Implementation Planning & Pilots

Implementation Planning & Pilots (CSSGB BoK V.B — Apply)

Quick Answer: Before full rollout, Green Belts validate solutions with proof of concept, try-storming, simulation, and pilot tests. Pilots prove the fix works under real conditions, limit risk, and produce data for a go/no-go scale-up decision.

In Improve, selected countermeasures must be implemented carefully. A statistically sound root cause does not guarantee a smooth deployment. BoK V.B is at the Apply level: choose and plan staged validation (proof of concept, try-storming, simulation, pilots) before enterprise-wide change.

Why Staged Implementation Matters

Full cutover on day one can:

• Disrupt customers if the solution fails under load.
• Create safety, quality, or compliance exposure.
• Waste capital on unproven equipment or software.
• Destroy trust when teams are forced into untested standard work.

Staged validation answers three questions: Does it work? Does it work here? Can we scale it without losing the gain?

Proof of Concept

A proof of concept demonstrates that a proposed solution can achieve the intended effect under simplified or laboratory conditions. Examples: a fixture that reduces assembly time on five units; a new query that correctly flags defects in a historical data set; a supplier trial lot that meets a tighter specification.

POC is not a full pilot. It may ignore training, IT integration, or full demand mix. Use POC early when technical risk is high (new technology, unproven supplier, novel measurement). If the POC fails, stop or redesign before spending on pilots.

Exam tip: POC proves feasibility; a pilot proves operational performance in the real process.

Try-Storming

Try-storming (try + brainstorm) means testing ideas quickly with rough prototypes, cardboard layouts, mock cells, or temporary process rules—learning by doing rather than only debating. Teams might rearrange a workstation with tape and crates, walk a temporary spaghetti path, or run a few orders with a temporary checklist.

Benefits:

• Surfaces practical constraints (reach, ergonomics, IT clicks) that whiteboards miss.
• Builds operator ownership early.
• Cheaply eliminates weak ideas before formal pilots.

Try-storming fits after root-cause analysis when multiple countermeasures compete. Document what was tried, what failed, and what advanced to pilot design.

Simulation

Simulation models process behavior—discrete-event simulation of a line, spreadsheet capacity models, or Monte Carlo of cycle-time distributions—to predict outcomes before physical change.

Use simulation when:

• Capital or safety risk of a live trial is high.
• Interactions are complex (multiple products, shared resources, random arrivals).
• You need to compare staffing, batch size, or layout scenarios quickly.

Limitations: models are only as good as inputs (cycle times, failure rates, demand). Always validate critical assumptions with Gemba data. Simulation supports decisions; it does not replace a pilot when human factors and real defects matter.

Pilot Tests

A pilot runs the improved process on a limited scale under real operating conditions. Limits may be one shift, one product family, one line, one clinic, or a percentage of volume.

Pilot design checklist

1. Scope: What is in / out (products, customers, sites, time window)?
2. Success criteria: Primary Y (defect rate, cycle time, cost) and consequential metrics (overtime, scrap, customer complaints).
3. Baseline: Pre-pilot performance on the same metrics and similar mix.
4. Standard work draft: Temporary procedures, visual aids, and who does what.
5. Training: Operators, supervisors, and support groups ready before go-live.
6. Data plan: What is measured, how often, who reviews, and decision dates.
7. Rollback plan: How to restore the old process if the pilot fails safely.
8. Communication: Stakeholders know this is a test, not a permanent cutover.

Example

A team reduces order-entry errors with a redesigned screen and poka-yoke fields. Instead of deploying company-wide, they pilot one customer segment for four weeks. Baseline error rate was 3.2%; pilot target is ≤1.5% with no increase in handle time. Results: 1.1% errors, handle time flat. After tollgate review, they train remaining teams and roll out in two waves.

From Pilot to Full Implementation

Scale when:

• Success criteria are met (or exceptions understood and accepted).
• Special-cause issues from the pilot are fixed.
• Documentation, training, and control plan are ready.
• Capacity and change management support the larger footprint.

Full implementation planning typically includes a phased rollout schedule, owners (RACI), risk mitigations, resource needs, and control metrics. Tie actions to a Gantt or project plan and to Control-phase sustainment (control plan, audits, SPC where appropriate).

Common Exam Traps

Green Belt sequence: select solution → try-storm or POC if needed → simulate complex flow if useful → design pilot with criteria → run and measure → go/no-go → phased full implementation → Control.

Implementation planning is risk management: prove the fix in small steps so Improve delivers lasting results rather than a one-time “big bang” that rebounds.