Standard WorkJapanese: 標準作業 — hyōjun sagyō

Standard work is the documented, currently one-best-known method for performing an operator task at takt — built by the team leader together with the operators who do the job, posted visually at the station, used as the training + audit + jidoka baseline, and replaced every time kaizen produces a better method. The Toyota convention is three artefacts side-by-side at every station: the Process Capacity Sheet (takt time, machine cycle times, manual cycle times, and resulting capacity per shift), the Standardised Work Combination Table (the sequence of operator + machine actions plotted against time within one takt cycle), and the Standardised Work Chart (the layout view — operator walk paths, standard WIP locations, quality-check points, safety hazards).

The crucial Toyota framing — easy to miss — is that the standard is provisional. It is the current best-known method, not the eternal correct method. Posted standards are expected to be replaced every few weeks as kaizen finds improvements. A standard that has not been updated in six months is either perfect (rare) or being ignored by the people doing the work (common). Either way it is a problem.

Toyota teaches standard work as three inseparable elements. Any one missing and the discipline collapses — a method without a takt-time anchor is just an SOP, a takt-time number without a method is an unaccountable target, and either without the standard-WIP rule produces a station that hides its inventory abuse.

Standard work is written collaboratively by the team leader and the operators who do the job. It is not written by an industrial engineer in an office. This rule has empirical history: standards written by IEs and handed down to the floor are almost universally ignored within weeks because they describe a method nobody on the floor recognises. Standards built with the operators are followed because the operators built them — and because the team leader knows the moment the standard drifts.

• Team leader leads the documentation. The team leader (1 per 4–8 operators, hands-on, on the floor) facilitates the time-and-motion study, drafts the three artefacts, posts them at the station and trains the operators against them. The team leader is the standard-work owner.
• Operators contribute their tacit knowledge. The IE-led 'one right method' fantasy ignores that experienced operators have already discovered a better way. Standard work captures their method, not a theoretical method.
• Industrial engineering supports, does not author. IEs provide the time-study tooling, the layout drawings, and challenge sequences that look inefficient — but they do not own the final standard.
• Quality + safety review for non-negotiables. QA verifies that quality-check points are included; EHS verifies that safety hazards + PPE requirements are explicit. Neither rewrites the method.
• Operators sign the standard before it goes live. Acknowledgement that this is the agreed method — not just leadership's preferred method.

In a regulated plant standard work coexists with several other documented-procedure artefacts. They are not duplicates — each serves a different audience. Plants that conflate them either drown the operator in paperwork (treating the SOP as standard work) or evade the regulation (treating standard work as the SOP).

The maturity test: walk a station and ask the operator how they perform the task. A mature plant produces an answer that matches all five artefacts above because they were authored coherently. A typical plant produces an answer that matches none of them, because the artefacts were authored in different offices for different audiences and never reconciled.

Toyota's operator training is the Training Within Industry (TWI) Job Instruction (JI) method — a four-step pedagogy developed by the US War Manpower Commission during WWII to qualify millions of new factory workers, adopted by Toyota in the 1950s and never abandoned. JI works against the standard-work document. Without the standard, JI has no script; with it, training is reproducible across trainers, shifts and years.

1. Prepare the worker. Put them at ease, find out what they already know, get them interested in learning the job, place them in the correct position.
2. Present the operation. Tell, show and illustrate one important step at a time — important steps + key points + reasons (the JI breakdown sheet). Stress the key points. Instruct clearly, completely and patiently. Cover no more than they can master.
3. Try out performance. Have the worker do the job, correcting errors. Have them explain each important step and key point as they perform it. Continue until you know they know.
4. Follow up. Put them on their own. Designate to whom they go for help. Check frequently. Encourage questions. Taper off coaching and close follow-up.

The training record (under 21 CFR 211.25 / 820.25 / 117.4 / ISO 13485 §6.2) is the regulated trail proving the worker was qualified — but the substance of the training is JI against standard work. A plant with a complete training-record system but no standard-work baseline has trained nobody to a reproducible method.

David Mann's Creating a Lean Culture (3rd ed., Productivity Press, 2014) extended standard work to leaders. The original Toyota practice was implicit; Mann made it explicit. Leader standard work is the documented daily / weekly / monthly cadence of activities a supervisor, area manager, plant manager and executive perform to sustain the operational system — gemba walks, andon-board reviews, kaizen-event approvals, monthly hoshin reviews, year-end retrospectives.

• Team leader — minute-by-minute presence at the station, andon response within 30s, daily start-of-shift hand-off, hourly process-confirmation walks against standard work, daily kaizen-board update.
• Supervisor — daily area walk (1–2h), weekly safety + 5S audit, weekly tier-meeting facilitation, monthly standard-work refresh sign-off, monthly area-improvement-plan review.
• Area / value-stream manager — weekly gemba walks (4–8h total), monthly hoshin-tactic review, monthly capability-development review with team leaders, quarterly cross-area kaizen review.
• Plant manager — daily 15-min andon-board review at the start of shift, weekly multi-area gemba walk, monthly hoshin review at site level, monthly customer-complaint trend review, quarterly executive review.
• Executive — monthly gemba walk at one site (not a tour — a structured walk with the local team), quarterly hoshin steering, annual True-North refresh.

Leader standard work makes the operational system sustainable. Without it, leadership presence at the gemba depends on individual personality, and the system decays whenever a strong leader rotates out. With it, presence is institutionalised: the new plant manager walks the andon board every morning because that is what plant managers do here — not because they chose to.

• 21 CFR 211.100 — Written procedures; deviations. Standard work is the operator-level expression of the 'written procedures' the SOP describes. Deviations are detected by jidoka against the standard; without standard work, deviation surfacing depends on individual judgement.
• 21 CFR 820.70 — Production and process controls. 'Process control procedures shall describe the methods to be used' — standard work is the gemba-level realisation of those methods at takt.
• 21 CFR 211.25 / 820.25 / 117.4 — Personnel qualifications + training. Standard work is the training substrate; the training record is the qualification trail. Auditors increasingly examine whether the trained method matches the documented method matches the observed method.
• 21 CFR 211.130 / 211.188 — Batch production + control records. The batch record's accurate-reproduction-of-the-MMR requirement (211.188) is impossible without standard work at each operator step — divergence at the station defeats the reproduction at the record level.
• ICH Q10 §3.2.2 — Process capability + control. Process capability is a property of the method-as-actually-performed, not the method-as-documented. Standard work is the only mechanism that makes documented + actual converge.
• ICH Q9(R1) — Quality risk management. Risk assessments assume a defined method; standard work defines it. Risk assessments performed against undocumented practice are guesswork.
• EU GMP Chapter 4 — Documentation. The SOP, work instruction and batch-record template define what must be done. Standard work defines how it is done at the station — and is the only way the documentation system stays coherent with practice.
• ISO 13485 §7.5.1 — Control of production. Production must be controlled through documented procedures, monitoring and product release. Standard work is the operator-level control; the SOP is the systemic control; both are required.
• FSMA 21 CFR 117.35 — Sanitary operations + 117.135 process controls. Operator-level method specification under the food preventive-controls rule.

• Coverage — % of operator tasks with a current standard work document. World-class >90%; typical Western plant <30%.
• Currency — average age of posted standards. World-class <60 days; aging standards = ignored standards.
• Adherence — % of process-confirmation walks where the observed method matches the posted standard. World-class >95%; below 80% means the standard is wrong, not the operators.
• Sequence-time variance — coefficient of variation of cycle time across operators on the same task. Falling CV = standard work is gripping; rising CV = ad-hoc methods returning.
• Training-to-qualification ratio — average time from new operator on station to qualified independent. With JI + standard work, halves vs ad-hoc training.
• Kaizen-to-standard ratio — % of implemented kaizens that produce a standard-work update. Below 70% means kaizens are not persisting; the system is leaking improvement.
• Leader standard work completion — % of leader cadence items completed on schedule per tier. Below 70% means leadership presence at the gemba is depending on personality, not system.

1. IE-written standards handed down. Operators ignore them within weeks; the plant 'has standard work' on paper and ad-hoc methods at the gemba.
2. Confusing SOPs with standard work. The SOP describes the compliance method; standard work describes the takt-anchored gemba method. Plants that conflate them either over-burden operators or under-document the actual practice.
3. Standards that never change. A standard untouched for 6+ months is either perfect or ignored. The latter is overwhelmingly more common.
4. No takt anchor. 'Standard work' written as a generic procedure without a takt-time number; operators run at whatever pace, capacity is invisible, abnormality is undefined.
5. Skipping standard WIP. Stations accumulate WIP whenever upstream runs faster than downstream; the accumulation is invisible because nobody specified the cap. Single-piece flow becomes impossible.
6. No leader standard work. Operator standard work is built but leadership presence at the gemba depends on personality. The first time a strong leader rotates out, the system decays.
7. Treating standard work as separate from training + jidoka + kaizen. The four disciplines are one system — standard is the baseline, JI is the training method, jidoka is the abnormality detection, kaizen is the improvement engine. Run any one without the others and the system fragments.

V5 makes standard work a first-class artefact in document control — not a separate spreadsheet world, not a copy of the SOP, but the gemba-level companion that drives the kiosk EWI, the training record, and the jidoka baseline.

• Three-artefact authoring workspace. Process Capacity Sheet + Standardised Work Combination Table + Standardised Work Chart in a single editor; takt-time field auto-derived from demand + available-time inputs; element-time inputs roll up to total cycle time + capacity per shift automatically.
• Team-leader-led collaborative authoring. Standard work is owned by the named team leader; operators are listed as contributors and sign the standard before it goes live (Part 11). IE / QA / EHS reviewers route in parallel.
• Living standard, versioned. Every revision carries an effective date, a delta-summary against the prior revision, and the named kaizen that drove the change. Aging standards (60+ days without review) auto-flag.
• Kiosk EWI synchronisation. The standard-work sequence drives the EWI step-list at the kiosk automatically; standard-WIP rules drive the in-line cap checks; quality-check points drive the in-line capture prompts. Change the standard work, the EWI updates the next cycle.
• Training-record integration. Every standard-work revision routes a training assignment to all qualified operators on that task; kiosk hard-block prevents starting work until the new revision is acknowledged. The training record links the operator to the revision they trained on, not just to the SOP number.
• Job-Instruction breakdown sheet generator. The standard sequence auto-drafts a JI breakdown (important steps + key points + reasons); trainer refines, attaches photos / videos, marks safety + quality key points; new operators train against the JI breakdown and qualify against the standard.
• Jidoka baseline. The kiosk knows the standard sequence + element time; deviations from sequence trigger andon by default; cycle-time variance triggers monitoring; abnormality is defined automatically.
• Leader standard work module. Configurable per-tier cadence (team leader / supervisor / area manager / plant manager / executive) with kiosk + mobile completion tracking, dashboard for completion rates by leader, and auto-rolled into management-review records.
• Process-confirmation walks. Supervisor / area manager walks with mobile kiosk; tap a station, see the posted standard vs the observed method, record adherence + findings + photos; non-adherence findings auto-route to standard-work review (is the standard wrong?) or training reinforcement (is the operator wrong?).
• Regulated overlay (211.100 / 820.70 / 211.25 / 820.25 / 211.130 / 211.188 / ICH Q10 §3.2.2 / EU GMP Ch.4 / ISO 13485 §7.5.1 / FSMA 117.35 / 117.135). Standard work routes through change control coherently with the SOP and MMR / DMR — operators do not see two parallel documentation systems.
• Part 11 + Annex 11 audit trail. Every authoring change, review approval, kiosk acknowledgement, training completion and process-confirmation finding is timestamped, attributed and retained.
• Mobile-safe (iPhone ≤390px). Authoring, review, process-confirmation walks and leader standard work all work on the floor — no desktop dependency.