Chartering a Six Sigma Portfolio When Demand Exceeds Belt Capacity
Northline Plastics (injection molding, $85M revenue) has trained 12 Black Belts and 40 Green Belts over 18 months. Scrap and rework run 4.2% of sales; the CFO calls that number 'accounting fiction' and has capped program funding at Q2. The VP of Operations wants 25 projects launched in Q1 to 'show momentum.' Only three belts are full-time; the rest carry roughly 20% project time. A key customer scorecard shows 2 defects per million and ranks Northline last among suppliers on delivery. The two strongest belts are already committed through June to a plant-transfer project, and the quality lab can support only four MSA studies per month.
1. Using the capacity and lab numbers given, how many projects should be chartered in Q1, and what selection logic do you apply? Show the arithmetic.
Read the worked answer
Capacity is the binding constraint, not enthusiasm.
- Full-time belts: 3, each realistically carrying one project at a time over a 5–6 month cycle (≈0.6 FTE per project) → 3–5 concurrent, and two of the three are already committed.
- Part-time belts: 9 at 20% time can each hold one charter, but at that load most will not clear Measure in a quarter. Count them as ~6 effective concurrent projects, not 9.
- Lab: 4 MSA studies/month × 3 months = 12 MSA slots. Every project reaching Measure needs at least one. Launching more than ~8 by March guarantees a queue and stalled projects at the gate most likely to reveal bad baselines.
Recommendation: charter 8–10, not 25.
Selection logic - a weighted matrix, scored by sponsor and belt together:
- Validated COPQ and customer impact. Delivery ranks last on the customer scorecard, so delivery-cycle projects get extra weight over internal scrap projects.
- Data availability and MSA feasibility for the specific metric.
- Named process owner and a sponsor who will sign the charter.
- Learning value for newer belts (bounded scope, clean data).
Merge duplicates, sequence MSA-heavy projects so they don't collide at the lab, and reject outright any charter lacking a validated baseline and an accountable process owner.
2. The CFO disputes the $3.57M cost-of-poor-quality figure and is withholding Q3 funding. What do you do in the next 30 days?
Read the worked answer
Do not argue the number in a steering meeting - rebuild it with the CFO's staff.
- Gross arithmetic is not the issue: 4.2% × $85M = $3.57M. The issue is composition. Split into hard (scrap material, rework labor at actual rates, warranty, expedite freight, credits) and soft (absorbed capacity, avoided capital, price erosion). Hard items should tie to general-ledger accounts line by line; soft items get listed separately and never counted as savings.
- Expect hard COPQ to land near $2.0–2.4M. Publish that as the defensible figure and treat $3.57M as a ceiling.
- Agree the savings-validation method before charters are signed: who signs off, at what point (project close plus 90 days), and how baseline is frozen.
- Bring this to the portfolio review as a joint finance/quality document. A CFO who co-authors the model stops being an opponent and becomes the control on inflated claims - which is exactly what you want, because one debunked savings claim costs more credibility than ten honest ones.
- Tie the funding ask to the 8–10 charters already selected, not to a program-level abstraction.
3. Design the governance cadence. Who owns what, and what happens at tollgates?
Read the worked answer
Three layers, each with a distinct decision right:
| Layer | Frequency | Owns |
|---|---|---|
| Belt + process owner | Weekly | Data collection, analysis, daily obstacles |
| Sponsor | Biweekly | Charter scope, resources, removing cross-functional blocks |
| Steering (VP Ops, CFO, Quality) | Monthly | Portfolio entry/exit, funding, validated savings |
Tollgates at each DMAIC phase with written exit criteria, not presentations. Define gate: charter signed by sponsor and CFO, baseline method agreed. Measure gate: MSA passed, sampling plan documented. Analyze gate: root causes verified with data, not voting. Improve gate: pilot results replicated, control plan drafted. Control gate: 90 days of sustained performance plus finance-validated savings.
Governance rules that actually prevent failure:
- No project enters the portfolio without a validated baseline and named owner.
- Belt loading is a published number; a belt cannot be assigned a fourth project because someone is impatient.
- Monthly dashboard shows validated savings against plan and projects past gate - not counts of belts trained.
- Escalation path is explicit: belt → sponsor → steering, with a decision requested, not a status update.