A checkpoint between stages. Useful as a backstop, weak as a strategy.
A quality gate is the cleanest example of a quality tool that everyone understands and most shops use too much of. The structure is simple: pick a stage boundary, define what "ready to move forward" means, and put a checkpoint in place that work cannot skip. Used surgically, gates are valuable. Used as the main quality strategy, they paper over upstream problems and add cost without removing causes.
"Every quality gate is a confession that the upstream process is not trusted yet."
A gate has four pieces.
A location. Usually a stage boundary where the cost of moving a defect forward goes up sharply. Before paint, before kit assembly, before shipment, before a costly finishing step. Gates are most useful where the next step adds value that is hard to recover if you find a defect later.
A criteria list. Explicit pass/fail conditions, ideally written in language the operator and the inspector both understand. Vague gates ("looks acceptable") produce drift; specific gates ("all four datum dimensions within tolerance, no surface defects larger than 0.5mm visible at 18 inches") produce consistency.
A staffing or instrumentation model. Someone or something has to enforce the gate. A QA tech at a bench, a vision system on a conveyor, an operator running a documented self-check with a sign-off. If the gate can be quietly bypassed, it is not a gate.
A fail path. What happens to work that fails the gate. Tagged. Segregated. Routed to rework or scrap. Logged as a nonconformance if appropriate. A gate without a defined fail path becomes a place where parts accumulate informally and quietly slip through when production pressure rises.
Gates work best when they are few and meaningful. A shop with 12 sequential gates spends more time at gates than at value-adding work, and operators start to treat the gates as someone else's responsibility for quality.
Picture a 35-person contract machine shop running medical parts for a regulated customer. The customer requires a final QA gate on every lot before shipment, including dimensional verification on a CMM and a documentation review. The shop has the gate but is also running three internal gates: one between rough machining and finish machining, one between deburring and washing, and one between assembly and packaging. The gates themselves are not the problem. The cumulative queueing at all four gates has added two days to the lead time.
A lean review of the gates would keep the final pre-shipment gate (it is required, and the cost of an escape is high) and the gate before packaging (the customer audits packaging, and packaging-related rework is expensive). The two internal gates between machining steps get replaced by self-inspection: the operator finishes their step, runs the dimensional check at their bench against the print, signs the inspection sheet, and moves the work along. The two gates disappear as queue points. Lead time drops by a day and a half. The defect escape rate to the final gate stays roughly the same because the self-inspection was catching the same issues earlier, with less queueing.
A quality gate is one specific location where inspection happens, structured as a pass/fail checkpoint. The lean preference is to push the same check earlier with quality at the source, so the operator catches the issue at the bench before any further value is added. Within a single operation, the same logic produces successive-inspection, where the next operator checks the previous step. The strongest companion to a meaningful quality gate is a credible stop the line discipline, so problems found at the gate trigger an upstream investigation, not just a rework ticket.
The questions we hear most about this term.
Long-form guides that pick up where this definition leaves off, written for manufacturers running Arda today.
Same-day setup. No distributor lock-in. Zero stockouts. Top teams double revenue in 9 months.
.svg)
.svg)
%201.svg){kind=logo}
.png)
.svg)
.svg)
