People Analytics: Job Hopping and The Science of Fit

Recently I sat down for dinner with a bunch of agency recruiters in Singapore. Being new to a talent function, I was curious to understand how they assessed candidates before submitting their resumes for review. One of the most interesting insights I heard was "I never forward resumes of "job hoppers - they never work out". If you are not familiar with the term, a “job hopper” or “job jumper” is a recruiting prospect who has had short tenures with several employers. The assumption was simple: job hoppers are disloyal, selfish and impatient and if they quit their previous companies quickly, they are likely to quit yours quickly too.

***********************************************************************

Working in People Analytics, I knew I had to verify this. Unfortunately, there are very few cross industry academic studies on this phenomenon so I could not challenge the validity of my findings against peer-reviewed literature. To test this I worked on a data-set of employee data from a company with 6000 employees across both Asia and Africa and had the following features:

1. Attrition Status: 1 if yes, 0 if no.
2. Tenure: Measured in months since hire.
3. Job Hopper: If an employee had less than two years of tenure in each of their two previous companies - 1 if yes, 0 if no.

The hypothesis I wanted to test was simple: Do job hoppers (defined by having less than two years of tenure in each of their two previous companies) have a higher attrition rate under 2 years of tenure that non-job hoppers?

If what my recruiter friends said stood to be true, there should be a significantly higher proportion of job hoppers quitting within 2 years of tenure than non job hoppers. Job-hoppers should keep hopping. However, at a 95% CI, this hypothesis did not hold true. At an aggregate level, job hoppers and non-job hoppers displayed similar rates of under-2 year attrition.

However, as with all research, further testing was required. We included moderating variables such as performance ratings, promotion rates, team attrition rates and manager 360 ratings to see if they showed us anything new. They did not. Except for one measure - fit.

**************************************************************************

Fit can be described as how well an employee's values align with a specific environment's values. We have all heard varying forms and descriptions of fit but scientifically it is described across three lines - PO (person-organization fit, also known as culture-fit), PJ (person-job fit) and PS (person-supervisor fit). For each employee we calculated a fit metric across these three lines using scientifically tested frameworks. The exact frameworks we used to quantify "fit" are worthy of an entire publication itself - but in this post, I will only skim the surface of them. If you are interested in how we did it, drop me a note and I'll be happy to share. We calculated the following fit measures:

1. Culture Fit: Both at a country level and supervisory level measured using Hofstede's corporate culture framework. An employee with high culture fit had strong value alignment with the organization they worked for.
2. Job Fit: A measure of how much an employee fits with the role they were doing. For example, one of the measures within job fit is "Skills Utilization" - an employee whose core skills that are not being utilized in their role are classified as having low skills utilization and in turn low job fit.

3. Manager Style Fit: A measure of how an employee's preferred style of being managed closely aligned with his or her manager's preferred style of management. For example, an employee who prefers to be innovative in the work they do will not "fit" well with a manager who is risk averse. Overall manager fit metrics took into account differences across: feedback, decision making, risk tolerance, rewards, and people over process orientation.

Results

1. Job-hoppers with low job-fit turned over faster than non job-hoppers with low job fit. Job-fit especially along "Skills Utilization" and "Learning Opportunity" metrics showed a significant difference between job-hopper and non job-hopper groups. Job-hoppers place a greater emphasis on learning opportunities and utilizing their full skills and was often a key driver of under 2 year tenure.
2. Job-hoppers with low culture-fit at a supervisory level turned over faster than non job hoppers with low culture-fit at similar supervisory levels. Quantifying culture is a gift that keeps giving. Our first finding was that culture was not singular - it manifested differently at different levels of the organization. Finance functions had different cultures to Marketing. HR had different cultures to IT. Job-hoppers with low culture fit within their own supervisory level across all lines turned over faster.
3. Employees with low manager-fit turned over at similar rates regardless of whether they were previously job hoppers or not. Manager-fit affected both job-hopper and non job-hopper groups in the same way - it was not necessarily a driver for job-hopper turnover.

Summary

At an aggregate level, job-hoppers and non job-hoppers show similar characteristics. However, job-hoppers place a greater emphasis on two types of fit - job and culture. They have a lower tolerance for roles that are not in line with the aspirations they have as well as a lower tolerance for teams with misaligned values. It is unclear why job-hoppers show this distinction - and merits its own research study. Any explanation I provide is mere speculation. But if I had one takeaway for you, it would be to not treat job-hopping as a blotch on a candidate's CV - rather, pay attention to the areas of fit during your interviews and assessments. Place a greater emphasis on fit to ensure who you hire has a strong alignment with the role they are applying for as well as the team they will be working in. A job-hopper with the right fit will not remain a job-hopper. Don't rule them out.

Thank you for reading! So why do you think job-hoppers show these distinctions? Leave a comment below!

Disclosures

This piece was written based on an analysis on a single company in a single region of operation. It is important you treat insights with caution and replicate similar studies in your own company. Further study and replication is needed to establish validity and reliability of our claims.

If you are interested in the frameworks we used to quantify fit, do drop me a message.