Root Cause Finder

Faced with a list of failure categories and their counts, the instinct is to treat them all as problems to solve. The Pareto principle says don't. Across an astonishing range of systems — and defect data is a textbook case — roughly 80% of the losses come from about 20% of the causes. A handful of categories dominate; the rest are noise. The job is to find that handful, and a Pareto chart finds it by ranking the categories and tracing the cumulative share.

That cumulative line is what turns a bar chart into a decision. Sorting tells you which cause is biggest; the running total tells you how many causes you must address to recover most of the loss. When the cumulative crosses 80%, you've found the vital few — the categories that justify nearly all the investigation and corrective effort. Everything past that line is the trivial many, where equal effort buys almost nothing.

The shape of the Pareto is itself a diagnosis. A steep one — one tall bar, then a cliff — means a single dominant cause, often an excursion or one bad tool, and a focused point fix recovers the yield. A flat one — all bars similar — means no single culprit; the loss is broad and the right response is a systemic process program, not a hunt for one villain. Reading the shape before diving in saves chasing a dominant cause that isn't there.

Pareto is the prioritization hinge of the whole yield loop: the Yield Trend dashboard detects the drop, the Wafer Map Analyzer reveals the signature, this Pareto ranks the categories to attack, and the Wafer Scrap console prices the recovery — turning an alert into a focused, justified action plan.