implementing six sigma: are you getting results fast enough?

Post on 07-Jul-2015

250 Views

Category:

Education

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Implementing Six Sigma: Are you Getting Results Fast Enough?

JMP, A Business Unit of SAS

PresenterChuck Boiler- Analytical Strategist

WCBF – Dallas, TXFebruary 15 – 17 2005

28 years

Wor

ldw

ide

Rev

enue

$1B

SAS - 28 Years of Growth

JMP, A Business Unit of SAS

- Founded by John Sall, Executive Vice President of SAS and Co-Founder of SAS Institute

- JMP founded in 1989

- Providing analytics to Engineers in Manufacturing with tools to rapidly impact the bottom line

Recursive Partition: Growing a Tree without digging a hole

Figure 6: Machining data with bad cuts highlighted

Figure 7: Machining data with good cuts highlighted

Figure 8: recursive partitioning windows before analysis

Figure 9: Predicting good and bad cuts with three splits

Figure 10: Predicting good and bad cuts with four rules

Figure 11: Predict lubricant amount

Figure One: Graphs that Live

Survey data distribution Figure 2

Survey data distribution Figure 3

Survey data Table 1, Table 2 E

ase

Resp

on

siv

e

Tim

ely

Pro

fess

ion

al

Qu

ali

ty

Th

oro

ug

h

Co

llab

ora

tio

n

Ex

pecta

tio

ns

Valu

e

Co

mm

un

icati

on

Ease 3 34 34 34 34 34 345 34 34 345

Responsive 34 3 34 34 34 34 345 34 34 345

Timely 34 34 3 34 34 34 345 34 34 345

Professional 34 34 34 3 34 34 45 34 34 345

Quality 34 34 34 34 3 34 345 34 34 345

Thorough 34 34 34 3 3 3 345 34 3 345 Collaboration 03 03 03 03 03 03 3 34 03 034 Expectations 034 034 3 30 03 034 345 3 3 345 Value 345 34 34 34 34 34 345 34 3 34

5 Communication 034 034 3 03 034 034 34 03 03 3

3 When respondent chose 3 on question shown on column… 03 Chose 0 and 3 almost equally 34 Chose mostly 3, sometimes 4 34 Chose 3 and 4 equally 034 Chose mostly 3, sometimes 0 or 4 345 Chose mostly 3, sometimes 4 or 5 03 Chose mostly 0, sometimes 3 345 Chose 3, 4, and 5 almost equally 345 Chose 3 and 4 almost equally, sometimes 5

Survey data Subsetting Figure 4

Survey data subset Figure 5

Designed to save money

Table 3, 4: Experimentation costs

Resource Cost Workers time $500 Materials $500 Down time/Opportunity costs $5,000 Total Cost per run $6,000

Number of Runs

Total Experiment Costs

1 $6,000 10 $60,000 50 $300,000 100 $600,000 150 $900,000

Experimental Cost

Total Experimentation costs

Figure 12: Response surface (weld, time)

Figure 13: Custom design, four hot-tool temps

Figure 14: Custom Design 8 Hot tool temps

Table 5: Costs of Experiments

Experiment

Runs Cost

($6,000 per run)

Cost ($50,000 per run)

Response Surface (weld, time) 8 $48,000 $400,000 Response Surface replicated 4 hot-tool temps

32 $192,000 $1,600,000

Response surface replicated 8 hot-tool temps

64 $384,000 $3,200,000

Custom Design, 4 hot-tool temps, minimum runs

15 $90,000 $750,000

Custom Design, 8 hot-tool temps, minimum runs

27 $162,000 $1,350,000

Custom Design, 8 hot-tool temps, user specific runs

42 $250,000 $2,100,000

Conclusion

Implementing Six Sigma: Are you Getting Results Fast Enough?

JMP, A Business Unit of SAS

PresenterChuck Boiler- Analytical Strategist

WCBF – Dallas, TXFebruary 15 – 17 2005

top related