csi study
TRANSCRIPT
DOES DEALERSHIP CSI
REALLY WORK?
ROB KEALEY
SO WHAT’S THE POINT?
OEMs CLAIM HIGH CSI EQUALS LOW WARRANTY
EXPENSE AND MORE VEHICLE SALES
DEALERS CLAIM NOT TRUE
OBJECTIVE
DETERMINE WHO IS RIGHT
Ho: High Sales CSI does not lead to more sales Ha: High Sales CSI does lead to more sales Ho: High Service CSI does not lead to more vehicle sales Ha: High Service CSI does lead to more vehicle sales Ho: High Service CSI dealers do not have low warranty
expense Ha: High Service CSI dealers do have low warranty
expense
Ho = Null Hypothesis
Ha = Alternative Hypothesis
Sample and methodology
611 Kia dealers sampled. Model Year 2007 vehicle sales and warranty cost
against Sales and Service CSI were measured. The objective was to determine correlation,if any,
between variables.
Sample selection
Histogram
0
5
10
15
20
25
2007 sales
Perc
ent
VEHICLE SALESFULL YEAR 2007
count 611
mean 359.18
sample variance 128,608.17
sample standard deviation 358.62
minimum 8
maximum 2762
range 2754
SALES SATISFACTION SCORES
FULL YEAR 2007
count 611
mean 88.9815
sample variance 59.4137
sample standard deviation 7.7080
minimum 48.31
maximum 100
range 51.69
Histogram
0123456789
10
KPI
Perc
ent
Sample selection
Histogram
0123456789
10
KSI
Perc
ent
SERVICE SATISFACTION SCORES
FULL YEAR 2007
count 611
mean 81.2153
sample variance 109.9525
sample standard deviation 10.4858
minimum 34.81
maximum 100
range 65.19
Sample selection
Histogram
0
2
4
6
8
10
12
14
16
18
SCPVS
Perce
nt
WARRANTY EXPENSEFULL YEAR 2007
count 611
mean 406.9122
sample variance 5,650.5844
sample standard deviation 75.1704
minimum 110.9
maximum 791.41
range 680.51
Sample selection
Ho Null Hypothesis: High Sales CSI does not lead to more sales
sales vs KPI y = 11.256 x - 642.406
R2 = 0.059
-500
0
500
1000
1500
2000
2500
3000
40 50 60 70 80 90 100 110
KPI
2007
sal
es
r² 0.059 n 611 r 0.242 k 1
Std. Error 348.252 Dep. Var. 2007 sales
ANOVA tableSource SS df MS F p-value
Regression 4,591,944.2091 1 4,591,944.2091 37.86 1.38E-09Residual 73,859,037.2606 609 121,279.2073
Total 78,450,981.4697 610
Regression output confidence intervalvariables coefficients std. error t (df=609) p-value 95% lower 95% upperIntercept -642.4064 163.3825 -3.932 .0001 -963.2679 -321.5450
KPI 11.2562 1.8293 6.153 1.38E-09 7.6637 14.8487
r value of 0.242 indicates very small correlation
r value of 1.0 = direct correlation
r value of 0.0 = no correlation
Ho Null Hypothesis: High Sales CSI does not lead to more sales
sales vs KSI y = 10.071 x - 458.765
R2 = 0.087
-500
0
500
1000
1500
2000
2500
3000
20 40 60 80 100 120
KSI
2007
sal
es
Ho Null Hypothesis: High Service CSI does not lead to more vehicle sales
Regression Analysis
r² 0.087 n 611 r 0.294 k 1
Std. Error 342.999 Dep. Var. 2007 sales
ANOVA tableSource SS df MS F p-value
Regression 6,803,170.9052 1 6,803,170.9052 57.83 1.09E-13Residual 71,647,810.5645 609 117,648.2932
Total 78,450,981.4697 610
Regression output confidence intervalvariables coefficients std. error t (df=609) p-value 95% lower 95% upperIntercept -458.7653 108.4544 -4.230 2.70E-05 -671.7553 -245.7752
KSI 10.0714 1.3244 7.604 1.09E-13 7.4704 12.6723
r value of 0.294 indicates very small correlation
Ho Null Hypothesis: High Service CSI does not lead to more vehicle sales
r value of 1.0 = direct correlation
r value of 0.0 = no correlation
warranty cost vs service satisfaction
y = -0.681 x + 462.240
R2 = 0.009
$0
$100
$200
$300
$400
$500
$600
$700
$800
$900
20 40 60 80 100 120
KSI
SC
PV
S
Ho Null Hypothesis: High Service CSI does not lead to low warranty cost
Regression Analysis
r² 0.009 n 611 r -0.095 k 1
Std. Error 74.892 Dep. Var. SCPVS
ANOVA tableSource SS df MS F p-value
Regression 31,127.5614 1 31,127.5614 5.55 .0188Residual 3,415,728.9163 609 5,608.7503
Total 3,446,856.4777 610
Regression output confidence intervalvariables coefficients std. error t (df=609) p-value 95% lower 95% upperIntercept 462.2399 23.6803 19.520 2.86E-66 415.7349 508.7449
KSI -0.6812 0.2892 -2.356 .0188 -1.2492 -0.1133
r value of –0.095 indicates almost no correlation
Ho Null Hypothesis: High Service CSI does not lead to low warranty cost
r value of 1.0 = direct correlation
r value of 0.0 = no correlation
Conclusion
SO WHO IS RIGHT? Ho: High Sales CSI does not lead to more
sales True – correlation is weak at best
Ho: High Service CSI does not lead to more sales True – correlation is weak at best
Ho: High Service CSI dealers do not have low warranty expense True – there is little evidence of correlation
Ho = Null Hypothesis
Ha = Alternative Hypothesis
WHAT DOES IT MEAN?
Dealership vehicle sales volume appears to be driven by factors other than dealership sales or service CSI
Manufacturer incentives to improve dealership CSI scores may be a misguided use of marketing resources.
Conclusion
WHAT DOES IT MEAN?
Warranty expense appears driven by factors other than dealership service CSI.
Dealership pay plans based on CSI may depress employee morale and appear to have little influence on the customers’ purchase decision.
Conclusion
WHAT DOES IT MEAN?
OEM CSI incentives and dealership actions to influence CSI scores may skew results, and diminish the value of the data as an improvement tool.
Conclusion
At least for these 611 Kia dealersAt least for these 611 Kia dealers