Process Standardization: Paper vs. Executed

What is process standardization in manufacturing?

Process standardization in manufacturing is the systematic definition and enforcement of a single, consistent way of executing a production process — across shifts, operators, lines, plants and countries. The textbook definition focuses on documents: work instructions, SOPs, quality specs, safety guidelines. The reality on the shop floor is different. In 25 years of rolling out manufacturing systems across seven countries and four continents, I have walked into very few plants that lacked written standards — and very few that followed them consistently. The gap between the standard on paper and the standard that actually runs is where most of the value of standardization either is captured or is lost.

At Johnson Controls I had global responsibility for MES and traceability across 900+ machines, 750+ users and 30+ production processes — soldering, assembly, injection moulding — distributed across China, Mexico, the US, Tunisia, Macedonia, France and Russia. Standardization was not an academic topic; it was a condition of shipping product. This article is about the distinction that matters operationally: the standard as written versus the standard as executed, and what it takes to make them the same thing.

The four levels of standardization — and where most plants stop

Standardization is not a single binary state. It exists in four progressive levels, and the level a plant actually reaches — not the level it claims — determines whether standardization delivers its promised benefits.

Most plants I have audited reach Level 1 or 2 and declare victory. The ISO binder is full, the training records are current, and the assumption is that the standard is therefore running. The MES data says otherwise. When a standard exists only as a document, compliance is assumed and never verified. When it is trained but not enforced at the point of execution, operators drift — not maliciously, but because every operator develops small preferences, every shift has its local interpretation, and over months those variations become the real standard while the written one becomes fiction. Only plants that reach Level 3 capture the operational benefits; only plants that reach Level 4 can improve the standard systematically because they can finally see where it is being followed and where it is not.

The paper-to-floor gap — where standards go to die

The most expensive mistake I watched managers make repeatedly, across every country I worked in, was confusing "we have written it down" with "it is happening." The gap between the paper standard and the executed standard is almost universal, and it has predictable mechanics.

Work instructions that live in a folder nobody opens. PDF SOPs in a document management system. Operators would have to leave the line, find a computer, log in, search for the current version, read it — and then return to a cycle that already started two minutes ago. Nobody does this. They rely on memory, on what the previous shift did, on what seemed to work last time.

Standards that are technically impossible to follow. The ideal cycle time says 42 seconds; the line has not run at 42 seconds in two years. The SOP says "verify temperature every 30 minutes"; the operator has eight machines and cannot physically do that. When the written standard is unreachable, operators invent a local one. The local one becomes reality; the written one becomes a ritual.

Silent drift between shifts. Night shift's "small improvement" becomes morning shift's bug. Day shift's workaround for a jammed sensor becomes permanent. Over months, three shifts produce three different processes — all of them claim to be running the same standard, and from the outside, nobody can tell which is which without event-level data.

Cross-plant variance that nobody measures. At Johnson Controls, the same process — the same product, the same equipment, the same SOP — ran differently in Changchun, Juárez and Trenčín. Not slightly differently. Measurably, consistently, reproducibly differently. Each plant believed it was following the global standard. Without cross-plant KPI comparison via the MES, the variance would have remained invisible for years.

The field observation from global MES rollouts: in the first month after a process is digitally instrumented, the measured variance across shifts, operators and plants running "the same standard" is typically 15-30 % on cycle time and 20-50 % on stop-reason distribution. This is not a result of bad standards or bad operators. It is the predictable consequence of standards that exist in documents instead of at the point of execution. The plants that act on this variance — that treat the data as information rather than as a threat — close most of the gap within six months. The plants that dismiss the data as "measurement noise" are still running the three different processes three years later.

How digital execution closes the gap

The mechanism that closes the paper-to-floor gap is not more training, more audits or stronger management pressure. It is moving the standard from the document system to the point of execution — to the operator's screen at the machine, at the moment the work happens. That is the operational insight behind every MES-based standardization initiative I have led.

Digital work instructions at the station. Not in a folder, not in a separate system — on the HMI next to the machine, tied to the active production order, updated when the SOP is updated, versioned and auditable. The operator cannot work from an outdated revision because the outdated revision is not on the screen.

Parameter sets pushed from the system. Machine settings — feeds, speeds, temperatures, pressures — transferred from the MES to the control system at order changeover, not set manually by the operator. This eliminates the single biggest source of cross-shift variance in most processes: the setup itself.

Enforced checklists with electronic signature. Setup verification, first-article inspection, safety checks — completed on the HMI, time-stamped, tied to the operator and the order. If the checklist is not complete, the line cannot release. This is unpopular with operators for the first week and essential thereafter.

Deviation detection at cycle level. The system knows the standard cycle time, the standard stop-reason distribution, the standard yield. Every deviation is flagged — not to punish the operator, but to drive root-cause analysis and SOP improvement. This is Level 4 standardization, and almost no plant reaches it without an MES.

A real case: Klocke

The Klocke Group is an international contract manufacturer in pharmaceuticals, cosmetics and nutritional supplements — a GMP-regulated environment where process standardization is not optional and where deviations have consequences that go well beyond operational KPIs. The engagement with SYMESTIC illustrates how fast a well-structured standardization initiative can propagate when the enforcement mechanism is right.

The start was deliberately narrow: one line in the packaging area at the Weingarten site, instrumented with digital I/O gateways, no LAN infrastructure needed, no PLC intervention. Count and stop events were captured automatically through DI-Gateway and mapped to the active production order. The system was designed from day one for a regulated environment, following GMP principles — electronic signatures, audit trail, version-controlled SOPs. A unidirectional interface to the Navision ERP via file exchange brought the order context and master data into SYMESTIC and mapped every machine cycle and every stop back to the originating manufacturing order.

The acceleration is the part that matters for the standardization story: within three weeks, the initial line had scaled to all lines at the Weingarten site. Not three months, not six months — three weeks. That scaling speed is only possible when the standardization pattern is right: digital I/O gateways installable in hours per line, standardized SOPs enforced through the same HMI across every line, the same stop-reason taxonomy everywhere, the same order-mapping logic. The modular SYMESTIC catalogue then allowed Klocke to continue extending the rollout independently — the self-service scaling pattern that distinguishes sustainable standardization programmes from one-time vendor projects.

The 12 % output gain and 8 % availability gain were not the goal of the project — GMP compliance was. They were the downstream consequence of what happens when the standard is enforced at the point of execution rather than written in a document: the variance that previously hid between shifts and between lines becomes visible, the recurring problems become addressable, and the process becomes measurably more stable. That is the practical case for standardization done properly.

FAQ

What is process standardization in manufacturing?
It is the systematic definition and enforcement of a single consistent way of executing a production process — across shifts, operators, lines and plants. The written standard is only the starting point. Real standardization requires that the same standard is actually being followed at the point of execution, verified by measurement, and continuously improved based on the deviations that measurement reveals.

Why do written SOPs often fail to deliver the expected benefits?
Because a written standard is not an executed standard. SOPs stored in document systems are rarely consulted at the moment of execution; operators rely on memory, precedent and local interpretation. Over time, shifts drift, plants diverge, and the written standard becomes a compliance artefact rather than an operating reality. The gap between paper and floor is almost universal and is rarely visible without event-level machine data.

What are the four levels of standardization?
Documented (SOPs exist on paper), Trained (operators have been trained and records exist), Executed (the standard is followed consistently in real conditions across all shifts), and Enforced & Measured (deviations are detected automatically and closed systematically). Most plants reach Level 1 or 2 and believe they are standardized. Only plants that reach Levels 3 and 4 capture the operational benefits — the quality, throughput and consistency that standardization promises.

How large is the variance between shifts and plants running "the same" standard?
In the first month after a process is digitally instrumented, cycle-time variance typically ranges 15-30 % across shifts and operators, and stop-reason distribution varies 20-50 % across plants running the same written standard. This is not the result of bad standards or bad operators — it is the predictable consequence of standards that exist only as documents. The variance is almost always invisible before automated measurement and highly actionable after.

How does an MES help with process standardization?
By moving the standard from the document system to the point of execution. Digital work instructions on the operator's HMI, tied to the active order and the current SOP revision. Parameter sets pushed from the system to the controls at changeover, eliminating manual setup variance. Enforced checklists with electronic signature, so the line cannot release without completion. Cycle-level deviation detection that flags variance for root-cause analysis rather than hiding it in averages. None of these require discipline from the operator — they make the standard the path of least resistance.

Is standardization compatible with continuous improvement?
Yes — and in fact, real continuous improvement requires standardization. Without a stable baseline, every "improvement" is just noise; without consistent execution, you cannot tell whether a change made the process better or simply replaced one variance with another. Toyota's original insight — standardization is the precondition for kaizen, not the opposite of it — remains true. The standard is what you improve against, and every improvement that sticks becomes the new standard.

How do I standardize across multiple plants in different countries?
Three non-negotiable principles from global rollouts. First, a single data model — same stop-reason taxonomy, same KPI definitions, same master-data structure across every plant. Second, a single enforcement mechanism — the same MES, the same HMI, the same workflow on every shop floor. Third, local adaptation only at the operator-interface layer (language, units, local labels) and never at the data-model layer. Plants that allow local variants in the data model cannot be compared, which means the variance between them cannot be managed, which means they are not standardized in any meaningful sense — they just share a logo.

How does SYMESTIC support process standardization?
Digital work instructions at the machine HMI, tied to the active production order, versioned and auditable. Parameter sets synchronised from the MES to the control systems via OPC UA at order changeover, eliminating manual setup variance. Enforced checklists and electronic signatures for setup verification, first-article inspection and shift handover. Standardised stop-reason taxonomy across every plant in a multi-site rollout, with no local variants at the data-model layer. Cross-plant KPI comparison with cycle-level drill-down so variance between "identically standardised" plants becomes visible and actionable. Bidirectional ERP integration so every standardised process is tied to order, product and customer context automatically. 15,000+ machines connected across 18 countries on this architecture. See SYMESTIC Production Metrics.

Related: Manufacturing Processes · Performance Measurement · OEE · Six Sigma · Kaizen · Lean Production · MES · SYMESTIC Production Metrics

About the author

Christian Fieg

Head of Sales at SYMESTIC. 25+ years in manufacturing industries — Six Sigma Black Belt and JIT engineer at Johnson Controls, expatriate plant lead in Changchun, China, global MES and traceability lead at JC Automotive Electronics (900+ machines, 750+ users, 30+ processes across seven countries), MES practice lead at Visteon, Sales Manager MES (DACH) at iTAC, Senior Sales Manager at Dürr. Author of the 2025 book OEE: Eine Zahl, viele Lügen. · LinkedIn