Operational excellence: responsibilities, working methods and follow-up

Distinguish excellence from a collection of tools

The Shingo Model connects cultural enablers, continuous improvement, enterprise alignment and results. It asks leaders to examine how operating rules affect people’s behavior. A team may use Lean, Six Sigma, maintenance methods or digital tools within that approach. Leaders still need to check how people use those methods during ordinary work.

Define the few principles that must guide trade-offs: safety, customer value, quality at the source, flow, respect, evidence, and system thinking. Translate each principle into expected leader and team behaviors. If a principle cannot be observed in daily decisions, it remains branding.

• Customer requirements and the people affected
• Expected actions and who may approve a change
• Work, improvement, and management systems
• Process measures linked to customer results
• Review of results with the teams affected

Align strategy with daily operating decisions

Convert strategy into a small set of outcomes, constraints, and guardrails for each value stream. Teams should understand how delivery, quality, cash, resilience, and safety connect to the work they control. Avoid cascades of targets that make every department optimize its own ratio.

Use a defined review rhythm. Frontline teams respond to current abnormalities. Value-stream leaders address cross-process constraints. Executives resolve policy, capacity, capital, and priority conflicts. Escalation should move the decision to the level that can change the system.

Name the model when using ten principles

The Shingo Model names ten guiding principles: respect every individual; lead with humility; seek perfection; embrace scientific thinking; focus on process; assure quality at the source; improve flow and pull; think systemically; create consistency of purpose; and create value for the customer. Attribute this list to Shingo when using it. Other operational-excellence models organize their guidance differently.

Give a principle an observable consequence. For a fictional dispatch team, a goal might be: reduce median accepted-order-to-dispatch time from five working days to four over the next eight weeks. The operations manager owns it. Track late-order share as well as the median, with no increase in first-pass errors, unsafe shortcuts or overtime beyond the agreed limit. Use the same start/end events and comparable order mix.

The baseline and target in that example are invented teaching values. Check an actual baseline before adopting the goal. The operational-excellence metrics guide helps connect a local measure to customer delivery and guardrails.

Design systems that make the desired behavior easier

Inspect what the current management, reward, scheduling, maintenance, quality, and data systems encourage. A leader cannot ask people to stop for quality while rewarding output without accounting for rework. An escalation process cannot work if responders lack time or authority.

Redesign the system around the behavior required at the decision point. Make standards visible, abnormalities safe to report, ownership explicit, and evidence accessible. Technology should shorten the path from signal to responsible action, not introduce another queue of alerts.

1. Clarify purpose

   Define customer and stakeholder value plus non-negotiable guardrails.

2. Observe behavior

   Compare stated principles with real decisions, incentives, and workarounds.

3. Map systems

   Identify policies and tools that reinforce or defeat the desired behavior.

4. Run experiments

   Change one management or work system and measure behavior and results.

5. Scale learning

   Standardize effective routines while adapting them to local context.

Use technology as an evidence layer, not the transformation

Connected PLC, sensor, ERP, MES, quality, maintenance, and manual data can reveal delays, interactions, and emerging risk. AI can detect patterns; simulation can test system consequences. Neither defines value, takes accountability, or repairs a misaligned policy.

Set decision ownership, acceptable latency, model validation, data lineage, cybersecurity, and human review before scaling automation. Start with a bounded operating decision and verify that the technology improves response, learning, or loss, not dashboard use.

Assess maturity through behavior and repeatability

Results can improve temporarily through demand, overtime, inventory, or individual effort. Assess whether the process produces the result repeatedly across products, shifts, leaders, and disturbances. Look for response discipline, problem visibility, verified learning, and cross-functional alignment.

Use assessments to choose the next system experiment, not to award a decorative level. Review gaps with the people doing the work, choose one or two changes likely to affect that gap, and compare observed behavior and operating outcomes after the change.

Do not confuse pillars, principles and an improvement sequence

In the Shingo Model, four dimensions organize the guiding principles: Cultural Enablers, Continuous Improvement, Enterprise Alignment and Results. These dimensions differ from the model’s ten principles. A five-element or five-step list from another author needs its own source; it is not a shorter version of Shingo.

For a five-element local assessment, this guide suggests customer purpose, leadership support, people’s capability, working methods and measurement with follow-up. These are practical review headings, not a universal five-element standard.

A useful five-part starting routine for this guide is to agree the customer outcome, observe current work, test one change safely, compare the result and maintain the proven method. This is our practical grouping of an improvement cycle, not an official “five steps of excellence” certification.

Likewise, “5Ps of operations” is used for different teaching lists. If using people, plant, parts, processes, and planning and control as a local checklist, define them: skills and staffing, facilities and equipment, inputs, working method, and release or scheduling rules. This proposed checklist can expose a missing part or an overloaded approver; it does not establish five universal pillars. Value-stream mapping connects these checks to the actual flow.