Introduction

This project report aims to examine the impact of a new retail store layout and signage design on sales. The trial was implemented in half of the stores at random, and sales data was collected for a whole period before and after the trial was implemented. To determine the layout and signage design's impact on sales, we used null hypothesis testing and the estimation approach. Additionally, we controlled for 3 different types of outlet store types and staff turnover in order to more accurately measure the effects of the trial.

Did the new store layout and signage change average sales?

Our analysis used t-tests and the estimation approach to assess the difference in sales gain for stores participating in the trial and stores not participating in the trial. The results of the t-test show that there is a statistically significant difference between the sales gain of stores participating in the trial and stores not participating in the trial. The mean sales gain for stores not participating in the trial was 18,642 GBP, while the mean sales gain for stores participating in the trial was 426,893 GBP. This represents a difference of 408,250 GBP in favor of stores participating in the trial. Therefore, it can be concluded that changing the retail stores' layout and signage design had a positive impact on store sales.

Was there a different effect in different outlet types, and does adding staff turnover as a predictor improve the model?

Using a t-test on a linear model of the percent sales gain as a function of stores in trial and the outlet type, it was found that percent sales gain differs by outlet type, t(2, 533) = 57.934, p \< 2.2e -16. The confidence intervals for the stores which are not intrial overlap zero, indicating no change in percent sales gain. For the stores which are intrial, 'community_convenience' and 'superstore' experienced an increase in percent sales gain, while 'city_centre_convenience' saw a decline in percent sales gain. This suggests that outlet type has a significant effect on the impact on sales gain.

Comparing models with and without 'staff_turnover' using ANOVA reveals it does not improve the model significantly, F(533) = 0.2894, p < 0.5908. The adjusted r-squares are also very similar for the two models. Furthermore, the effects of 'staff_turnover' using the estimation approach show that 'staff_turnover' does not influence the sales gain at different outlet types.

Results

This project report aimed to determine the impact of a new retail store layout and signage design on sales. Our analysis used t-tests and estimation approach to assess the difference in sales gain for stores participating in the trial and stores not participating in the trial. The results showed that the mean sales gain for stores participating in the trial was significantly greater than for stores not participating in the trial, with a difference of 408,250 GBP in favor of stores participating in the trial. Furthermore, it was found that outlet type has a significant effect on the impact on sales gain. Additionally, it was concluded that 'staff_turnover' does not significantly improve the model.

GitHub Repository