8.11 Implementation Planning, Piloting, and Solution Selection

The Improve phase transitions a project from root cause identification to solution design, risk evaluation, pilot testing, and full-scale implementation. Black Belts apply structured decision tools to select optimal solutions that maximize customer value while minimizing implementation risks.

---

Systematic Solution Selection Tools

1. Solution Selection Matrix (Pugh Concept Selection Matrix)

Compares candidate solution concepts against baseline current process performance across key evaluation criteria (CTQs, cost, feasibility, risk):

• Baseline Scoring: Current process is assigned a rating of $0$ across all criteria.
• Candidate Evaluation: Alternatives are rated as $+1$ (Better), $0$ (Same), or $-1$ (Worse) relative to baseline.
• Weighted Total: Sum of weighted scores guides selection of winning solution concept.

2. Effort vs. Impact Prioritization Matrix

Ranks potential solutions into four action quadrants:

• Quick Wins (High Impact, Low Effort): Implement immediately.
• Major Projects (High Impact, High Effort): Plan carefully; allocate capital resources.
• Fill-Ins (Low Impact, Low Effort): Implement if resources permit.
• Thankless Tasks (Low Impact, High Effort): Discard.

---

Failure Mode and Effects Analysis (FMEA) for Risk Mitigation

FMEA is a proactive risk assessment tool used to identify potential failure modes in proposed solutions before full-scale deployment.

Calculation of Risk Priority Number (RPN)

Every failure mode is scored on three 1-to-10 ordinal scales:

$\text{RPN} = \text{Severity (S)} \times \text{Occurrence (O)} \times \text{Detection (D)}$

• Action Threshold: Failure modes with RPN $> 100$ or Severity $\ge 8$ require immediate corrective action plans.

---

Pilot Implementation Strategy

A Pilot Study is a small-scale, controlled trial of the proposed solution conducted prior to full rollout.

Objectives of a Pilot Study

1. Confirm real-world solution effectiveness without exposing entire customer base to risk.
2. Validate predicted financial COPQ savings.
3. Identify unexpected operational friction, operator training gaps, and measurement flaws.
4. Refine standard operating procedures (SOPs) based on empirical pilot feedback.

---

Poka-Yoke (Error-Proofing) Taxonomy

Poka-yoke mechanisms prevent human errors from turning into customer defects:

1. Prevention-Type Poka-Yoke: Physical design mechanisms that make it impossible for an error to occur (e.g., asymmetric electrical connectors that fit only one way).
2. Detection-Type Poka-Yoke: Automated sensors that immediately halt the process and alert operators if a non-conformance occurs (e.g., photoelectric light curtains verifying part placement).
3. Control-Type Poka-Yoke: Automatic shutoff switches that interrupt power if operating parameters exceed safety limits.

---

Pilot Validation & Statistical Conformance Metrics

Before launching full-scale implementation, pilot results are evaluated using formal statistical tests:

• Two-Sample $t$-Test / Paired $t$-Test: Verify that mean process performance during pilot ($Y_{\text{pilot}}$) is significantly better than baseline ($Y_{\text{baseline}}$) at $\alpha = 0.05$.
• $F$-Test / Levene's Test: Verify that process variance during pilot ($\sigma^2_{\text{pilot}}$) is significantly lower than baseline.
• Process Capability Re-Assessment: Confirm that pilot $C_{pk}$ achieves $\ge 1.33$ target.

---

Detailed Cost-Benefit Analysis (CBA) & Financial ROI

Before final solution selection, Black Belts compute financial payback metrics:

1. Net Present Value (NPV)

Calculates the present value of future cash flows generated by the solution discounted at cost of capital ($r$):

$\text{NPV} = \sum_{t=1}^{n} \frac{\text{Cash Savings}_t}{(1 + r)^t} - \text{Initial Investment}$

2. Payback Period

Calculates the exact elapsed time required for cumulative net financial savings to equal initial solution implementation capital:

$\text{Payback Period} = \frac{\text{Total Initial Capital Investment}}{\text{Annualized Net COPQ Savings}}$

• Benchmark Criterion: Industrial Six Sigma projects typically require a payback period of $< 12\text{ months}$ and a positive NPV.

---

Implementation Planning, Pilots, and Simulation

Body of Knowledge section VII.C asks Black Belts to develop plans for implementing proposed improvements, including conducting pilot tests or simulations, and evaluate results to select the optimum solution -- at the Evaluate level, the highest in this section.

The implementation plan

The two elements most often missing are the rollback plan and the acceptance criteria. Without a rollback plan, a failed implementation becomes an argument rather than a decision; without acceptance criteria agreed in advance, the pilot result is negotiated after the fact.

Simulation as an alternative to a physical pilot

Where a physical pilot is impossible, prohibitively expensive, or unsafe, simulation substitutes.

Simulation earns its cost when it explores many alternatives cheaply -- comparing six buffer configurations in an afternoon -- or when the interaction between variation and queueing makes the outcome genuinely unpredictable by inspection. Its limitation is that it can only be as good as its input distributions and its assumptions, so a simulation result should be treated as a prediction to be confirmed, exactly like a DOE model.

Designing the pilot

A pilot is an experiment, and it should be designed like one.

• Define the acceptance criteria first, in the project's primary metric, with the practical threshold.
• Size it. A pilot too short or too small to detect the improvement claimed proves nothing; use the sample-size calculation, not a convenient duration.
• Cover the sources of variation the full process contains: all shifts, both material lots, the difficult product variants. A pilot run only on day shift with the best operator predicts nothing about Sunday night.
• Measure the consequential metrics too, so a gain bought at another metric's expense is visible.
• Include a control where possible: an unchanged parallel line or the same line before and after with the same measurement definition.
• Plan the duration to cover at least one full cycle of the dominant time pattern.

Evaluating pilot results to select the optimum solution

Compare the pilot outcome against the acceptance criteria using a hypothesis test, not an eyeball comparison of two averages. Then decide among four outcomes:

Where several candidate solutions were piloted, select using the same weighted criteria that were agreed before the pilots -- effectiveness, cost, implementation time, risk, sustainability -- rather than re-weighting after seeing the results. Recording the criteria and weights in the charter is what makes that discipline possible.