Gap Analysis
Key Takeaways
- Gap analysis compares current performance to a predefined future-state target using agreed metrics so the team can size the problem and prioritize Analyze work.
- A performance gap is Target − Current (or Current − Target when lower is better); always state the unit, time window, and data source with the number.
- Build gap analysis from a clear Y metric, a validated baseline, a documented future-state goal, and stratification so you know where the gap is largest.
- Gap analysis identifies and quantifies what must close; it does not by itself prove root causes—use root-cause tools next.
- CSSGB Analyze items often test whether you can read a gap scenario, choose the right metric comparison, and avoid treating the gap statement as a solution.
Gap Analysis
Quick Answer: Gap analysis compares current-state performance to a predefined future-state (or customer/specification) target using agreed metrics. The gap is the numeric difference the DMAIC project must close—not a solution list and not a root cause by itself.
Why Gap Analysis Belongs in Analyze
By Analyze, Measure has usually given you a baseline Y and a sense of variation. Gap analysis turns that baseline into a decision-grade delta: How far are we from the goal? Where is the shortfall worst? How large must Improve be to matter?
ASQ’s CSSGB BoK places gap analysis at IV.C.1 (Analyze): you analyze scenarios to identify performance gaps—not merely write a charter slogan. Use data, clear metrics, and current-vs-future comparison, then interpret what the gap implies for prioritization.
Current State vs Future State
| Element | Meaning | Green Belt practice |
|---|---|---|
| Current state | How the process performs now | Validated Measure baseline (mean, median, rate, cycle time, etc.) |
| Future state | The predefined project target | Spec limit, VOC CTQ, management goal, or charter primary metric |
| Gap | Difference that must close | Target − Current (higher is better); Current − Target (lower is better) |
| Metric | Unit and definition of Y | Same definition, window, and inclusion rules on both sides |
Higher is better: First-pass yield baseline 92%; goal 98%. Gap = 6 percentage points.
Lower is better: Mean quote cycle time 4.8 days; goal 2.0 days. Gap = 2.8 days.
Never mix definitions: “orders with no rework” vs “orders with no customer complaint” are different Y’s. A gap on mismatched definitions is fiction.
Predefined Metrics—Not Gut Feel
Fix metrics before storytelling:
- Primary Y from the charter (plus consequential metrics if needed).
- Operational definition — what counts as a defect, cycle, or complete order.
- Time window and stratification — e.g., last 90 days by plant, product, shift, channel.
- Baseline method — mean, median, proportion, DPMO, or capability—consistent with Measure.
- Future-state source — customer CTQ, regulatory limit, management stretch, or benchmark documented with the sponsor.
If the target was never predefined, you are inventing a goal under pressure. Reset the goal with the sponsor, then measure the delta.
How to Run Gap Analysis
- Confirm Y and goal from the charter after Measure still makes sense.
- Refresh current state with the baseline or a stable updated window; note sample size and dates.
- Compute the gap in the correct direction and unit.
- Stratify by product, site, shift, failure mode, or segment. One overall number often hides a small high-gap stratum and a large low-gap stratum.
- Translate into focused problem statements (“Plant B is 11 points below target; Plant A is 1 point below”).
- Size opportunity when useful—annual units, CoPQ, or hours—so prioritization is business-real.
- Hand off to root-cause tools. Fishbone, 5 Whys, fault trees, and data analysis explain why; gap analysis only sizes the shortfall.
Worked Scenario — Claims Lead Time
Primary metric: median end-to-end lead time (submit → paid/denied), lower is better.
| Slice | Current median (days) | Future-state target | Gap (days) | Volume share |
|---|---|---|---|---|
| Simple claims | 3.1 | 2.0 | 1.1 | 55% |
| Complex claims | 12.4 | 5.0 | 7.4 | 30% |
| Appeals | 18.0 | 10.0 | 8.0 | 15% |
| Overall | 7.2 | 4.0 | 3.2 | 100% |
Interpretation: Overall gap is 3.2 days, but complex + appeals dominate the absolute shortfall. Chasing only high-volume simple claims may close little of the blended gap. Analyze should prioritize drivers of long complex and appeal paths while still watching simple-claim waste. When useful, convert the gap into annual claim-days or CoPQ so sponsors see why the stratified focus is not “ignoring” simple claims.
Exam Patterns and Pitfalls
Common patterns use the same skill—same metric, two states, quantified difference, stratified insight—whether the Y is capability vs. spec, on-time or AHT service levels, scrap/CoPQ, or DPMO/sigma goals.
Avoid:
- Treating “we need new software” as a gap (that is a solution guess).
- Moving the future-state target to erase bad news.
- Aspirational targets with no sponsor ownership.
- Omitting the direction of better when reporting the number.
- Skipping stratification so one plant or SKU hides inside an enterprise blend.
- Treating gap size as root cause—causes, tests, and controls still follow.
DMAIC fit: Define states the goal; Measure locks current state; Analyze quantifies and stratifies the gap then hands off to RCA; Improve must close a material share of the gap; Control watches whether it stays closed.
Expect items that compute Target vs. Current, choose the right comparison, size shortfalls with stratification, and refuse to confuse a gap statement with a root cause or Improve action. On scenario questions, first name the metric and direction of better, then compute or select the gap, then decide what to stratify next. Measure the distance to the goal first, then dig for why.
A call center’s primary metric is average handle time (AHT), where lower is better. Current AHT is 8.6 minutes and the predefined project target is 6.0 minutes. What is the performance gap, and what does gap analysis alone establish?
A Green Belt finds overall first-pass yield is 3 points below target. Stratification shows Product Line A is 1 point below target at 80% of volume, while Product Line B is 12 points below target at 20% of volume. What is the best Analyze conclusion?