AD. Test hypothesis

Are there more services pushed if advisor generated maintenance menu (clicked button MENU - not at the end of the process)?

Data conditions

• period: 120 days, approximately from 2017/01/10 to 2017/05/10
• analyzed sample has size 5000 and obtained randomly from uniform distribution
• data source: Application Insigts, PROD2, ad.dealer-fx.com

Analysis dependency MENU and total count of DRS+FRS submited

ANOVA

av = aov(drs_plus_frs ~ menu, dataset)
summary(av)

              Df Sum Sq Mean Sq F value Pr(>F)  
menu           1    717   717.2   5.617 0.0178 *
Residuals   4998 638170   127.7                 
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Analysis dependency MENU and total sum of DRS+FRS submited

ANOVA

av = aov(drs_plus_frs_sum ~ menu, dataset)
summary(av)

              Df   Sum Sq Mean Sq F value  Pr(>F)    
menu           1   527238  527238   48.17 4.4e-12 ***
Residuals   4998 54699150   10944                    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Conclusion

• Generating MENU affects the number of DRS+FRS submited
• The average number of pushed services increases ~ two times if MENU genereted
• In 2% of cases MENU is generating
• The average total price of FRS+FRS if MENU generated is about $128
• The average total price of FRS+FRS if MENU not generated is about $46
• There are many outliers of total sums if menu was not generated.