Let’s Talk About Resume Screening
With apologies to?Bruno , we don’t talk about resume screening. But we should.
Resume screening is usually the first step in the hiring process for active candidates. After a candidate submits an application for a job posting, resume screening determines whether the application moves forward or abruptly terminates. Resume screening, whether performed by humans or algorithms, is generally opaque to candidates, who mostly see it as a hurdle they need to overcome to have the opportunity to get an interview, or at least to interact with a human being.
There are good reasons for resume screening.
Many job seekers “spray and pray”, especially after a more targeted approach fails to deliver results. Not surprisingly, this leads to a lot of clearly unqualified applicants. Employers want to spend as little time as possible on indiscriminate applicants, so they rely on automation or cursory human review to weed them out.
Other applicants simply lack required qualifications. If a job description enumerates non-negotiable requirements, such as location, work authorization, or specific credentials, it’s incumbent on applicants to show that they qualify. Resume screening should filter out those who don’t.
So far, so good. If employers enumerate?clear job requirements , applicants should self-select themselves accordingly. And unqualified applicants shouldn’t be surprised when employers don’t follow up with them.
But resume screening often goes beyond filtering out candidates who lack explicit qualifications. Heuristics favor well-known schools and employers. When people screen resumes, they bring their own biases to the process. When algorithms do it, those?biases are baked in to their heuristics or machine learning models . And the opacity of the process means that candidates can only guess why their resumes are being screened out.
Part of the problem is that employers write terrible job descriptions. Instead of listing clear job requirements, they rely on resume screening to “know it when they see it ” — a subjective evaluation that is ripe with bias.
领英推荐
Nonetheless, it’s unreasonable to expect an employer to interview every applicant. Interviewing requires an employer to make a significant investment, and that investment only makes sense if it provides a sufficient expected return. Even if employers recognize that their resume screening process is biased, they can’t afford to simply get rid of it.
That’s no excuse for employers to engaged in biased hiring practices. Employers should continuously monitor their end-to-end hiring processes to make them more fair. But it’s not unreasonable for employers to invest more resources into the applicants they are more likely to hire.
What does resume screening mean for job seekers? It’s certainly a good reason for job seekers to invest effort into optimizing their resumes. That might mean seeking professional help, though there’s a lot of great advice out there available for free. It also helps to be at least somewhat targeted as a job seeker, rather than taking a “spray and pray” approach.
But resume screening bares an inconvenient truth of the hiring process: some applicants offer higher expected return to employers than others, based solely on their resumes. You can’t fault employers for seeking to reduce risk and improve expected return. But is there any recourse for the applicants who find themselves on the losing end of this process?
There is. Since employers are trying to reduce their risk and improve their expected return on investment, candidates can share the risk and make their own investments into the process. Specifically, candidates can supplement their resumes with evidence from a trusted third party that can vouch for their qualifications.
Companies like?Karat ?(which I advise) and?interviewing.io ?offer candidates the opportunity to not only learn from practice interviews, but also bypass resume screening by demonstrating their skills through those practice interviews. An increasing number of companies are embracing this approach, and hopefully we will start to see a more level playing field for candidates whose resumes undersell them.
We should all want hiring to be effective, efficient, and fair. But we can’t ignore the reasons that organizations use resume screening. So we should remove bias where we can, and work to find other ways to help candidates whom resume screening unfairly disadvantages.
A CTO and futurist who demystifies the impact of emerging technologies and pursues their opportunity to transform product offerings and create new markets. I love hairy problems!
2 年Great summary. I guess it is only a matter of time until a creative developer with some spare time on their hands writes a job submission bot to find out what things the AI is filtering on. It'll certain ensure no human ever has time to be in-the-loop. Taking it up a notch, maybe someone will then develop a GAN that will generate the ideal CV/resume given (hopefully) factual inputs. So then we'll have AIs filtering AIs to get back to the facts - "what can you do and how well do you work with people." Like most arms races - intelligence itself being one - it doesn't really benefit anyone, but once the race starts it'll never end.
Director, Consumer Data Science at Doordash
2 年Agree there is a real fairness issue with this step of the interview process. But I’m skeptical that existing 3rd party solutions can provide sufficient signal to overcome inherent biases. That’s because those signals are weak (Karat works well for standardized interviews, but it’s much harder to test a candidate’s problem solving or thought process with that format) and because candidates can easily learn to game the system. For data analytics interviews, I find that take homes can be a good unbiased indicator - the downside, of course, is that they require a non trivial investment from the candidate. But it could be a way to give some candidates a way to prove themselves. And I think cos should be willing to compensate candidates for the time spent if they don’t advance to the later stages.