Decoding Turnover

Data-Driven Insights to Retain Engineering & Sales Top Talent

Background

Before I got into HR, I thought it was hiring, discipline, and firing. But, after being a practicing HR professional in marketing & advertising, technology, and non-profit industries, I’ve come to realize the true power that HR can have. At its most powerful, HR is a consultative business partner that uses data and subject matter expertise to guide organizations to make better people-centered decisions.

For this capstone project, I will be stepping into an HR role where the COO has come to me with a problem: there are rising concerns about employee turnover in key departments, specifically Engineering and Sales. Turnover in high-impact departments is costing the business money and time, as well as negatively impacting productivity. The COO asks to analyze factors driving employee attrition across the company.

Key Questions

1.    How do performance ratings correlate with termination rates?

2.    What departments have the highest termination rates?

3.    What departments have the most overtime?

4.    How does overtime affect termination rates?

5.    How do gender and education play into termination rates?

6.    What locations have the most turnover?

Key Insights

1.    Engineers and Sales folks who received a performance rating of “Good” left the company the most.

2.    The departments with the most turnover are Engineering and Marketing.

3.    The departments with the most overtime are Human Resources and Sales

4.    The departments that had the most terminations of employees who worked overtime are Engineering and Marketing.

5.    Out of the 16,828 terminated employees, 54% were male and 46% were female. Employees with bachelor’s degrees had the highest termination rates.

6.    Texas was the location with the most terminations of Human Resources employees.

The Data

After looking at Kaggle and other resources, I wasn’t finding a dataset that really spoke to me. So I decided to create one using ChatGPT and a prompt by Baraa Khatib Salkini (Data With Baraa). The prompt asked for a realistic dataset of 150,000 records for HR at a tech company using a Python script which I then turned into a CSV file.

This data set contains 16 columns of employee descriptors and 150,000 rows of unique employee data.

SQL Analysis

First, I wanted to see the employees who exited and what their performance ratings were. Interestingly it was the employee who had Good or Satisfactory performance ratings that left the company the most.

I wanted to break this down by department to see how employees in each department were being rated and how that impacted terminations. In Engineering and Sales, the employees who received “Good” left the most followed by Satisfactory.

I then wanted to see the termination rates by the department to understand if the rising concern about Engineering and Sales is valid. The departments with the highest number of terminations are Engineering and Marketing.

To understand the effects of overtime, I first wanted to understand how many employees in each department had overtime through the following query. Based on this analysis, the department with the most overtime is HR followed by Sales.

To further this analysis, I wanted to understand the impact overtime had on terminations in each department.  This query showed that terminations were highest in Engineering and Marketing for employees who worked overtime.

Visualization – Tableau

After doing my SQL analysis, I created a Termination Dashboard in Tableau to visualize and present the data to the COO.

Here is a look into my presentation and the beginning of the meeting with the COO.

Conclusion

The company is right to be concerned about the number of terminations as they are increasing year over year. Interestingly, it is Marketing that is experiencing the most and so, I would want to explore Marketing, Engineering, and Sales to see if there is a correlation there. Are gaps in Marketing, due to terminations, causing terminations in the other departments? How does their work interrelate?

The locations with the most turnover for Engineering and Sales are Texas and New York. My recommendation is to look into the culture of these locations and the sentiment employees are providing through surveys and listening sessions. There is probably more data there about why Engineering and Sales folks are leaving that the company can then action on.

Engineering and Sales folks who were rated as “Good” were the ones leaving the most and my recommendation is to dig deeper into performance review and conversation data. I hypothesize that these employees may want more career progression and development so being able to dig into what was being said during reviews and performance conversations would be helpful.

Ultimately, the data has provided us with a few factors that are influencing turnover, and the next steps would be to align on the top factor to actioning on for improvements to the company’s turnover rates.

This project was done as part of the DAA Boot Camp Projects. Educational purpose.