2. Reactive – Operational Reporting
Operational Reporting for Measurement of Efficiency and Compliance, Data
Exploration and Integration, Development of Data Dictionary
Level 1
Predictive Analytics
Development of Predictive Models, Scenario Planning, Risk
Analysis and Mitigation, Integration with Strategic Planning
Level 4
Strategic Analysis
Segmentation, Statistical Analysis, Development of “People Models”,
Analysis of Dimensions to Understand Cause & Delivery of Actionable
Solutions
Level 3
Proactive – Advanced Reporting
Operational Reporting for Benchmarking and Decision-making,
Multidimensional Analysis and Dashboards
Level 2
Prescriptive Analytics
Machine Learning, Prescribe Recommended Actions,
Specify Interrelated Effects of Decisions
Level 5
Evolution of Analytics
Source: Bersin & AssociatesSource: Bersin & Associates
3.
4. State of the Art Data Analytics System
Ad-Hoc
initial analysis
Expensive
Software
Taxonomy-Building &
Semantic Grouping
Unsupervised
Learning Clustering
Supervised
Learning Optimization
Deep Learning
Algorithms
Natural Language
Processing
Developers
Predictive
Applications
xxx
xxx
Hardware Upgrades
Prescriptive
Applications
Versioning
Machine
Learning Algorithms
Data Cleansing/
Transformation
Multi-Source
Data Integration
Expensive
Software
Data
Scientists
Creating a State of the Art Analytics System
VERY HARD AND EXPENSIVE!
How hard can it be?
Data
Aggregation
Data
Classification
Implementation
of Algorithms
Application
Creation
Continuous
Machine Learning
Application UpdatesMaintenance
5. Taxonomy-Building &
Semantic Grouping
Ad-Hoc
initial analysis
Unsupervised
Learning Clustering
Prescriptive
Applications
Versioning
Data Cleansing/
Transformation
Multi-Source
Data Integration
Maintenance
Expensive
Software
Supervised
Learning Optimization
Deep Learning
Algorithms
Natural Language
Processing
Developers
Predictive
Applications
xxx
xxx
Hardware UpgradesMachine
Learning Algorithms
Expensive
Software
Data
Scientists
Application Updates
Data
Aggregation
Data
Classification
Application
Creation
Continuous
Machine Learning
Implementation
of AlgorithmsInsight Apps
7. HCM Example: Top Performer Retention
Business Question Addressed: Which of my top performers are at risk of leaving?
To help VPs of HR and lower-level managers retain top talent by:
• Building insights from performance metrics
• Integrating with Recommended Candidates to ensure prospective talent has
strong potential
• Suggesting improvements to reduce talent flight risk
Key Objectives
Performance Reviews
Identify top performers
With high flight risk
Take actions to
keep performers
Reduce
Flight Risk Cycle for Maintaining Top Talent
9. HCM Example: Top Performer Retention
Normal Department Abnormal Department
10. ▪ Model was 40% more accurate than
the manager
▪ With 25% overlap between the
manager and the model
Tangibly more Accurate Predictions
High Risk
(By Model)
Terminated
High Risk
(By Manager)
55
215
137 550
11. Retention Prediction Quality
WORKDAY CONFIDENTIAL
Period Total Random Prediction Model Prediction
Termed (Apr – June) 300 6 110 (38%)
Termed (July-Nov) 585 22 (3.7%) 169 (30%)
12. Workday Analytics
Connect
Any type – structured and
unstructured
Any size, including Big Data
Inform
Intuitive, interactive visualizations
Contextual
Predictive and Prescriptive
Act
Same global platform
Same security model
Accessible anywhere