Process Capability Console

Process capability answers a deceptively simple question: will this process reliably make parts inside the specification? The honest answer needs two things — how tightly the output is grouped, and where that group sits relative to the limits. Cp captures the first; Cpk captures both, and that's why Cpk is the number that gates production while Cp is just the best case.

The mechanism is in the tails. A normal process produces most of its output near the mean, but the tails extend toward the spec limits, and any part beyond a limit is a defect. Cpk works by measuring, in standard deviations, the distance from the mean to the nearest limit, and taking the worse of the two sides — because the closer limit's tail produces most of the defects. Drift toward one limit shrinks that distance and Cpk drops, even if the spread never changed. The shaded tails in this console are literally the defect rate.

That tail behavior is exponential, which has a profound practical consequence: capability improvements pay off faster than they look. Moving Cpk from 1.0 to 1.33 isn't a 33% defect reduction — it's from about 2,700 defects per million down to about 63, because the normal tail collapses as you push the limit further out. This is the whole logic of Six Sigma: chase capability into the far tail where defects become vanishingly rare.

The actionable insight is to diagnose before you fix. If Cp is healthy but Cpk is low, the problem is centering — often a cheap mean adjustment recovers it without touching variation. If Cp itself is low, you must reduce variation, which is harder. This console's Cp-versus-Cpk gap and Cpu/Cpl split tell you which. Then connect capability to fab output via the Defect Density and Yield consoles.