- 25th Feb 2024
- 06:03 am
Introduction:
RACV Solar is a wholly-owned subsidiary of the RACV (Royal Automobile Club of Victoria). While the RACV is best known as a provider of car and road-user services, the organization has been active in expanding the range of services it offers to its members, including insurance and travel services.
The purpose of this report is to analyze the 2011 client data and come up with the factors influencing client satisfaction. The data comprises different types of information like the client_id, the salesperson, customer introduction, inverter capacity, etc along with the target variable “Client Satisfaction” on a scale from zero (completely unsatisfied) to 100 (completely satisfied) of the RACV Solar Clients.
The objective is to under the significant variables that impact the client’s satisfaction, and their relationship with the client satisfaction and accordingly take measures to increase/ reduce or optimize such variables.
Section 1 – EXPLANATION OF THE REGRESSION ANALYSIS
After assumption testing of multicollinearity and linear relationship of dependent and independent variables, the multilinear regression model fitting was initiated.
In the 1^{st} iteration of the regression analysis, there were 12 variables taken into consideration for multiple linear regression model fitting. Out of these 12 variables, we started eliminating the variables at each iteration that had a p-value > 0.05 (5% level of significance) and ended up with the 6 significant variables.
Summary Output (Regression Analysis)
The R squares of the above-fitted model is 0.6287 which means 62% of the variation in the data can be explained by the independent variables and the overall model is significant at 5% level of significance.
Provided below is the list of final variables along with their estimated values and p-values for reference:
Variable Name |
Estimate |
P-value |
PV System Size (kW) |
1.20441599 |
3.63221E-08 |
PV System Cost ($) |
-0.0014462 |
5.65942E-20 |
CI_Digital |
-15.29668501 |
5.07473E-30 |
CI_RACV_Store |
-7.6607314 |
5.91695E-07 |
CI_Event |
-14.87281074 |
5.76720E-26 |
Inv_Medium |
5.214742053 |
2.39719E-07 |
Here, all the variables except for PV System Size and Inv_Medium are inversely related to Client Satisfaction. The reduction of these variables would increase client satisfaction and vice versa.
As the p-value of the model is 1.54850022282092E-49 < 0.05, the F-test conducted to test the significance of the overall model rejects the null hypothesis at a 5% level of significance.
Hence, the overall model along with the above 6 variables is significant.
Regression Statistics
Provided below is the description of the summary statistics in the regression output:
Column 1: This column comprises of the list of independent variables along with the intercept. The intercept represents the value of client satisfaction when all independent variables are 0 (In our case, the client satisfaction is 80 if there is no improvisation in any of the 6 independent variables).
- Coefficients: Weight assigned to each of the independent variables. The sign of the sign of the estimated value is to be observed. +ve sign implies a direct relationship with respect to client satisfaction and vice versa. There is no threshold on the estimated value.
- Standard error: It is the standard deviation of the coefficient value that is used during the computation of the t-statistic.
- T stat: This value represents the statistic value of the one-sample t-test applied to each of the independent variables in the model basis their coefficient values. This test determines how significant/insignificant an independent variable individually is.
- P-value: This is the probability value of falling under the rejection criteria of the hypothesis being tested. In our case, p-value < 0.05 for the corresponding independent variable would make it significant.
- Lower 95%, Upper 95% à Confidence intervals of the coefficients.
- Lower 95.0% Upper 95.0% à Prediction intervals of the coefficients.
- ANOVA: This is the F test performed on the entire data to check the significance of the overall model.
- R Squared: The value of R squared represents the amount of variation explained by the independent variables. The higher the R squared, the better the mode.
- Adjusted R squared: This is the R squared value adjusted for every single inclusion/exclusion of an independent variable.
Practical Interpretation (Regression Analysis)
The results obtained from regression output can be further utilized for understanding the impact of each independent variable on client satisfaction and how an increase or decrease of any one of the independent features can influence client satisfaction.
It can be further used for building predictions basis certain values of the independent variable in the future when a new client comes in.
Conclusion:
From the outputs and analysis, it can be concluded that the factors driving the client satisfaction significantly are PV System Size, PV System Cost, Digital Customer Introduction, RACV Store related customer introduction, Event related customer introduction and Medium level of Inverter capacity with an accuracy of approximately 80%.
Section 2 – recommendations TO MAXIMISE CLIENT SATISFACTION
Basis the analysis carried out there are 2 recommendations that can be taken ahead for maximization of client satisfaction concerning RACV Solar’s Promotion with Geelong Sustainability.
Recommendation 1
Targeting a group of customers RACV Solar could pursue that maximizes Client Satisfaction. This group would comprise of the customers not being introduced via Digital, Event, and RACV Store, along with the ones having higher PV System Size and lower PV Cost with more priority to medium level of inverter capacity.
Explanation of Recommendation 1
The usage of the customers being introduced through Digital, RACV Store, and Event is negatively impacting Client Satisfaction. This means, that the customers introduced through the above platform reduce the score of client satisfaction and hence the customers must be introduced through Referral mode only. This is also intuitive as the client is coming through another client who was highly satisfied with RASV Solar Energy's service.
Also, as recommended the higher PV System Size (kW) with lower PV System Cost ($) increases client satisfaction. Th board can look up to a solution that can find a perfect trade-off between the above 2 parameters to maximize client satisfaction.
Lastly, is has been drawn from the analysis that the medium size of inverters, where the PV System Size is greater than 5kW and less than 10kW leads to an increase in client satisfaction. Therefore, more focus and priority should be provided to that category of investors.
Recommendation 2
Exclusion of the customers insignificantly impacts client satisfaction like the ones from the location of Geelong or Surf Coast or having Claris as the sales person and the ones negatively impact client satisfaction like small inverter size.
Explanation of Recommendation 2
The rate of change of client satisfaction has been negligible with respect to the rate of change of the customers belonging to locations like Geelong or Surf Coast. It can be noticed that client satisfaction is least impacted with these 2 locations. Hence, they can be deprioritized and the right audience can be targeted for sales purposes. Having Claris as the salesperson has also not led to any significant impact on client satisfaction according to the tests. Hence, a replacement or more improvement is required at Claris’s end to gauge better client satisfaction. Having a small size of inverters has negatively impacted client satisfaction according to the tests. This means that the client satisfaction reduces with every 1 unit increase of small-sized inverters. Hence, the production of such inverters should be reduced to not impact the overall client satisfaction.
Conclusion:
The overall client satisfaction at RACV Solar is influenced by multiple factors. Prioritizing or deprioritizing the factors on the basis of the tests computed and analysis performed can maximize client satisfaction and provide better results.