11 Appendix 4
11.1 How Were the Main Areas of Focus Selected?
11.2 Figures 3 and 4
Figures 3 and 4 each show the correlation of Intent to Stay with each of the other constructs measured by the survey. A correlation matrix is shown because the concept of correlation is generally understood by business audiences. However, a correlation matrix does not prove that low ratings of the highly correlated constructs cause turnover. For that, Multiple Regression and Relative Weights Analysis were performed.
The strongest predictors of intent to stay for Operations and Sales are the strongest correlated constructs as shown in figured 3 and 4 but the order of the strongest is slightly different after performing the more advanced analyses. However, since the practical differences in weights is neglible, the correlation matrix is the best tool to use for communication.
11.3 Why Aren’t We Looking at Sales Goals as the Outcome Variable?
At this time, the more pressing concern is high performers leaving; not only in the Sales department but also Operations. Multivariate RWA was run, and the top weighted predictors did not change much. Therefore, we can still increase sales overall by focusing on these predictor variables. In the future, we can focus solely on increasing sales for the lower performers, but right now lets make sure we don’t lose our best people.
11.4 Multiple Regression:
11.4.1 Operations Employees & Intent to Stay
##
## Call:
## lm(formula = Intent_Stay ~ CONF_Mean + FEED_Mean + ALGN_Mean,
## data = ops)
##
## Residuals:
## Min 1Q Median 3Q Max
## -7.6973 -1.1249 0.3866 1.6074 5.1798
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -7.4904 0.7731 -9.689 < 2e-16 ***
## CONF_Mean 1.8781 0.2385 7.874 1.86e-14 ***
## FEED_Mean -0.1678 0.2749 -0.610 0.542
## ALGN_Mean 1.5105 0.3237 4.667 3.85e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.465 on 547 degrees of freedom
## (12 observations deleted due to missingness)
## Multiple R-squared: 0.4239, Adjusted R-squared: 0.4207
## F-statistic: 134.2 on 3 and 547 DF, p-value: < 2.2e-16##
## Call:
## lm(formula = Intent_Stay ~ CONF_Mean + FEED_Mean + ALGN_Mean,
## data = ops)
##
## Standardized Coefficients::
## (Intercept) CONF_Mean FEED_Mean ALGN_Mean
## 0.00000000 0.44622761 -0.03891434 0.27939500The dependent variable is Intent to Stay for Operations and the independent variables are Confidence in Direction, Manager Feedback, and Alignment of Priorities. These three variables account for 42% (p < .05) of variance in Intent to Stay.
Confidence in Direction and Alignment of Priorities are both significant predictors of Intent to Stay.
Manager Feedback is not significant but was included in the model because of its impact on the Sales department.
11.4.2 Sales Employees & Intent to Stay
##
## Call:
## lm(formula = Intent_Stay ~ CONF_Mean + FEED_Mean + ALGN_Mean,
## data = sales)
##
## Residuals:
## Min 1Q Median 3Q Max
## -8.2093 -1.2811 0.5548 1.7902 4.2824
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -4.8394 0.7449 -6.497 3.67e-10 ***
## CONF_Mean 0.8278 0.3385 2.446 0.015065 *
## FEED_Mean 1.1144 0.2915 3.823 0.000162 ***
## ALGN_Mean 1.0003 0.3809 2.626 0.009108 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.197 on 284 degrees of freedom
## Multiple R-squared: 0.4556, Adjusted R-squared: 0.4499
## F-statistic: 79.23 on 3 and 284 DF, p-value: < 2.2e-16##
## Call:
## lm(formula = Intent_Stay ~ CONF_Mean + FEED_Mean + ALGN_Mean,
## data = sales)
##
## Standardized Coefficients::
## (Intercept) CONF_Mean FEED_Mean ALGN_Mean
## 0.0000000 0.2096819 0.2718424 0.2476441The dependent variable is Intent to Stay for Sales and the independent variables are Confidence in Direction, Manager Feedback, and Alignment of Priorities. These three variables account for 46% (p < .05) of variance in Intent to Stay.
All three variables are significant predictors of Intent to Stay.
11.4.3 Sales Employees & Sales Revenue
##
## Call:
## lm(formula = Sales_vs_PlanQ219 ~ CONF_Mean + CWSP_Mean + FEED_Mean +
## ALGN_Mean + DEV_Mean + Engagement_Mean + Perform_Rating,
## data = sales)
##
## Residuals:
## Min 1Q Median 3Q Max
## -53344 -17703 -2839 14884 71136
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -390.6 13168.2 -0.030 0.9764
## CONF_Mean -8008.5 4021.8 -1.991 0.0474 *
## CWSP_Mean 14226.7 3369.3 4.223 3.27e-05 ***
## FEED_Mean -2172.7 3500.1 -0.621 0.5353
## ALGN_Mean -5323.5 4557.2 -1.168 0.2437
## DEV_Mean -3074.4 2237.0 -1.374 0.1704
## Engagement_Mean 1003.8 3773.4 0.266 0.7904
## Perform_Rating 7514.0 940.6 7.988 3.58e-14 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 25290 on 280 degrees of freedom
## Multiple R-squared: 0.3015, Adjusted R-squared: 0.284
## F-statistic: 17.27 on 7 and 280 DF, p-value: < 2.2e-16##
## Call:
## lm(formula = Sales_vs_PlanQ219 ~ CONF_Mean + CWSP_Mean + FEED_Mean +
## ALGN_Mean + DEV_Mean + Engagement_Mean + Perform_Rating,
## data = sales)
##
## Standardized Coefficients::
## (Intercept) CONF_Mean CWSP_Mean FEED_Mean
## 0.00000000 -0.20108887 0.29483800 -0.05253431
## ALGN_Mean DEV_Mean Engagement_Mean Perform_Rating
## -0.13063674 -0.08304684 0.01817327 0.41243302The dependent variable is Sales vs Q2 Goal and the independent variables are all surveyed constructs plus performence ratings. These three variables account for 30% (p < .05) of variance in sales revenue.
The only three significant predictors of sales revenue are Confidence in Direction, Coworker Support, and Performance Ratings (which is also an outcome variable).
11.5 Relative Weights Analysis
R-squared For the Model
## [1] 0.4543085The Raw and Rescaled Weights
| Variables | Raw.RelWeight | Rescaled.RelWeight |
|---|---|---|
| CWSP_Mean | 0.04869222 | 10.717874 |
| CONF_Mean | 0.13733973 | 30.230497 |
| FEED_Mean | 0.07567470 | 16.657116 |
| ALGN_Mean | 0.10746875 | 23.655454 |
| DEV_Mean | 0.02397672 | 5.277629 |
| Engagement_Mean | 0.03461320 | 7.618875 |
| Perform_Rating | 0.02654323 | 5.842556 |
The above output shows RWA analysis for Operations with all constructs still included. This shows that Confidence in Direction, Alignment of Priorities, and Manager Feedback were the strongest drivers of Intent to Stay, accounting for 70.5% of the 45% of variance that all these constructs contribute to Intent to Stay.
11.5.1 Sales
R-squared For the Model
## [1] 0.511The Raw and Rescaled Weights
| Variables | Raw.RelWeight | Rescaled.RelWeight |
|---|---|---|
| CWSP_Mean | 0.05620720 | 10.999774 |
| CONF_Mean | 0.09110371 | 17.829035 |
| FEED_Mean | 0.10631215 | 20.805333 |
| ALGN_Mean | 0.09675283 | 18.934570 |
| DEV_Mean | 0.02727298 | 5.337335 |
| Engagement_Mean | 0.10342528 | 20.240371 |
| Perform_Rating | 0.02991093 | 5.853582 |
The above output shows RWA analysis for Sales with all constructs still included. This shows that Manager Feedback, Alignment of Priorities, and Confidence in Direction were the strongest drivers of Intent to Stay, accounting for 57.58% of the 51% of variance that all these constructs contribute to Intent to Stay.
11.5.2 Why aren’t we focusing on engagement?
Engagement should really be considered an outcome variable as its highly correlated with all of these constructs. By focusing on the other major concerns, we will also increase engagement.
11.5.3 Sales Multivariate Analysis (Intent to Stay with Sales Revenue)
Multivariate R-squared
## [1] 0.3747646The Raw and Rescaled Weights
| Variables | Raw.RelWeight | Rescaled.RelWeight |
|---|---|---|
| CWSP_Mean | 0.04596070 | 12.263884 |
| CONF_Mean | 0.05429577 | 14.487966 |
| FEED_Mean | 0.05617254 | 14.988752 |
| ALGN_Mean | 0.05476716 | 14.613747 |
| DEV_Mean | 0.01597801 | 4.263478 |
| Engagement_Mean | 0.05420647 | 14.464137 |
| Perform_Rating | 0.09338399 | 24.918036 |
The above output shows Multivariate analysis for Sales where the dependent variables are Intent to Stay and Sales Revenue; with all constructs still included. This shows that Manager Feedback, Alignment of Priorities, and Confidence in Direction were still the strongest drivers of Intent to Stay and sales revenue (engagement and performance excluded), accounting for 44.09% of the 38% of variance that all these constructs contribute to Intent to Stay/Sales Revenue. Therefore, we do not think it makes since to initiate interventions specifically to increase individual sales revenue at this time. Our proposed interventions focused on Intent to Stay will still have a positive impact on sales revenue.