240 Kaizen Events and No Result: What to Scale and What to Refuse

The programme's unit of improvement was the event, not the value stream. Events end at the event boundary; there is no owner of the end-to-end flow, so local gains are absorbed back into the system - capacity is freed, then immediately filled with WIP because the scheduling rules, batch sizes and MRP logic were never touched. That is exactly why inventory rose 9% while everyone was 'doing Lean': freed capacity without a pull signal becomes overproduction.

Specific failure modes:

• Activity metrics masquerading as results. 240 events, 4.2/5 on 5S and 96% trained measure effort and tidiness. None measures flow, lead time or delivery. The 5S score is the most dangerous: it is auditable, it always rises, and it is uncorrelated with OTIF.
• No order-to-cash VSM means the group never agreed on what the problem was, never located the constraint, and never sized the gap between current and future state.
• Event selection by volunteering or consultant preference optimized for enthusiasm, not for constraint or cash impact.
• Structural weakness: coordinators report to plant managers and two double as maintenance managers, so there is neither dedicated capacity nor authority to change cross-functional rules.
• No daily management. Without tiered accountability, nothing holds a gain after the consultant leaves.

OTIF falling means due-date performance and the constraint were never addressed; Plant C is the counter-example that proves the point - same toolset, but connected along a value stream.

Refuse the 20-event × 12-site calendar. With two additional engineers that is 240 events in nine months - the same volume that just failed, at higher cost, and it repeats the activity-not-results design. Also refuse to scale before the holding mechanism (daily management) exists; replication of an unheld gain just multiplies rework.

Negotiate to a value-stream-led deployment:

Selection criteria for the first waves: streams with the largest OTIF or cash gap, a single accountable value stream manager with authority over scheduling, stable enough demand to see the effect, and a site leader who will personally run leader standard work. Pick two or three sites and roughly four streams - not twelve sites.

Per stream, 90 days: an order-to-cash current and future-state map; a value stream manager named and empowered; two or three KPIs only (OTIF, lead time, WIP); a plan that changes rules (scheduling, batch size, pull, changeover) not just workplaces; and tiered daily accountability installed in the first 30 days so gains are held.

Capability: coaching on the line, not more classroom. Two coaches can realistically support one new stream each per quarter - so four to six streams in year one, with internal coaches being certified as they go.

Structure: make the CI coordinator role dedicated, trained and dual-reporting to the group CI lead, not to the plant manager alone.

The trade-off to present to the COO: publish it in units he cares about - 'three sites and four streams this year, targeting 8–10 OTIF points and 20% WIP reduction on those streams; the remaining nine sites begin in year two once the gate is met.' A committed, gated plan beats a symbolic 12-site mandate.

Gate (all conditions, not a score):

1. The stream's OTIF has held at or above target for eight consecutive weeks after the CI team stepped back.
2. Tiered daily meetings run at every level without a CI person in the room, and escalated items close with a named owner and a date.
3. A local person has led a full order-to-cash VSM and taken at least one A3 to a verified countermeasure.
4. Leader standard work exists and is audited by a peer, not self-reported.
5. Inventory and lead-time gains are confirmed by physical or ERP audit, not by the event report.
6. The value stream manager has scheduling authority in writing.

Measure set:

• Lagging (outcomes, one set for the group): OTIF, order-to-cash lead time, inventory turns, cost per unit, safety.
• Leading (system health): percentage of streams with daily management operating; escalation closure time; A3s closed with verified countermeasures; certified internal coaches; percentage of improvement plans that changed a rule rather than a workplace.
• Explicitly not used as a success measure: event count, training completion, 5S audit score, newsletter frequency.

Anti-theatricality: audit the data source against the ERP and the shop floor on a sample basis each quarter; require every stream to report a baseline before its first intervention; and publish failures alongside successes - a deployment where no stream failed a gate is not being measured. If the leading measures improve for two quarters without lagging movement, the diagnosis is wrong and the deployment should stop and be re-scoped, not accelerated.