Data Transparency in Manufacturing: Why It Usually Fails

What data transparency in manufacturing actually means

Data transparency in a production environment means that the numbers people use to run the plant can be trusted — not just that they exist on a screen. Concretely: raw signals are captured automatically at the source, each data point is linked to order, product, shift and operator context, KPI definitions are the same across lines and plants, and anyone with a legitimate reason can trace a reported value — an OEE number, a scrap figure, a changeover duration — back to the underlying events that produced it.

The distinction that matters: a dashboard is not transparency. A dashboard is a presentation layer. Transparency is the chain underneath — from sensor or manual input through context enrichment and KPI logic to the figure the plant manager reads at 7:30 in the morning. If any link in that chain is missing, inconsistent or quietly edited, the dashboard gets prettier but less honest.

Why data transparency usually isn't there — even when people think it is

Having rolled out MES across roughly forty plants on four continents, the same handful of failure modes recur. They rarely present as "we have no data." They present as "we have plenty of reports and somehow still argue about what actually happened last shift."

Any one of these is enough to undermine a data-driven decision culture. In practice most plants have two or three of them running quietly in the background at the same time.

What actually makes data transparent

Four components, in the order they matter:

• Capture at the source, automatically where possible. Machine cycles, stop events, quantities, process values via OPC UA, PLC reads, or digital I/O gateways. Manual entry is reserved for things that genuinely require a human judgement (root-cause category, rework flag), and it is structured — picklists, not free text.
• Context in one place. Each data point links to order, product, variant, shift, operator and workstation. Without context a cycle count is a number; with context it is a KPI input.
• Shared KPI definitions, centrally governed. Availability, performance, quality, changeover time, scrap — defined once, applied everywhere. This is unglamorous work and usually the single biggest determinant of whether cross-plant comparisons are meaningful.
• An auditable chain from KPI to raw event. Any headline figure can be drilled down to the underlying sequence of events. Changes to classifications are logged, not silently overwritten. This is what separates "transparency" from "a well-produced report."

Observation from MES rollouts in Europe, North America, Mexico, China and Tunisia: the headline problem is almost never missing data. The headline problem is that the numbers that exist are optimised for reporting, not for learning. On several sites I worked on before SYMESTIC, the OEE posted on the wall looked healthy while the actual pattern of micro-stops, minor quality defects and unrecorded changeovers told a very different story. None of that was malicious — it was the natural result of people being measured on a single figure and given the means to influence how it was calculated. That experience is what prompted me to write "OEE: Eine Zahl, viele Lügen" in 2025. The practical conclusion I keep coming back to is simple: a lower but honest KPI is more valuable than a higher one you had to polish to get. A transparent data layer is the thing that makes the honest number possible.

How transparent data changes daily operations

The effects tend to show up quietly rather than dramatically. Shift handovers get shorter because there is no argument about what happened. Problem-solving sessions spend less time establishing facts and more time analysing causes. Improvement initiatives — Lean, Six Sigma, Kaizen — become easier to evaluate because before-and-after comparisons rest on the same measurement logic. Cross-plant benchmarking stops being a political exercise and becomes an operational one.

The effect on KPI numbers themselves is worth calling out honestly: when automated capture is switched on for the first time, measured OEE often drops several points compared to the previous manual figure. That is not the system performing worse. That is the first time the plant is looking at an unedited version of its own reality. The second round of measurable improvement usually starts from there.

How the SYMESTIC platform supports the chain

SYMESTIC is a cloud-native MES built around the four components listed above. Machine connectivity runs over OPC UA for modern controls and over IoT gateways with digital I/O for brownfield assets — typically without any PLC intervention. Order and master-data context comes from bidirectional ERP integrations (SAP R/3 via ABAP IDoc, Microsoft Dynamics/Navision, Infor/InforCOM, proAlpha). KPI definitions — OEE and its availability / performance / quality components, First Pass Yield, scrap, changeover time, throughput — are configured centrally and applied consistently across lines and sites. Classifications and reason-code changes are logged. Drill-down from an aggregated KPI down to the specific stop event on a specific machine in a specific shift runs on the same underlying data, not on a separately maintained export. The customer-facing modules most relevant to the transparency layer are production metrics, process data, alarms and production control.

FAQ

Is data transparency the same as having a dashboard?
No. A dashboard is the output. Transparency is the chain underneath — source capture, context, consistent KPI logic, traceability back to the raw events. A dashboard on top of a broken chain produces confident-looking but misleading figures.

Why do our OEE numbers drop when we introduce automated data capture?
In most cases because the earlier figure was based on incomplete capture, generous classification of stops as "planned", or gaps in micro-stop detection. The drop is usually the first honest reading the plant gets. Treat it as a new baseline, not a regression.

How is this different from manufacturing visibility?
Visibility is about seeing current state in real time across lines, shifts and sites. Transparency is about being able to trust what you are seeing — that the signals are captured correctly, classified consistently and traceable back to source. Visibility without transparency is a well-lit view of an unreliable reality.

How much of this can be done without an MES?
For a single line with one or two signals, a lot can be done with a historian, a BI tool and disciplined master-data handling. Beyond that, the combination of machine connectivity, ERP context, shared KPI logic and auditable drill-down is exactly what an MES is designed for. Plants that try to recreate it from loose components typically end up with something MES-shaped but harder to maintain.

What is the biggest risk in a transparency programme?
Starting with visualisation rather than data hygiene. If the underlying capture, classification and definitions are not in order, a better dashboard makes the problem less visible, not smaller. Investing in the chain first — even when it is the slower part — is what separates programmes that stick from ones that get quietly replaced two years later.

Does transparency mean exposing every number to everyone?
No. Role-based access is part of the design. Operators, shift leaders, production management and group controlling need different cuts of the same underlying data, not the same screen. The point is that all of those cuts come from one reconciled source, not from four parallel reports that happen to agree most of the time.

How does SYMESTIC handle cross-plant KPI consistency?
KPI definitions are configured centrally in the platform rather than set per site. Stop categorisations, availability rules, performance benchmarks and scrap logic are shared across plants, which makes group-level dashboards comparable rather than negotiated. Data retention, ERP integration and cloud architecture are covered in Cloud MES vs. on-premise, vendor and cost context in MES software compared. See also about SYMESTIC and pricing.

Related: MES: Definition, functions & benefits · OEE: Definition, calculation & practice · OEE software · MES software compared · Cloud MES vs. on-premise · Production metrics module · Process data module · Alarms module · Automotive · Metal processing · For COOs & plant managers · For operational excellence.

About the author

Christian Fieg

Head of Sales at SYMESTIC. 25+ years in manufacturing operations — maintenance engineer and Six Sigma Black Belt at Johnson Controls, PLC engineer for JIT / Just-in-Sequence centres of excellence, expatriate in Changchun (China), later Team Leader Business Analyst with global MES and traceability responsibility across 900+ machines, 750+ users and plants in China, Mexico, USA, Tunisia, North Macedonia, France and Russia. Manager Center of Excellence at Visteon, Sales Manager MES (DACH) at iTAC, Senior Sales Manager at Dürr. Author of "OEE: Eine Zahl, viele Lügen" (2025). Expertise: MES, OEE, shopfloor digitalisation, Six Sigma (Black Belt), traceability, Smart Factory, global MES rollouts, cloud MES. · LinkedIn