ORIGINAL COMPANY-CONTEXT PRACTICE
Calculate yield without averaging incompatible percentages
Original practice scenario: A hypothetical memory manufacturing dashboard averages daily pass percentages across test stations. Stations process different quantities and some units are retested. Design a corrected metric.
Written for independent practice using company context; this is not a verified question from a company interview.
A REASONED APPROACH
Model answer.
Define the tested-unit population and retest policy before aggregating. An unweighted average of percentages gives a low-volume station the same influence as a high-volume station.
Work through it
- Retain unit, test attempt, station and outcome identities, selecting the intended first-pass or final-pass measure with an explicit rule for incomplete attempts.
- Aggregate passed and eligible counts at compatible product and time grains, deriving the combined percentage from summed counts rather than averaging station percentages.
- Reconcile units entering and leaving the test population, showing retest frequency and coverage alongside yield to preserve operational meaning.
Tradeoffs to discuss
- First-pass yield exposes rework while final-pass yield describes eventual acceptance; reporting both avoids conflating those purposes.
Common pitfalls
- Do not count a unit once for every successful retest in the numerator.
- A station with no eligible tests has an undefined rate, not automatically zero or one hundred percent.