How a scientific approach to selection can reduce costs, improve efficiencies and prove return on investment
Would it surprise you to know I’m a secret statistics geek? People tend to think psychology is a ‘soft’ subject but it’s not, it’s a science. A science based on statistics. Statistics enable you to make predictions about your recruitment activity and provide the evidence you need to make confident hiring decisions. But you don’t need to be a scientist to use statistics – here’s how you can too.
Picture the scene. You recruited a number of people six months ago using a range of recruitment exercises. Feedback from line managers suggests varying levels of performance from those employees, some of whom have already left the business. You wonder whether your selection process is up-to-scratch but how can you be sure? And what can you do about it? More worryingly, the finance director wants to know too.
Recruitment costs money. So does making the wrong hiring decision. But if all you’ve got to go on is a feeling that things aren’t quite right with your selection process, how can you make any changes with confidence?
Through the appliance of science.
Rather than rely on gut feeling or anecdotal evidence from managers, take a scientific, systematic approach to evaluating your selection process – by gathering and analysing data. It will provide the evidence you need to:
- make confident decisions about changes to your process
- ensure better quality hires (which will please the line managers)
- prove return on investment (which will appeal to your finance director).
How do you evaluate your selection process?
To evaluate your selection process, you need to measure the link between how people performed at assessment and how they are performing now. This will enable you to see which elements of the selection process are successfully predicting performance in the role, and which recruitment exercises aren’t working as well.
First, gather your data
- What data should you collect from the selection process?
In order to make an effective, useful evaluation of your selection process, you need to be collecting the scores of successful candidates from every stage – from sifting to interviews to assessment exercises including every competency, every score.
- How will you capture that data?
It can be hard, if not impossible, to collect this data after the event. Candidates’ scores may not have been recorded electronically or in a way that can be analysed effectively.
It’s crucial therefore to plan this into the design of your selection process right from the beginning.
For instance, I’m designing a selection approach for a client at the moment. At this early stage, we’re already talking about what success looks like six months down the line in terms of key performance indicators. Meaning we know what we want to evaluate at that point so we can make sure we gather the relevant data now, in a way that will mean it’s easy to analyse.
- When would you expect to see performance in the role?
Would you expect a new employee to be up to speed and performing effectively in their role within 6 months, perhaps sooner? This is your benchmark timescale for gathering data on their performance.
- How are you going to evaluate employee performance?
What measures are you going to use to evaluate performance e.g. turnover, absence, customer satisfaction scores etc? One of my clients had a significant issue with people dropping out during initial training so that was one good area to measure against.
- How many people do you need to include in your analysis?
To ensure the statistical stability of your findings, you need to gather data about at least 50 people (smaller numbers mean one person can have a big influence on your results. In those circumstances a more cautious approach is required, and a greater amount of qualitative or descriptive data is gathered).
Next, compare the scores successful applicants achieved during the selection process against their performance in the role
Now you have your data you can assess the correlation or link between scores during the selection process and the data you have gathered about performance in the role.
You should start to see which elements of the selection process are statistically more effective at predicting performance than others.
In other words – if those candidates who scored highly in the role-play exercise are now successful performers and those who did not are underachieving, then the role-play is very likely to be predictive of job performance.
Here’s a real life example – a client of mine found that one of their assessment exercises – a role-play – was statistically predictive of job performance. The interview, on the other hand, had the weakest correlation. Given that evidence, they placed more importance on the role-play and reduced the number of interview questions so they just focused on motivation and attitude. This meant they saved 45 minutes from their interview and they could be more confident about their hiring decisions.
Taking a scientific approach to selection can help you to evaluate whether your recruitment activity is giving you a return on investment. If you’ve spent £5k on a hiring campaign but you can prove that the new starter achieved 20% more sales than those hired in the original way, you’ve shown a return on investment. This approach can also help you make cost savings by improving how candidates are assessed. And it can help you be more confident about your hiring decisions.
Thank you for reading this article. I post regularly about recruitment and people development issues – to help you create a brighter workplace. To get future blogs sent direct to your inbox, sign up here.