Data & Analytics

Group 6 employee_attrition

Description
1. Predicting Employee Attrition ã Bhavin Shah ã Krishan Dadlani ã Mrityunjay Gunjiyal ã Nirali Shah ã Yamini Toki ã Yash Gupta How it will help if we know…
Published
of 17
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  • 1. Predicting Employee Attrition • Bhavin Shah • Krishan Dadlani • Mrityunjay Gunjiyal • Nirali Shah • Yamini Toki • Yash Gupta How it will help if we know the top reasons of attrition? Group 6
  • 2. Agenda 1. What is employee attrition? 2. Why do we care? 3. Objective and business value 4. Analysis 5. Modeling results 6. Take away
  • 3. Employee Attrition • Attrition in business can mean the reduction in staff and employees in a company through normal means. • It is imperative for every organization to understand factors driving the employees to look for better work opportunities and provide better satisfaction in terms of salary and kind of work.
  • 4. Why do we care? Cost: Replacing an employee costs as much as 210% of the employee’s annual salary Talent Retention: Since executive and C-level employees are especially hard to replace, these roles typically take longer to fill
  • 5. Objectives and business value • Predict drivers of employee attrition. • Predict potential cases where employee might leave the company. • The factors analyzed in the study will help the organization to identify the weak areas that need attention in order to improve employee satisfaction and retain valuable employees. • Thus, save business costs that range from quantifiable numbers to hidden costs.
  • 6. • Employee Number: Unique id of employee • Age: Age of employee • Distance from home: Approximate distance of workplace from home • Education: Employee’s education degree • Environment Satisfaction: Satisfaction from the surrounding • Gender: Male or Female • Hourly rate • Job Involvement: Employee’s involvement in the job • Job level: Level of job performed • Job role: Designation • Job satisfaction: Satisfaction from job • Martial Status: Married or Single • Years with current manager: Years completed working with current manager Target Variable: Attrition • Monthly Income: Employee’s monthly income • Number of companies worked • Overtime: Expressed as ‘yes’ or ‘no’. • Percent Salary hike: Increment factor in salary • Performance rating: Rating ranging between 3 - 5 • Relationship satisfaction: Expressed in terms of number. • Total working years: Work experience • Years at company: Years completed serving the current company • Years in current role: Years completed working in current role • Years since last promotion: Number of years completed after last promotion • Is Promotion Due: Recoded Field -Has the employee ever been promoted after joining this company. Predictors and Target Variable
  • 7. Education 1 'Below College' 2 'College' 3 'Bachelor' 4 'Master' 5 'Doctor' Environment Satisfaction 1 'Low' 2 'Medium' 3 'High' 4 'Very High' Job Involvement 1 'Low' 2 'Medium' 3 'High' 4 'Very High' Work Life Balance 1 'Bad' 2 'Good' 3 'Better' 4 'Best' Relationship satisfaction 1 'Low' 2 'Medium' 3 'High' 4 'Very High' Performance Rating D 'Low' C 'Good' B 'Excellent' A 'Outstanding' Job Satisfaction 1 'Low' 2 'Medium' 3 'High' 4 'Very High' Stock Level Option 1 'No stock options' 2 ‘Fixed Stock Option' 3 'Based on Job Position' 4 'Based on Performance' Data Definition
  • 8. Variables Data Transformation Years At Company Recoded from the date field ‘Date of Joining’ where, Years At Company = 14 - RIGHT(A2, 2) Years Since Last Promotion Recoded from the date field ‘Date of Last Promotion’ where, Years Since Last Promotion = 14 - RIGHT(A3, 2) Years Promotion Due Since Joining Recoded from ‘Years At Current Company’ and ‘Years Since Last Promotion’ Recoding has been done as follows: =(Years At Current Company – Years Since Last Promotion) Is Promotion Due Recoded from ‘Years At Current Company’ and ‘Years Since Last Promotion’ Recoding has been done as follows: =IF(Years At Current Company – Years Since Last Promotion = 0, "N", "Y") Data Re-Coding
  • 9. Impact of Attrition False Negative: When an employee has not been predicted to leave by the model, but still leaves the company. For this unexpected case of attrition the company will have to bear the following costs: ● Hiring Costs of new Employee ● Training cost of new employee ● Productivity issues during the learning curve Training and hiring costs usually depend on the Job Level of employees. Typically these costs are 150% of the annual salary of the employee. Weighted average cost to company of an Employee leaving = ∑ (Attrition rate * Cost) = approx. $ 76,000 Job Level Attrition Rate Average Salary at this level Average cost of replacement (% in terms of annual salary) Cost of Attrition to company C-level 4/65 = 6.1% $ 228,229 210% $ 479,280 Managers 6/89 = 6.6% $ 192,190 150% $ 288,285 Consultant 14/149 = 9.3% $ 119,880 120% $ 143,856 Mid-level Exec 31/389 = 7.9% $ 64,260 80% $ 51,390 Executive 75/433 = 17.3% $ 32,616 70% $ 22,831
  • 10. Impact of Attrition False Positive: When an employee has been predicted to leave by the model, but was actually not going to leave the company. In this case the company would be bearing the cost of incentives offered: ● Better Stock level options ● Increased pay for overtime or additional hiring of staff to take load off from existing employees ● Increase in pay. An increase in 1 level of Stock option would typically cost the company at least 10% of the annual salary offered. Also an increase of 10% in the workforce will lead to 10% more cost to the company in terms of salary. Similarly a 5 to 10% hike might be required in annual salary to retain an employee. So if an employee is offered any one of the incentives the cost to company would be 10% of their salary.
  • 11. Modeling Result Models AUC FN FP Recall/ Sensitivity Accuracy Precision F1 Forest 0.872 44 64 0.672 0.908 0.584 0.625 Jungle 0.854 36 169 0.731 0.826 0.367 0.489 Boosted Tree 0.875 53 79 0.604 0.888 0.506 0.551 Logistic 0.777 48 227 0.642 0766 0.275 0.385 Neural 0.744 1 928 0.993 0.210 0.126 0.224
  • 12. Decision Tree
  • 13. Significant Predictors • Over Time: In case of Overtime the attrition rate jumps to 22% from 11% and amongst the employees not doing overtime it drops from 11% to just 7%. • Stock Option Level: The employees with Stock Options are less likely to leave based on Tree decision model. Employees with Stock Options 3 and 4 in the company have only a 3% chances of leaving while employees with lesser stock option levels at 1 and 2 have attrition rate of 22% • Monthly Income • Marital Status • Employee Satisfaction • Work Life Balance
  • 14. Data Visualization - Tableau
  • 15. Data Visualization - Tableau 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 0 500 1000 1500 2000 2500 3000 1 2 3 4 STOCK OPTION LEVEL Number of Employess Attrition
  • 16. Data Visualization - Tableau 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 0 200 400 600 800 1000 1200 1400 1600 1800 2000 4 'Very High' 3 'High' 2 'Medium' 1 'Low' ENVIRONMENT SATISFACTION Number of Employess Attrition 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 0 500 1000 1500 2000 2500 3000 3500 4000 4 'Best' 3 'Better' 2 'Good' 1 'Bad' WORK LIFE BALANCE Number of Employess Attrition
  • 17. Thank You Questions?
  • We Need Your Support
    Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

    Thanks to everyone for your continued support.

    No, Thanks
    SAVE OUR EARTH

    We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

    More details...

    Sign Now!

    We are very appreciated for your Prompt Action!

    x