Data Mining and Predictive Modeling for HR Data Mining Techniques for Analyzing Voluntary Turnover. Jason Noriega

Transcription

1 Data Mining and Predictive Modeling for HR Data Mining Techniques for Analyzing Voluntary Turnover Jason Noriega 1

2 Agenda Introduction Background and experience. What is Data Mining? High level overview. Decision Tree Examples Overview of the decision tree technique, and how I have applied it in different ways. Final Thoughts Lessons learned. 2

3 Agenda Introduction Background and experience. What is Data Mining? High level overview. Decision Tree Examples Overview of the decision tree technique, and how I have applied it in different ways. Final Thoughts Lessons learned. 3

4 Introduction Jason Noriega SanDisk Mobile Computing Senior Workforce Analyst at SanDisk Consumer Experience: ebay, Lawrence Livermore National Laboratory, NASA, and National Geospatial Intelligence Agency. 4

5 Agenda Introduction Background and experience. What is Data Mining? High level overview. Decision Tree Examples Overview of the decision tree technique, and how I have applied it in different ways. Final Thoughts Lessons learned. 5

6 What is Data Mining? Data mining is a field that utilizes methods of machine learning, traditional statistics, database technology, and data visualization. Explores Your Data Finds Patterns Performs Predictions Typical Uses Targeted Customer Retention WF Analytics Targeted Employee Retention Data Mining Predict 12 Months of Sales of a New Product Predict 12 Months of Employee Separations Building Effective Marketing Campaigns Building Effective Recruiting Initiatives 6

7 High Level Process Historical Data Predictive Model R Program Weka SAS EG SAS JMP SPSS Modeler Predictive Model Data with Predictions 7

8 Agenda Introduction Background and experience. What is Data Mining? High level overview. Decision Tree Examples Overview of the decision tree technique, and how I have applied it in different ways. Final Thoughts Lessons learned. 8

9 Decision Tree Technique Decision tree algorithms applied to employee attrition data provides you with a laser precision focus on employee retention. Recursive Partitioning Decision Rules - Example Predictors 1. Job Function 2. Gender 3. Ranking 4... Predictors 1. Age 2. Rating 3. Commute 4... Predictors 1. Performance 2. Tenure 3. Commute Leavers Top Performer Stayers Leavers Age < 40 Software Engineer 50% Stayers Leavers Stayers 65% 50% 35% Stayers Leavers 75% 25% 90% 10% Rule 1: 50% Likelihood of Leaving If Job Function = Software Engineer And Age < 40 And Performance = Top Performer Rule 2: 60% Likelihood of Leaving If Manager Effectiveness < 70% Favorable And Pay Grade < 17, And Region = Americas Rule 3: 70% Likelihood of Leaving If Not Promoted in Prior 2 Years = No And Tenure > 3 Years 9

10 Example 1 Improving the Retention of Employees Decision Tree Analysis Where are the high rates of voluntary turnover? Exit Survey Why do these patterns exist? Prioritization Where should we prioritize areas of improvement? Action Planning What can do to reduce high rates of voluntary turnover? 10

11 Example 1 - Variables Variables for Attrition Analysis Demographics/ Tenure Age Gender Tenure Job Content Reasgned Diff. Pers Area Job Function Exempt/Nonexempt Latest Perform. Rating Perform. Rating Delta Manager Flag Location Region Country Location Compensation Pay Grade Comp Ratio Promotion Flag Spot Award Flag Tuition Reimbursement Flag Manager Mgr Tenure Mgr Perform. Attrition Analysis Timeframe June May Model Timeframe for Finding Patterns 11

12 Example 1 - Tree Less than 1.1 yrs Stayed: % Left: % Stayed: % Left: % Stayed: % Left: % Employee Tenure > 7.3 Yrs Stayed: % Left: 250 8% <= > 22 Stayed: % Left: % Stayed: % Left: % Business Unit Organization #1 Organization #2 Stayed: % Left: % <= 85 Stayed: % Left: % Stayed: % Left: % Compa Ratio > 85 Stayed: % Left: % Pay Grade Stayed: % Left: % Stayed: % Left: 50 5% 12

13 Example 1 Survey Linkage Stayed: % Left: % Less than 1.1 yrs Stayed: % Left: % Pay Grade <= Stayed: % Left: % Low Importance/ High Performance Stayed: % Left: % Exiting Employees Stayed: % Left: % High Importance/ High Performance Employee Tenure > 7.3 Yrs Stayed: % Left: 250 8% > 22 Stayed: % Left: % Stayed: % Left: 50 5% Uncompetitive Pay Job matched expectations Opportunity for growth Low Importance/ Low Performance High Importance/ Low Performance 13

14 Example 1 - Tree DNM, Meets Some Stayed: % Left: % Stayed: % Left: % Stayed: % Left: % Stayed: % Left: % Employee Tenure Less than 1.1 yrs yrs > 7.3 Yrs Performance Rating Meets Stayed: % Left: % Stayed: % Left: % Yes Stayed: % Left: 50 5% Exceeds Some Exceeds Most Stayed: % Left: 150 7% Stayed: % Left: 50 7% Promotion No Stayed: % Left: % Big Data Forbes Article -big-data-is-transforming-the-hunt-for-talent/ 14

15 Example 2 Grouping Variables Location, Job, Org, Other Impact of People Mgr Recruitment Other Where does the problem exist? Location/Job/Org Performance Rating Region State Country Work Location Pay Grade Job Function Job Level Business Unit What impact does the mgr have? Manager Impact Overall Engagement Score - Overall Satisfaction - Feel Proud - Feel Motivated - Recommend as Great Place Overall Job Sat. Score - Work feels important purpose - Clearly understand indv. Objectives - Job makes good use of skills - Oppty to learn and grow Overall Mgr Sat. Score - Mgr clearly explains expectations - Mgr provides ongoing coaching - Mgr actively supports prof. dev. - Mgr involve people with decisions - Mgr instills positive attitude - Mgr is someone I trust Mgr Performance Mgr Tenure How can we target those most likely to stay? Recruitment Education Level Total Yrs Experience Total Yrs Relevant Experience Recruitment Source Are there any other important variables? Age Gender Compa Ratio Other 15

16 Location/Job/Org - Perform. Rating - Region - State - Country - Work Location - Pay Grade - Job Function - Job Level - Business Unit Where Does the Problem Exist? Does Not Meet Stayed: % Left: % Stayed: % Left: % Performance Rating Stayed: 84 76% Left: 26 24% Americas Stayed: % Left: % Stayed: % Left: 6 4% Stayed: % Left: % Region Asia Pacific Stayed: 83 87% Left: 12 13% Job Function A, B, and C All Other Meets Some, Meets, Exceeds Some/Most No Rating EMEA Stayed: % Left: 20 7% 16

17 Mgr Impact - Overall Engagement Score - Overall Satisfaction - Feeling Proud - Feeling Motivated - Recommend great place - Overall Job Sat. Score - Work feels important purpose - Clearly understand indv. Obj - Job makes good use of skills - Oppty to learn and grow - Overall Mgr Effectiveness - Mgr clearly explains expectations - Mgr provides ongoing coaching - Mgr actively supports prof. dev. - Mgr involve people with decs - Mgr instills positive attitude - Mgr is someone I trust - Mgr Performance What Impact does the Manager Have? Does Not Meet Stayed: % Left: % Mgr Effectiveness Stayed: % Left: % Performance Rating Stayed: 84 76% Left: 26 24% Americas Stayed: % Left: % Stayed: % Left: 6 4% <= 75% Favorable > 75% Favorable - Mgr Tenure Stayed: % Left: % Stayed: % Left: 60 16% Stayed: % Left: % Region Asia Pacific Stayed: 83 87% Left: 12 13% Job Function A, B, and C All Other Meets Some, Meets, Exceeds Some/Most No Rating EMEA Stayed: % Left: 20 7% 17

18 Example 3 Quarterly Workforce Dashboard with Deep Dive Quarterly Dashboard Deep Dive 25% 20% 15% 10% 5% 0% Voluntary-ORG Involuntary -ORG Voluntary-Companywide 17.1% 19.1% 14.2% 15.4% 12.3% 2013-Q Q Q Q Q2 Predictors 1. Tenure 2. Age 3. Region 4... Predictors 1. Region 2. Rating 3. Pay Grade 4... Predictors 1. Grade 2. Performance 3. Age <= 7 Stayers Leavers India Leavers Leavers 10% 19% <= 2.5 yrs Tenure > 2.5 yrs 50% Stayers Leavers Stayers 65% 50% 35% Grade Stayers Region 75% 25% > 7 90% 81% 18

19 Software R Program for Statistical Computing SPSS Modeler