McDonald’s Case Study

Measure
Select CTQ Characteristic
You are the McDonald’s Green Belt Project team for this
area. In an effort to increase customer satisfaction and
capture more market share, you performed customer surveys
and a QFD analysis. Through your thorough analysis, you
found service time to be a major CTQ. You have also broken
service time into two parts:
1. wait time in line
2. order time (time from when you begin to order your food
until you are given your food and change)
 Measure
Data Collection
• Your first efforts will focus on order time. You have a hunch
that order time may differ for drive-thru versus counter
service and for ordering by numbers versus off menu. Two
classes of Six Sigma trainees were sent out for lunch, half to
the Miami Lakes McDonald’s and half to Miami Gardens.
About half of the people were asked to order at the drive-thru
and half at the counter. The following data was collected:
Order Time [seconds]
Location [1 Miami Lakes, 2 Miami Gardens]
Service [1 Drive-thru, 2 Counter]
Order Method [1 By Number, 2 Off Menu]
 Measure
Define Performance
Standards
• To determine what the performance standard
should be, you surveyed 100 customers, asking
them: “To ensure repeat business, how much
time are you willing to spend from when you
start to order your food until you are given your
food and change?” Based on these results, you
determined the performance standard for order
time to be 90 seconds.
 Analyze

• In your analysis of the McDonald’s data,

you will
– compute the current order time process
capability
– determine which factors (location, service,
order method) have a significant affect on the
output (order time)
 Instructions
• Minitab file: Mcd_st4.mtw contains the McDonald’s data.
• On the following page you will find a list of all the tools we have covered.
You will use these tools to perform the analysis.
• Before starting the actual analysis of the McDonald’s data, take some
time as a team to determine the objective, hypotheses (when applicable)
and potential conclusions for each of the statistical tools.
• At the end of this section, you will find questions about the McDonald’s
data.
• Answer the questions assigned to you.
• Hint for comparing the data subgrouped by 1 factor: unstack it. For
instance, unstacking by location will result in two columns, one for Miami
Lakes and one for Miami Gardens. This enables you to do the Run
Charts.
• Present your findings with a Power Point Presentation.
 Questions
for Complete Data Set (All Teams)

MINITAB FILE Mcd_4.mtw: explanation of subscripts

Location Service Method
1 Miami Lakes Drive Thru By Number
2 Miami Gardens Counter Off menu
1. Visualize the distribution of the data. Is the data normally distributed? NO

2. What is the order completion process capability, given an

upper specification limit of 90 seconds?
 Graphical Summary of service ti
Summary Report

Distribution of Data
Examine the center, shape, and variability.
Descriptive Statistics
N 54
Mean 66.019
StDev 29.324
Minimum 28
5th percentile 31.75
25th percentile 44.75
Median 60
75th percentile 80
95th percentile 137.5
Maximum 167
0 30 60 90 120 150

95% Confidence Intervals

Mean (58.015, 74.022)
Median (50.357, 69.643)
StDev (24.651, 36.201)

Data in Worksheet Order Normality Test

Investigate any outliers (marked in red).
Decision Fail
P-value <0.005
150
service ti

100

50

0 10 20 30 40 50 60
 Capability Analysis for service time
Summary Report

Customer Requirements
How capable is the process?
Upper spec 90
0 6
Target *
Z potential = 1.07 Lower spec *
Low High
Z actual = 0.82 Process Characterization
Mean 66.019
Standard deviation (overall) 29.324

Actual (overall) capability

Pp *
Actual (Overall) Capability Ppk 0.27
Are the data below the limit? Z.Bench 0.82
USL
% Out of spec 20.67
PPM (DPMO) 206734

Comments

• The defect rate is 20.67%, which estimates the percentage of

parts from the process that are outside the spec limits.

Actual (overall) capability is what the customer experiences.

Potential (within) capability is what could be achieved if process

shifts and drifts were eliminated.

0 30 60 90 120 150
 Analyze the Data to determine if Location is an
X
Miami Lakes vs. Miami Gardens

3. Graph the differences between median service time and service time variation for the x’s. Slide
11

4. Graph the differences between average service time and service time variation for the x’s.
Slide 12

5. Are there any non-random patterns over time? (2 charts) Slide 13 & 14. Gardens is out of
Control!!!

6. Is the data normally distributed for the Y? 15 & 16. Yes to Lakes, No to Gardens.

7. Do the two groups have the same consistency (st dev) in speed of service? No. Gardens is
significantly higher. Slide 11.

8. Is there a statistically significant difference between the average speed of service? No. Slide
17.
 Boxplot of Lakes, Gardens
180

160

140

120
Data

100

80

60

40

20
Lakes Gardens
 Boxplot of Lakes, Gardens
180

160

140

120
Data

100

80

60

40

20
Lakes Gardens
 I-MR Chart of Lakes
Summary Report
Is the process mean stable? Comments
Evaluate the % of out-of-control points.
The process mean is stable. No data points are out of control on the I
0% > 5% chart.

Yes No
0.0%

Individual and Moving Range Charts

Investigate any out-of-control points.
UCL=114.7
100
Individual Value

_
X=63.3
50

LCL=11.8
0

UCL=63.21

50
Moving Range

25 __
MR=19.35

0 LCL=0
1 3 5 7 9 11 13 15 17 19 21 23

N: 24 Mean: 63.25 StDev(within): 17.152 StDev(overall): 18.878

Control limits are estimated using the StDev(within).
 I-MR Chart of Gardens
Summary Report
Is the process mean stable? Comments
Evaluate the % of out-of-control points.
The process mean may not be stable. 2 (6.7%) data points are out of
0% > 5% control on the I chart. Keep in mind that you may see 0.7%
out-of-control points by chance, even when the process is stable.

Yes No
6.7%

Individual and Moving Range Charts

Investigate any out-of-control points.
160
UCL=144.9
Individual Value

80 _
X=68.2

0 LCL=-8.4
Moving Range

100 UCL=94.2

50
__
MR=28.8

0 LCL=0
1 4 7 10 13 16 19 22 25 28

N: 30 Mean: 68.233 StDev(within): 25.556 StDev(overall): 35.742

Control limits are estimated using the StDev(within).
 Graphical Summary of Lakes
Summary Report

Distribution of Data
Examine the center, shape, and variability.
Descriptive Statistics
N 24
Mean 63.25
StDev 18.878
Minimum 39
5th percentile 39
25th percentile 45
Median 61.5
75th percentile 73.25
95th percentile 106
Maximum 108

20 40 60 80 100
95% Confidence Intervals
Mean (55.279, 71.221)
Median (49.134, 70.173)
StDev (14.672, 26.481)

Data in Worksheet Order Normality Test

Investigate any outliers (marked in red).
Decision Pass
P-value 0.278
100
Lakes

75

50

0 5 10 15 20 25
 Graphical Summary of Gardens
Summary Report

Distribution of Data
Examine the center, shape, and variability.
Descriptive Statistics
N 30
Mean 68.233
StDev 35.742
Minimum 28
5th percentile 29.65
25th percentile 37.75
Median 59
75th percentile 82.5
95th percentile 156.55
Maximum 167

0 40 80 120 160
95% Confidence Intervals
Mean (54.887, 81.580)
Median (49.229, 79.543)
StDev (28.465, 48.049)

Data in Worksheet Order Normality Test

Investigate any outliers (marked in red).
Decision Fail
P-value <0.005
150
Gardens

100

50

0 5 10 15 20 25 30
 2-Sample t Test for the Mean of Lakes and Gardens
Summary Report

Do the means differ? Individual Samples

0 0.05 0.1 > 0.5 Statistics Lakes Gardens
Sample size 24 30
Yes No Mean 63.25 68.233
P = 0.514 95% CI (55.28, 71.22) (54.887, 81.580)
Standard deviation 18.878 35.742
The mean of Lakes is not significantly different from the mean of
Gardens (p > 0.05).
Difference Between Samples
Statistics *Difference
Difference -4.9833
95% CI for the Difference
Is the entire interval above or below zero? 95% CI (-20.247, 10.280)
*Difference = Lakes - Gardens

Comments

-20 -10 0 10 • Test: There is not enough evidence to conclude that the means
differ at the 0.05 level of significance.
• CI: Quantifies the uncertainty associated with estimating the
difference in means from sample data. You can be 95% confident
that the true difference is between -20.247 and 10.280.
Distribution of Data
Compare the data and means of the samples. • Distribution of Data: Compare the location and means of samples.
Look for unusual data before interpreting the results of the test.
Lakes

Gardens

20 40 60 80 100 120 140 160

 Questions for
Comparing All 8 Subgroups

9. Graph the differences between average service time and between

service time variation. Slide 19.

10. Is the data normally distributed for each subgroup? No, 1 is a fail.
Slide 20.

11. Do the eight groups have the same consistency (st dev) in speed of
service? No, 1 differs from 8 slide 21.

12. As a corporation, McDonald’s has a goal to provide, on average,

the same speed of service to customers regardless of location,
service type or order method. Based on this sample, does
McDonald’s reach that goal? No!
 Histogram of service time
Normal
0. 06 sub grp
111
112
0. 05 121
122
211
0. 04 212
221
Density

222
0. 03
Mean StDev N
49.22 10.88 9
83.33 24.03 3
0. 02 62.67 11.09 9
87 14.73 3
72.5 14.25 6
0. 01 107.3 36.64 9
38.57 6.949 7
47 13.98 8
0. 00
30 60 90 120 150 180
service time
 Graphical Summary of service _1, service _2, service _3, service _4,...
Summary Report

Distribution of Data Data in Worksheet Order

Compare center, shape, and variability. Investigate any outliers (marked in red).
service _1 service _2 service _1 service _2
150

100

50

service _3 service _4 service _3 service _4

150
100

50

Y Data
service _5 service _6 service _5 service _6
150

100

50

service _7 service _8 service _7 service _8

150

100

50

30 60 90 120 150 180 30 60 90 120 150 180 0 5 100 5 10

Y Data

Normality Test
Group N Mean 95% CI StDev 95% CI Min Median Max P Decision
service _1 9 49.222 (40.857, 57.588) 10.883 (7.3511, 20.850) 39 45 74 0.013 Fail
service _2 3 83.333 (23.645, 143.02) 24.028 (12.510, 151.01) 60 82 108 0.618 Pass
service _3 9 62.667 (54.142, 71.192) 11.091 (7.4912, 21.247) 39 65 78 0.308 Pass
service _4 3 87 (50.406, 123.59) 14.731 (7.6698, 92.580) 71 90 100 0.505 Pass
service _5 6 72.5 (57.544, 87.456) 14.251 (8.8958, 34.953) 50 79 87 0.129 Pass
service _6 9 107.33 (79.169, 135.50) 36.640 (24.749, 70.194) 58 93 167 0.693 Pass
service _7 7 38.571 (32.145, 44.998) 6.9488 (4.4778, 15.302) 31 37 53 0.064 Pass
service _8 8 47 (35.313, 58.687) 13.980 (9.2429, 28.452) 28 49.5 69 0.714 Pass
 Standard Deviations Test for service ti_1, service ti_2, service ti_3,...
Summary Report

Do the standard deviations differ? Which standard deviations differ?

0 0.05 0.1 > 0.5 # Sample Differs from
1 service ti_7 8
Yes No 2 service ti_1
P < 0.001 3 service ti_3
Differences among the standard deviations are significant (p < 4 service ti_8
5 service ti_5
0.05).
6 service ti_4
7 service ti_2
8 service ti_6 1
Standard Deviations Comparison Chart
Red indicates standard deviations that differ.

service ti_7

service ti_1
Comments

service ti_3 • Test: You can conclude that there are differences among the
standard deviations at the 0.05 level of significance.
• Comparison Chart: Red indicates standard deviations that differ
service ti_8 from each other. Comparison intervals cannot be calculated
because the sample sizes are too small or vary by too much.
service ti_5 Without the comparison intervals, the only standard deviations
you can determine are different are the two that are furthest
apart.
service ti_4

service ti_2

service ti_6

10 20 30 40
 Bonus Question
13. Given your preference for a combination of service type
and order method, would you go to the Miami Lakes or
Miami Gardens McDonald’s?
(meaning I give you the last two numbers I.e 11, and
you must tell me if you’ll go to 111 or 211)
2-sample t-test

Drive Thru, By Number Yep, better head on to Miami

Lakes
 2-Sample t Test for the Mean of service ti_1 and service ti_2
Summary Report

Do the means differ? Individual Samples

0 0.05 0.1 > 0.5 Statistics service ti_1 service ti_2
Sample size 9 6
Yes No Mean 49.222 72.5
P = 0.009 95% CI (40.86, 57.59) (57.544, 87.456)
Standard deviation 10.883 14.251
The mean of service ti_1 is significantly different from the mean of
service ti_2 (p < 0.05).
Difference Between Samples
Statistics *Difference
Difference -23.278
95% CI for the Difference
Is the entire interval above or below zero? 95% CI (-39.089, -7.4668)
*Difference = service ti_1 - service ti_2

Comments

-40 -30 -20 -10 0 • Test: You can conclude that the means differ at the 0.05 level of
significance.
• CI: Quantifies the uncertainty associated with estimating the
difference in means from sample data. You can be 95% confident
that the true difference is between -39.089 and -7.4668.
Distribution of Data
Compare the data and means of the samples. • Distribution of Data: Compare the location and means of samples.
Look for unusual data before interpreting the results of the test.
service ti_1

service ti_2

40 60 80