I First Pass Yield (FPY) vs. (RTY) Rolled-Throughput Yield
i. First Pass Yield (FPY)- common business metric, (aka Final Test Yield, or First Time Good)
1. FPY= (#Processed-#Defective/(#Processed) * 100%
ii. In-Process Yield (IPY) – First Pass Yield for the Individual Process Step (fairly common)
1. IPYn=(#Processed - #Defective)/#Processed*100%
iii. Rolled Throughput Yield (not-so-common) – The probability a part or transaction will move from beginning to end error-free.
1. approximated by = IPY1*IPY2*IPY3*IPY4*.....IPYn
2. More accurately measured by RTY= e^(-DPU)
II Defective Rate
a. Defective – A unit that contains at least one defect (i.e. a defective part)
b. Defect – An output of a process that does not meet a defined specification - Must be Measurable!
i. Relationship #Defects >/= #Defectives
ii. Nicks, Dents, Scratch, Width out of Tolerance, Incorrect information, etc.
c. Defects Per Unit (DPU)– Average number of defects on a single unit or part
i. Example: One “Defective” part has a dent, and scratch, and missing hardware. This is 3 Defects on one unit. DPU=3.
III First Pass Yield (FPY) vs. (RTY) Rolled-Throughput Yield (continued)
a. RTY tells the whole story that FPY misses, a more honest operations business metric.
b. RTY takes into account rework and the “hidden factory”, not just Scrap.
c. RTY </= FPY
i. 5-step process with rework
ii. Batch Chemical Manufacturing with Rework
e. Challenges of RTY
i. First Pass Yield is the long-standing incumbent standard
ii. Math more difficult and not easily proven as valid without greater familiarity with statistics
iii. In-Process Yields may not be captured
iv. Defects per Unit are often not captured. Most pieces are rejected after the first defect is spotted, and no more time is spent on the pieces.
IV Other defect measures: DPPM vs. DPMO
a. DPPM - Defective Parts Per Million - Count of defective parts per 1 million parts produced. One defective part may have multiple defects, but only counts once.
b. DPMO – Defects Per Million Opportunities – Count of total defects per million opportunities to create a defect.
i. Benefits: More precise DMAIC project metric. Analysis can allow honing in on a key process step or specific opportunity associated with a high defect rate.
ii. Challenges: Defining what constitutes an “opportunity” can be difficult, and not universally recognized.
V Final Note: DMAIC projects should focus on eliminating individual defects. As such, DPU and DPMO make better DMAIC project metrics. This allows the project to scope down and focus on the key defects. The reality of DMAIC projects is that the number of defects can be reduced without actually decreasing the count of defective parts. Project success may not register an impact on business metrics. If DPU is not well measured at the start, the most obvious defects can be replaced by other defects that were previously undocumented. A sampling plan may be necessary in these cases at the start of a project to capture these data.