Page 174 - Six Sigma Demystified

Chapter 7 i m p r o v e S tag e 155

j. Column J contains the order completion time; for example, row 2 con-
tains the expression “=E2+i2.”
k. Column K contains the order processing day. Row 1 contains the
value “1”; row 2 contains the expression “=iF(J2>480,K2+1,iF(D3>K2,K2
+1,K2)),” which increments the day count when the order completion
time in column J exceeds 480 minutes or the order arrival day has been
incremented.
l. Column L contains the order wait time (exclusive of that portion of
waiting owing to incoming calls, which is included in the regression
model), calculated as the difference between the order start time and
arrival time. Row 2 contains the expression “=iF(K2>D2,E2+(480-C2),
E2-C2).”
m. Column M contains the order cycle time, calculated as the sum of the
process time and the wait time. Row 2 contains the expression “=i2+L2.”
4. The simulation provided the following information, based on 5,000 simu-
lated trials:
a. Using the model developed in the designed experiment, the team
verified that the simulation provided a reasonable estimate of the origi-
nal baseline data. Since the baseline data did not include the two main
factors, the team reviewed a subset of the data and, using the records of
the actual orders, noted the number of line items in the order and esti-
mated the call arrival rate using the phone records for the particular date
and time of the order. The model approximated the data fairly well,
based on a residuals analysis. The average process time, including the
wait time associated with incoming calls, was 36 minutes, of which the
actual processing time (excluding the incoming call wait time) was 22
minutes. all orders were processed on the same day, averaging 72 min-

utes of total cycle time, including 35 minutes of wait time not associated
with incoming calls. although this model for the as- is process will not be
used to estimate the processing time for the suggested process, it will be
used as a baseline to estimate the improvement potential for the sug-
gested process.
b. Using the design of experiments (DOE) model, the impact of vari-
ous changes in the process inputs can be estimated relative to the
baseline:
(1) if the order arrival rate increased by 50 percent (from 1 order per
hour to 1.5 orders per hour), the total order cycle time increased 93 per-
cent, with 2 percent of the orders exceeding eight hours.