Parr and Daniel Winter
of Accenture UK about what is beginning to change, and what remains stubbornly unresolved.
Digital transformation in UK manufacturing is no longer an abstract idea, yet it remains unevenly embedded in day-to-day operations. Many organisations can point to individual initiatives that promised better visibility or efficiency, but far fewer can demonstrate sustained improvement across sites or production lines. The difficulty lies less in access to technology than in translating fragmented data, ageing assets and limited investment headroom into decisions that consistently improve performance.
Much of this challenge stems from the way manufacturing systems have evolved. Layered architectures separating sensors, control systems, supervisory platforms and enterprise applications remain technically sound, but they have also reinforced organisational and technical silos. Operational technology and information technology have historically developed in parallel, and although convergence has been discussed for years,
delivering it in practice has often proved slow, disruptive and costly.
There is growing acceptance that progress does not always require wholesale architectural change. Daniel Winter, Manufacturing & Supply Chain System Manager at Accenture UK, points to a shift towards more pragmatic data strategies, describing the value of “getting raw data from the shop floor straight to the cloud … without having to go through all that complexity”. Rather than removing control systems, this approach selectively bypasses parts of the traditional automation stack when the objective is to surface insight for higher-level systems, not to interfere with machine control.
That distinction is particularly relevant in brownfield environments, which account for much of UK manufacturing capacity. Many plants rely on legacy equipment with limited connectivity, restricted access to source code or controllers never designed to feed modern analytics platforms. Winter characterises the opportunity as a retrofit challenge, noting the role of “overlaying additional hardware” to connect assets that are “obsolete or don’t even have the capability” to expose usable data. The appeal lies in improving visibility while leaving established production systems untouched.
A similar emphasis on pragmatism underpins the position taken by Tony Coghlan, Managing Director at Turck Banner. He is clear that digitalisation should not default to replacing existing systems or rebuilding control architectures. “We are just not that hardware supplier,” he says, framing Turck Banner’s role around enabling outcomes rather than simply providing components. Central to that position is the ability to deliver “actionable data” while avoiding unnecessary complexity, particularly at a time when, as he observes, “people are really struggling to get capex approved”.
Full lifecycle
In practical terms, Coghlan describes a scope that extends well beyond device supply. Turck Banner positions its involvement across the full lifecycle of shop-floor data enablement, beginning with the selection of appropriate sensing and connectivity hardware from its industrial IoT portfolio, often working with equipment already installed. That is followed by early definition of requirements and system architecture through a Functional Design
Specification, intended to establish clarity before technology is deployed.
Deployment itself is treated as a critical phase rather than an afterthought, particularly in brownfield environments where installation, connectivity and software configuration must coexist with live production. The integration focus is deliberately aligned with standard industrial and enterprise interfaces rather than bespoke automation solutions. Coghlan points to the use of widely adopted protocols such as REST APIs, MQTT and OPC UA, alongside other established industrial communication standards, as a way of ensuring that operational data can be consumed by IT systems without introducing unnecessary complexity.
The objective of this end-to-end approach is to make shop-floor data usable beyond the machine or maintenance silo. Information is intended to flow into enterprise environments such as MES, CMMS, EAM and OEE platforms, as well as enterprise dashboards, where it can support production management, maintenance planning, asset performance analysis and utilisation decisions. The underlying argument is about reducing supplier
fragmentation: limiting the need for manufacturers to assemble and manage multiple vendors to move data from sensors to enterprise consumption.
This focus on integration highlights a common weakness in many Industry 4.0 programmes. Proof-of-concept projects often succeed in demonstrating that data can be captured but fail to show how it will be embedded in existing operational processes. Coghlan’s criticism is direct: data that cannot be acted upon adds little value. The opportunity lies in ensuring that information collected on the shop floor can trigger work orders within CMMS or EAM systems, inform production decisions through MES, or contribute meaningfully to OEE analysis, rather than remaining confined to isolated dashboards.
Artificial intelligence has intensified attention on these fundamentals. While AI is widely promoted as transformational, its effectiveness in manufacturing depends entirely on the quality and context of the data it consumes. Jonathan Parr, who leads the Manufacturing Operations practice at Accenture UK, is explicit on this point, stating that manufacturers “need to have the availability of the data … to be able to plug AI into and be able to utilise that”. Without reliable, contextualised shop-floor data feeding enterprise systems, AI risks producing outputs that are mistrusted or
operationally unsafe.
Concerns around cyber security inevitably accompany greater connectivity, particularly as data moves closer to enterprise platforms. The discussion, however, is shifting away from fear and towards design discipline. Winter emphasises that principles such as encryption, network segregation and careful protocol selection must be treated as integral to system design rather than afterthoughts. In retrofit scenarios, parallel data-acquisition architectures can, when properly designed, limit exposure of legacy control systems while still enabling enterprise-level data flows.
Skills shortages
Skills shortages continue to shape what is practical. Many manufacturers are losing experienced engineers faster than they can replace them, taking decades of tacit knowledge with them. There is cautious optimism that better integration between shop-floor data and systems such as EAM and
CMMS can help capture operational knowledge, standardise responses and support less experienced staff. However, this depends on disciplined data definition and deployment, not on technology alone.
Levels of digital maturity across UK manufacturing remain highly variable. Some operations are heavily automated and already feeding data into MES and OEE frameworks, while others rely on manual processes with limited visibility. That disparity creates different development paths. Highly automated sites face the complexity of retrofitting legacy systems at scale, whereas less automated facilities may be able to progress more quickly
by deploying modern sensing and connectivity technologies that are now cheaper and easier to integrate with enterprise systems.
Taken together, these perspectives point towards a more grounded understanding of digital transformation. Rather than pursuing idealised architectures, manufacturers are increasingly focused on value, scalability and speed to insight. Winter’s emphasis on simplification and Parr’s focus on data readiness, alongside Coghlan’s insistence on actionable data and end-to-end responsibility, reflect a broader shift towards pragmatism.
Lasting progress will depend not on digitising poor processes or accumulating data for its own sake, but on ensuring that shop-floor information can move reliably into MES, CMMS, EAM and OEE environments where it supports decisions that materially affect performance.
For further information please visit:
https://www.turckbanner.co.uk
https://www.accenture.com