Lean Manufacturing Transformation Checklist

Value Stream Mapping

    Pick a single product family that shares routings and shares roughly 70% of its work centers — not the whole plant. Use the product-family matrix (process steps × part numbers) to confirm. The CI lead and production manager co-own this selection; mapping the wrong family burns the entire kaizen budget.

    Walk the value stream backwards from shipping to receiving with a cross-functional team (operator, supervisor, planner, quality, maintenance). Capture cycle time, changeover time, uptime, scrap rate, and WIP at each work center on the spot — don't trust ERP standards without floor verification.

    Takt = available production time ÷ customer demand. Compare cycle time at each station to takt to find the bottleneck (CCR). Total lead time = process time + queue time across the stream — typically 95%+ is queue, which is where the kaizen leverage lives.

    Design the future state around continuous flow where possible, supermarket pull where flow isn't feasible, and a single pacemaker process. Mark the kaizen bursts (5S, SMED, line balance) on the map — these become the work packages for the rest of this checklist.

5S Implementation

    Tag any item not used in the last 30 days and stage it in a red-tag holding area for 30-day review. Common finds: obsolete fixtures, old prints at superseded revs, abandoned WIP, expired chemicals (check SDS dates). Plant manager dispositions the holding area at day 30.

    Shadow-board hand tools at point of use; tape footprints for carts, totes, and trash. Each item gets one home — and the home is labeled. Aim for any tool retrievable in under 30 seconds. Operators do the layout, not engineering — they own the cell.

    Shine isn't cosmetic — it's TPM cleaning-as-inspection. Operators clean while looking for oil leaks, air leaks, loose guards, missing bolts, frayed cables. Log findings in the CMMS (Fiix, eMaint, MaintainX) as PM tickets, not as a separate punch list that disappears.

    Use the standard LPA scoring sheet — operator self-audit weekly, supervisor weekly, plant manager monthly. A passing audit means every zone scores above the threshold; one failing zone fails the cell. Score below threshold triggers the remediation step that follows.

    Triggered when the baseline audit fails. Schedule a half-day blitz with the cell team to fix the failing zone — don't paper over it with a system note. Re-audit within one week and confirm closure before moving to the continuous-flow phase.

Continuous Flow

    The CCR is the work center whose cycle time exceeds takt — that's where throughput is set. Confirm with OEE data from MachineMetrics or your MES; don't trust the lunchroom consensus on which machine is the bottleneck.

    Build a yamazumi (work-balance) chart by station and shift work elements until each station runs at or just under takt. Industrial engineering owns the time studies; the cell lead validates that the rebalanced work is actually executable on the floor.

    Halve the EPEI (every-part-every-interval) and watch what breaks — usually changeover time, which feeds directly into the SMED phase. Don't drop batch sizes without addressing changeover, or you'll just create more downtime.

    Size kanban quantities from average demand × replenishment lead time + safety. Use two-bin or card kanban at the cell; reserve electronic kanban for cross-plant or supplier loops. Materials planner adjusts ERP min/max to match — otherwise MRP will keep over-ordering.

Setup Reduction (SMED)

    Record from last-good-part to first-good-part — the full clock, not just the wrench-turning. Review the video with the setup operator and the manufacturing engineer; the time is always longer than people remember.

    Internal = machine must be stopped (die change, alignment). External = can happen while the machine runs the prior job (staging tooling, pre-heating, pulling the kit). Most baseline changeovers have 30-50% internal time that's actually externalizable.

    Write a setup standard work sheet with sequence, time, tools, and torque values. Post it at the machine. Setup operators across all shifts run the same sequence — variance kills the gain. Roll the standard into the operator's annual training matrix.

    Run the new standard for five consecutive changeovers and time each one. SMED target is typically 50% reduction by the first iteration. If the average misses target, schedule the follow-up workshop in the next step.

    Triggered when the verification round misses the changeover target. Re-watch the video with a focus on the slowest 20% of steps; common second-pass wins are quick-clamp fixtures, color-coded tooling, and pre-set die heights. Document the second-pass standard before closing.

Quality at the Source

    Publish a written stop-the-line policy signed by the plant manager: any operator can stop production for a quality concern with no retaliation. The policy is hollow without management response — define a 5-minute supervisor response SLA and audit it for the first 30 days.

    Pull the top three defect modes from the last 90 days of NCR data. For each, design a mistake-proofing fixture, sensor, or check that makes the defect physically impossible — not just visible. Examples: keyed connectors, presence sensors, weight checks, contrast vision systems.

    The cell's process flow has changed — PFMEA must reflect the new steps, new failure modes from the rebalance, and new controls from poka-yoke. Quality engineer leads; manufacturing engineer and operator sign off. Update the control plan in lockstep so they don't drift.

    Plant manager and CI lead review the before/after metrics: lead time, WIP, OEE, scrap rate, changeover time, audit score. Capture the decision, the lessons learned, and the photos of the transformed cell. Schedule the 30-day sustainment audit before closing.

Use this template in Manifestly

Start a Free 14 Day Trial
Use Slack? Start your trial with one click

Ready to take control of your recurring tasks?

Start Free 14-Day Trial


Use Slack? Sign up with one click

With Slack