Four Plants, Four Flavors of Lean
Meridian Components has four plants and a five-year-old lean program with no common definition of success. Plant A runs 5S with quarterly audits and scores 4.6 out of 5; it has logged 142 kaizen events in two years, but on-time delivery is 86% and customer complaints are up 22%. Plant B staffed nine Six Sigma black belts; scrap fell 18% while order-to-cash lead time stayed flat at 26 days. Plant C runs no formal projects at all, only daily huddles on every line, a supervisor standard work routine, and a single A3 board per value stream; its on-time delivery is 97% and inventory turns are 8.1. Plant D started four months ago, hired a lean director, and has already issued an RFQ for an electronic andon system. Corporate has now chartered a council to write a single lean playbook and a common maturity scorecard benchmarked quarterly across all four sites. The operations VP asks you to advise on scope.
1. What explains the pattern of results across the four plants?
Read the worked answer
The results track management system maturity, not tool sophistication.
Plant A is tool-rich and system-poor. A 5S audit score measures whether the plant is good at 5S audits. Nothing connects the 142 events to delivery or quality, so the plant is optimizing conditions inside the four walls while the value stream still runs on expediting. Complaints rising alongside a 4.6 score is not a paradox; it is what happens when improvement activity is decoupled from customer-facing measures.
Plant B is optimizing a local metric. Scrap is a plant metric; lead time is a value-stream metric. Nine black belts can drive scrap down at individual operations while inventory piles up between those operations. Flat lead time with lower scrap is exactly the predicted outcome.
Plant C looks unsophisticated and is the only one running a management system. Daily huddles, supervisor standard work, and value-stream A3s create a daily loop where abnormalities surface, get an owner, and get escalated. Ninety-seven percent on-time delivery and 8.1 turns are the output of that loop, not of any particular tool.
Plant D is buying a symptom. An andon system installed before the escalation routine and response standard exist will produce a light nobody answers, and the plant will conclude lean does not work.
The common thread: the three weak sites purchased or deployed tools, while the strong site deployed routines with owners, cadence, and escalation.
2. Recommend the boundary: what should corporate standardize, and what should each plant keep local?
Read the worked answer
Standardize the system that governs how problems are found, escalated, and closed. Leave the choice of improvement tools local.
Standardize, with corporate owning the definition and the audit:
- The daily huddle: cadence, duration, board content, who attends.
- The escalation path and its clock - how long an open abnormality may stay at team level before it moves to the value stream manager.
- Leader standard work for supervisors and value stream managers, including the gemba walk and its questions.
- The A3 format and the review cadence, including what evidence closes an action.
- Standard work documentation conventions for core processes.
- Metric definitions. On-time delivery must be calculated identically at all four sites or the scorecard is fiction.
- Value stream mapping cadence and the gate review that authorizes future-state work.
Leave local: which improvement methods a value stream uses (SMED, Six Sigma, 5S, kaizen events), which projects get selected within the future-state plan, and how improvement is resourced and celebrated.
Sequencing: define the routines and the metric definitions first, run them for two quarters, and only then build the maturity scorecard. Otherwise corporate will spend a year scoring the scorecard, and every site will learn to manage the score rather than the value stream. Give corporate ownership of definitions and plant general managers ownership of results, and say so in the charter.
3. How do you bring Plant A into the standard without destroying the participation that is its one real asset?
Read the worked answer
Plant A's engagement is genuine and worth keeping; its problem is direction, not energy. Treat it as a redirection, not a correction.
Do not open with the score. Leading with "your 5S score is meaningless" guarantees defensiveness and loses the room. Open with the customer data: 86% on-time, complaints up 22%, and ask the plant to explain the gap. They already know.
Use their strength as the vehicle. Put their 5S audit data against the complaint and late-order pareto. The exercise is not to abandon 5S but to point it at the constraint. Ask the kaizen program to run its next events inside the value stream of the top-complaint product family, with a customer metric attached to each event charter.
Change the audit's purpose, not its existence. Shift from a 4.6 score to condition-plus-countermeasure: what did the audit find, who owns it, when is it closed. Keep the audit rhythm; change what it produces.
Give them a visible pilot. Make Plant A the first site for leader standard work, with a 90-day target on one customer-facing measure. Pair them with Plant C as a buddy site rather than a corporate auditor - peer credibility does more than a mandate.
Protect the people. If the plant manager who built the 142 events is publicly demoted in status, the organization learns that engagement is punished. Reframe his program, keep him visible, and hold him to the new measure.