2.6 Organizational Barriers and Common Causes of Six Sigma Failure

Why deployments fail with good tools in the room

Six Sigma failures are rarely statistical. The methodology is mature, the tools are validated, and the training is standardized. Deployments fail because the organization around the project makes success structurally difficult. ASQ tags this topic at the Apply level, which means the exam will hand you a symptom and expect the diagnosis.

Structural barriers

Functional silos. Most DMAIC projects cross departmental boundaries; most organizational charts do not. When a value stream runs through purchasing, production, quality, and logistics, no single manager owns the process, and each owns a metric that can be improved at the others' expense. The countermeasure is structural, not motivational: a named executive sponsor with authority over the whole value stream, a charter signed by every affected function, and a system-level project metric that no single department can move alone.

Hierarchy and decision latency. In steep hierarchies, a Black Belt's recommendation queues behind several approval levels, and the project's improve phase stretches past the sponsor's attention span. The countermeasure is a pre-agreed decision-rights matrix in the charter: what the team may decide alone, what the sponsor decides, and what escalates.

Geographic and shift dispersion. A process running across three shifts and two sites has three or six versions of standard work. Data collected without stratifying by shift and site hides the largest source of variation in the study.

Resource barriers

ASQ names lack of resources as a core cause of failure. In practice it takes four forms:

A Black Belt assigned "50% to projects" is, in effect, assigned 0% during any operational crisis -- which is exactly when the process data is most interesting. The full-time Black Belt model exists because part-time assignment is the most common quiet cause of project failure.

Cultural barriers

Lack of management support is the failure ASQ names first, and it is usually visible in behaviour rather than statements. Executives who publicly endorse Six Sigma but never attend a toll gate, never ask about a control plan, and never fund an implementation have withdrawn support in the only currency that matters.

Fear and blame. If the data will identify whose area causes the defects, people will manage the data. Under-reporting, redefining defects, and "cleaning" records before the study are all rational responses to a blame culture and all fatal to a Measure phase.

Not-invented-here and change fatigue. Organizations that have run several improvement programs carry scar tissue. Frontline staff who watched the last initiative disappear will comply visibly and revert quietly. This is a sustainment failure that shows up months after the project closes, which is why VIII.D exists in the Body of Knowledge.

Reward system misalignment. This is the most under-appreciated barrier because it is invisible on an org chart. If operators are paid for units produced, they will run through a marginal condition rather than stop and adjust. If a plant is measured on machine utilization, it will overproduce and build inventory. Improvement dies wherever the reward system pays for the behaviour the project is trying to eliminate.

A diagnostic sequence

When a project stalls, work down the list in order; the first "no" is usually the real problem.

1. Is there a named sponsor with authority over the entire process in scope?
2. Is the project metric one that no single department can move alone?
3. Is the Black Belt's time formally protected, and the SME time allocated?
4. Can the team obtain trustworthy data without exposing an individual to blame?
5. Does any existing reward or target contradict the improvement?
6. Is there an approved funding path for the eventual solution?
7. Does the process owner accept accountability for the control plan after handover?

Worked example

A hospital charters a project to cut emergency-department boarding time. Six weeks in, the team cannot get inpatient discharge timestamps, the sponsor is the ED director, and nursing unit managers are measured on bed occupancy. The technical work is competent and the project is doomed.

The diagnosis is not "resistance to change". It is three specific, nameable barriers: the sponsor's authority stops at the ED door (silo), the data lives in a system owned by another function (resource and information), and the inpatient reward system pays for high occupancy, which is precisely the condition that creates boarding (reward misalignment). The correct action is to re-sponsor the project at the chief operating officer level, add inpatient nursing to the charter, and put a system-level measure -- total time from decision-to-admit to bed occupancy -- in place of the ED-only metric.