Enterprise Resource Planning Blogs by Members
Gain new perspectives and knowledge about enterprise resource planning in blog posts from community members. Share your own comments and ERP insights today!
Showing results for 
Search instead for 
Did you mean: 

You may have identified some of your employees that you would like to predict for possible employee turnover. These could be new hires  or customer service employees or selected by some other criterion. 

For example consider following test data of 5 active employees

In order to make predictions, let us understand the actual historical data of employee turnovers for example we can take employee turnover data of  (say)last 2 years and use it to make some rules.

For example following is snapshot of one such historical records data

The actual historic data & rules built analyzing it can be used for data mining to make decisions.

Predicted results for possible turnover for test data employees could be as follows


Value of YES in Turnover column is just indicator of high chance of employee turnover , it does not guarantee that it will happen in actual . It is just indicator of potential or likelihood based on data mining , rather than analyzing all , this smaller data can be used for further analysis or investigation .

Let ‘s cover some background , motivation or rationale behind the working of above described scenario.

HRIS Implementations using SAP HCM solution have  great reporting capabilities which allows users to answer questions about demographics, performance , characteristics or costs of employees. For example Questions like How many employees have started/left ? How many will retire in next 24 months ? Number of Training done last quarter  ?

Such analysis provides business view on what happened and what is currently happening. Though it easily provide answers to questions like how long an employee has been with us , whereas if the questions were which Employees are most likely to leave ?  or how long will employee stay ? we need Predictive analysis which provides future looking insights on the business.

These analysis techniques are based on mathematical statistical  models used in data mining and these models generally work very well specially on large amount of data, predicting what is likely to happen based on historical data. Although predictive analysis addresses outcome or what will happen next , further analysis can be done to answer why it's likely to happen and what if these trends continue ?

The questions requiring predictions in future could be varying in nature and importance to business i.e. instead of answering how many sales representatives left us last year ? the question which sales representatives are most likely to leave ? which employees are most likely to perform or promoted ? who are our most valuable employees ?

Traditional HRIS alone can not answer such type of questions.

Predictive analysis are not new and have been in use in various industries  and now SAP HANA Platform has the computing capabilities to build real time predictive analysis .

Predictive analysis can help in forecasting sales demand or help in product proposals for cross selling to customers and have many different uses in other business scenarios .

In context of Human Capital Management, Employee turnover is a major problem for many organizations. It is important to understand drivers of employee dissatisfaction as happy employees  also mean happy customers. Great talent is rare, hard to keep and in high demand. Predictive analytics can be a very useful tool to set up a well targeted employee retention campaign. Employee turnover predictive application can be used as a tool in context of Strategic Workforce Planning.

For employee turnover predictions, first we need to start from a historical data set of staff turnover.

Perhaps the most important step here is to collect the right variables for the staff turnover data set and it is important make sure they have the necessary data quality.

Few Examples of such predictor variables are personal or demographic information (age, marital status etc), income ,  departmental information (e.g. some departments may have high turnover rates), any other information regarding employee training or assessment data, etc. For the example shared here I have used Classification Decision Tree Predictive analysis for Data mining.

The application allows us to identify rules on the variables , for example we may have information on Employees Age, Educational Qualification, Occupation or position, Salary, that determine an action or decision of turnover.

The algorithm to identifies rules on the variables that determine an action or decision of turnover for example employee with age  less than 40 with a Phd degree working in Sales and having annual Salary less than 60,000 may have a greater chance of turnover  than some other criterion.

I am sharing below video demonstrating working of this POC predictive employee turnover analysis application built  over SAP HANA Cloud Platform. Feel free to contact me for details. It utilizes Single sign on capabilities & by default anyone with SAP S user ID or (Free) SAP on demand Id can be easily given access to use it.  Only comma delimited csv  data files , one for historical employee turnover data as step to create rules and other file for employee data that need to be analyzed are needed to test run it.

1 Comment
Labels in this area