Managing Employee Retention

PURPOSE OF ANALYSIS

This analysis aims to:

 Determine the relationship between employee tenure (manager, crew) and store performance

 Understand the impact of additional store factors on store performance, versus employee tenure

 Recommend a human resource strategy to optimise performance of Store24 stores

ANALYSIS APPROACH

1. Select an appropriate measure of store performance (dependent variable)

2. Analyse the relationship between employee tenure and store performance and optimise linearity

3. Define additional factors that could be affecting store performance

4. Determine the relationship between individual factors (independent) and store performance (dependent)

5. Transform individual factors to generate the best linear relationship to store performance

6. Determine collinearity between transformed independent variables

7. Build a linear multi-variable regression to understand which factors most impact store performance

8. Define human resource strategy for Store24 to optimise store performance

1. Select Store Performance Measurement

There are 2 measures of store performance; sales and profit. These two measures have a significant positive

correlation (see Appendix 1), therefore we only need to select one as a dependent variable. For this analysis, we

have selected profit.

2. Relationship between Employee Tenure and Store Profit

Management Tenure had a stronger correlation to store profit (0.438, see Appendix 2), versus Crew Tenure

(0.258, see Appendix 3). However both analyses had residual plots which were not evenly distributed (i.e.

residuals were concentrated against the Y axis), indicating the relationship could be improved. The two

independent variables were therefore transformed using both a square root and a logarithmic method. For

Management Tenure, the best transformation was to take the square root of the independent variable (low

impact on R2 and delivered the most evenly distributed residual plot). For Crew Tenure, the best transformation

was to take the log of the independent variable

...