SMED: externalize safely, verify first good output

SMED: externalize safely, verify first good output

Observe the complete changeover boundary, including the point at which acceptable output is restored. Separate activities that require stopped equipment from preparation that can safely occur beforehand or elsewhere. Improve preparation first, then ask whether any internal activity can be redesigned as external work. Simplify the remaining steps without bypassing isolation, technical settings or quality checks. Keep the first-good-output verification in the new method. Record measured timing rather than treating reduced searching as the entire changeover gain. SMED names an ambition for single-digit minutes; it does not promise every process can change in one minute.

Follow the method

1. Before
2. External preparation
3. Internal work
4. First good output

Move preparation without moving the acceptance boundary

Fictional packaging changeover: the team currently stops, fetches an approved kit, performs the controlled internal change and verifies first good output. Leader Nina proposes preparing the approved kit while the preceding run is still operating. No step durations or achieved reduction are supplied.

Follow the method

1. Approved-kit preparation
2. Internal change
3. First-good verification
4. Elapsed improvement
5. Missing kit item

Fetch the kit before the stop, if it is truly ready

Separate work that genuinely requires a stopped process from preparation that can occur externally, and preserve first-good-output verification when proposing a changeover trial.

Fictional packaging changeover: the team currently stops, fetches an approved kit, performs the controlled internal change and verifies first good output. Leader Nina proposes preparing the approved kit while the preceding run is still operating. No step durations or achieved reduction are supplied.

Role: Changeover team with equipment, quality and operating representatives.

Normal condition

The next configuration is correct, safe and producing accepted output before the changeover is considered complete.

The gap

The machine is stopped while the approved kit is located, but nobody has confirmed which preparation can safely occur during production.

• Internal work requires the process stopped; external preparation can occur while it operates only under approved conditions.
• No target time justifies bypassing isolation, safeguards or acceptance checks.

1. Define the changeover boundary

   Nina states the start and finish events before observing: the agreed last good output of the prior configuration and first accepted output of the next, with the exact local time convention recorded.

   Why: A reported saving can be manufactured by stopping the clock before quality verification or moving the start.

   Evidence: Boundary includes transition and accepted restart; actual timestamps are still required.

2. Classify by operating condition

   List kit preparation as a candidate for external work and retain the controlled configuration change as internal. Ask the equipment and quality owners to verify what may occur while the preceding run operates.

   Why: The distinction is about whether the process must stop, not whether a task seems easy or quick.

   Evidence: Candidate external preparation is explicitly provisional; internal access controls remain intact.

3. Make readiness visible before the stop

   Propose a checked kit at the defined preparation point: correct identity, complete approved contents and known next configuration. Record who verifies readiness and what happens if anything is missing.

   Why: Fetching an unchecked kit earlier could move the delay or create a wrong-configuration risk rather than improve the changeover.

   Evidence: A readiness record replaces the assumption that “kit nearby” means “kit ready.”

4. Preserve the controlled restart

   Retain internal work, restoration checks and first-good-output acceptance in the proposed sequence. If a check fails, use the existing hold/response process.

   Why: Changing preparation timing does not remove the requirements for a correct process and suitable output.

   Evidence: Proposed sequence: prepare approved kit → stop → controlled internal change → verify first good output.

5. Measure before claiming a gain

   Observe representative before and trial changeovers using the same boundary, noting mix, staffing, missing items, adjustments and quality results.

   Why: SMED is an improvement method; its name does not establish a measured result for this unmeasured example.

   Evidence: No duration or percentage saving is reported until comparable records exist.

The kit is incomplete at the readiness check

A required approved component is absent before the planned stop.

Escalate the readiness failure to the changeover owner and follow the approved production plan; do not substitute an unapproved component to protect a time target.

The earlier check has made the problem visible. It has not created authority to change the configuration requirements.

Missing item, owner, decision and revised plan are recorded.

Run a practice session

Materials

• Before/proposed sequence cards and a readiness checklist
• A missing-kit exception card

1. Define completion · 4 minutes

   Does the clock stop before or after acceptance?

2. Classify the original steps · 8 minutes

   What evidence makes preparation genuinely external?

3. Repair the printing proposal · 8 minutes

   Which shortcut changes quality authority?

4. Plan measurement · 5 minutes

   How will you compare unlike changeovers honestly?

Debrief

• Challenge unsupported savings figures.
• Use the incomplete kit to discuss early visibility rather than blame.

Write the two sequences and annotate approval, readiness and acceptance conditions before reading the solution.

Transfer into the work

Owner: Changeover process owner with quality and equipment representatives

Record: Element classification, readiness check and comparable timing/quality record

Review: At the controlled trial and before approving revised standard work

Evidence: Correct first output, unchanged necessary controls and measured elapsed performance

Retain the approved method and address readiness or technical constraints through their owners.