You are ready to hire. The job post goes out and the candidates flood in. Dozens or hundreds of candidates pile up demanding your precious time.
So, you use a bit of HR technology magic. With a few clicks you have a stacked ranked list – in order of the best candidates. Now, all you have to do is go top down looking at the highest ranked candidates to find that perfect fit.
Fast, fair, effective.
Or… is it?
When done incorrectly, stack ranking is fast but not at all fair and therefore not effective. When done incorrectly, you risk missing great candidates while simultaneously driving down diversity and losing all the benefits of a diverse workplace.
Here are some of the common pitfalls plaguing stack ranking processes and technologies.
Ranked with keyword match
Some processes, especially ones automated by technology, use keywords to feed the rank score (i.e. calculating who is best matched). Keywords are great… but only for measuring candidate’s knowledge of keywords. Candidates can have all the knowledge, skills, experience that you are looking for, but don’t use the same words to describe it, don’t use them as frequently, or that one detail didn’t make it into the one-page resume.
With keyword stack ranking logic, the better candidates know how keywords work and which keywords you are looking for. The more candidates stumble upon the right words, the better fit they are for the job.
Yes!
Wait… no. That doesn’t make sense.
Great fit candidates have the skills, knowledge, and abilities, not just the keywords.
So how should keywords be used?
They shouldn’t. Except perhaps to test keyword knowledge.
Ranking with preferred requirements
Job requirements tend to come in two flavors:
Must have: I won’t hire you without it
Preferred: I will hire you without it, but it would be great if you had it
Both are commonly used for ranking candidates where candidate’s score higher when they meet more must have and preferred requirements.
On the surface, this is logical… but we aren’t going to remain at the surface, are we?
There are two pitfalls with using requirements to rank candidates.
1) Misidentifying requirements
For must have requirements, there is no need to rank – candidates either meet them, or they are not really candidates. Therefore, the trick is to know if the requirements are really must have.
Many organizations confuse ‘must have’ and ‘preferred’ (or ‘nice to have’) requirements. Must have requirements means that a candidate must have them all to be successful therefore you won’t hire the candidate if they don’t meet them all.
There is a simple test to see if all your requirements are ‘must have’ – can you easily explain why it is required for the job? If you can’t – it’s not a must have requirement.
2) Misleading preferred requirements
Even if the requirements are properly categorized as ‘preferred’, many organizations use preferred requirements interchangeably with ‘must have’. Candidates get higher rankings for meeting the preferences which effectively means that candidates who don’t meet the preferred requirements aren’t considered for the job.
But that’s how it’s supposed to work, right?
No.
If the preferred requirements are inappropriate, unimportant, or unimpactful, they remove great candidates and add bias.
For example, “Master’s degree preferred.” Will the candidate necessarily be a better choice for having one? If not, all this does is add a socio-economic bias (those who have the means to get a Master’s degree over those who don’t).
Or, “experience at a director level preferred.” What is the title doing (versus the job responsibility or experience) that makes one candidate more effective than another? This requirement is biased as not everyone has equal access to promotions.
So, if not in rankings, how should preferred requirements be used?
Tie breakers. Rather than using them to rank candidates (effectively making them must have requirements) use them to help select from the finalists, along with other information like references and interview results.
Ranked with assessment results
Assessments can be a great tool to help evaluate candidate’s potential, if used correctly. However, using assessment results to rank candidates gets a bit sticky and may add, not reduce bias.
Assuming you are assessing the right traits for the job and that the assessments are valid with low adverse impact (i.e. not inappropriately favoring one group of people over another), the results are still not exact.
1) To be human is to be flawed
Consider when you were in school and took a math test. The results were likely largely reflective of your ability in math. However, did you ever accidentally switch two numbers? Or did you ever misread the question? Or have you ever taken a test during a really bad day and you were just off?
Candidates will also accidentally switch numbers, misread questions, and have bad days.
2) To be an assessment is to be flawed
All assessments have an error range. In other words, the accuracy of the measurement correlating to the thing it is measuring. So, even if every candidate reads every question correctly, doesn’t make a single silly mistake, and is in a perfect state of mind, the results are still not precisely measuring what they intend.
What does all this mean? Even with the best assessments a 95% is not necessarily better than an 88%. Therefore, if you use assessments to rank candidates, there’s a good chance great candidates are being ignored just because they scored an 88% rather than a 95%.
So, if not in rankings, how should assessments be used?
Set a minimum threshold for early stages of candidate evaluation (either candidates meet the threshold or not) and use scores alongside preferred requirements as tie breakers in final stages.
Bringing it all together – effective ranking
If keywords, requirements and assessment scores should not be used to rank candidates, how are you supposed to get through those dozens or hundreds of candidates?
With a combination of filters and evaluations.
Candidates that don’t meet the basic requirements or thresholds as measured by assessments aren’t right for the job. For those that do, evaluate them on their capability to do the job – job scenarios, work samples, situational interviews. Even better, have a process that allows you to evaluate their responses while keeping the candidates anonymous so that everyone is truly treated fairly and equally.
You can do it. We can help. Career.Place offers a structured process that filters in candidates who meet basic requirements, pass thresholds of assessments and enables employers to ask and evaluate responses to job relevant questions while keeping candidates anonymous. Contact us today to learn more.
Together we can remove bias one hire at a time.