The Tangled Web We’ve Weaved in Hiring Automation

Written By Geoff Kreller, CRCM, CERP

The effort of companies to attract the best candidates in the most economical way possible has had the unforeseen consequence of making the hiring and recruiting process excruciatingly inefficient for both HR departments and prospective candidates. While each step to hiring expediency and automation made sense, they also created new challenges that uncovered significant fundamental weaknesses in the system and ultimately to sub-optimal hiring decisions overall.

The rise of job search engines such as CareerBuilder, ZipRecruiter, and Monster.com (may it rest in peace) and job networking sites such as LinkedIn created a streamlined and cost-effective path to posting and viewing available positions. However, the byproduct of that efficiency was HR departments that were quickly overrun by the sheer volume of received applications and resumes. Look at any position you see on LinkedIn today, and you’ll notice that hundreds (if not thousands) of people have applied within two weeks of the position being posted.

Faced with the challenge of how to efficiently vet a large number of applicants, many hiring departments transferred this responsibility to third party companies who promised to distill their enormous applicant pool into a few qualified candidates. What started as keyword detection evolved into full applicant tracking systems (ATS) designed to automatically parse, sort, and rank applicants entering the hiring system[1].

While ATS are often necessary to resolve the challenge posed by large numbers of initial applicants, the model for sorting and ranking may not be created in a way that artificial intelligence can logically interpret and produce the same rankings as a human who could consider the criteria provided and substantially apply it to a resume without specific logic or keyword matching[2].   

The result is a very broken, inefficient system where job seekers become incredibly frustrated, and companies miss out on candidates that could have been better fits for the role or the enterprise overall. Hiring apps have become this generation’s dating apps – to obtain an interview (your “first date”) it could take hundreds or thousands of job applications. While there are potential solutions to increased efficiency, they require companies, job search engines, and applicants to resolve the noted gaps below.

1.       The position never existed.

The lower cost and higher visibility of creating a job listing have created a moral hazard for companies. According to a survey in 2024, 40% of companies using Resume Builder admitted to posting fake job listings[3].

We’ve heard about job scams where criminals create spoof companies to obtain personal information (which businesses should periodically search for and remove)[4]; this actually feels worse. Whether it’s because a company wants to create a favorable impression about expansion to stakeholders or reassurance (or a threat) to existing employees that additional help is on the way – there shouldn’t be a place for these fabricated positions outside of a 10th grade class learning resume and interviewing skills for the first time.  This feels a lot like a “dating bot” – has anything positive resulted from those conversations?

Beyond the inefficiencies created in the hiring system, ghost jobs can lead to reputational damage with stakeholders, partners, and employees if discovered. Businesses that post ghost jobs may also be susceptible to civil lawsuits under different legal theories such as fraudulent misrepresentation, unfair business practices, and false advertising[5].

While I appreciate the videos and content dedicated to detecting and avoiding ghost jobs, there are opportunities to hold both companies and job posting boards more accountable for the content they disseminate and to cure this concern on a broader level.

2.       The ATS is never credibly challenged.

Just like any other model, an ATS needs to be tested, refined, and recalibrated. AI and ATS are only as good as the logic they are programmed with and the people responsible for its programming. There is often not enough collaboration between a company’s hiring function and IT/Engineering to develop a model that accurately identifies an institution’s top candidate pool. Communication is key to resolve fundamental gaps in keyword matching, strict algorithms, layout parsing, and identifying key context and nuance[6].

If the logic is too broad, the applications sent to the hiring department will be filled with white noise and false positives (people who don’t match the job criteria). If the logic is too narrow or rigid, dismissed applicants will be filled with false negatives - applicants who were qualified for the role and could have made significant impacts on your enterprise.

While humans understand synonyms, career context, and nuance, these things are very difficult to embed within an automated set of rules and logic. Instead of adjusting the ATS or requesting a larger selection for manual review, there are a plethora of companies (Jobscan, Enhancv, Resume Nerd, etc.) who instead modify resumes to score higher in the ATS ranking system. Keep in mind that if the ATS isn’t performing its task well, all that means is that it’s presenting a company applicants who happen to score well within the ATS criteria of key word matching, logic, and parsed history – but that score is not accurately reflecting the candidates who might perform best in that role.

Compliance and Model Risk Management (MRM) wouldn’t allow other critical models to run unmonitored, and (arguably) the most important decisions a company makes is the people they hire. On a regular basis, companies should take a sample of both the applications dismissed and those sent forward for further review – use those findings to refine and iterate on your ATS.

3.       The friction in the process is not calibrated to the position’s complexity.

In theory, LinkedIn’s “Easy Apply” option is a godsend for saving time. Click the button, make sure your most recent resume is attached, upload a cover letter with the company and job title (if you even have the option to upload it in the first place), and submit!

In practice, these low-friction flows lead to applications coming from applicants who are underqualified, not qualified at all, or are simply not a great fit because either they didn’t read the job requirements or the system didn’t have the capability to adequately review for those prerequisites. Regardless of how you feel about in-office, remote, and hybrid workforces, if a position is listed as in-office or hybrid in Fargo, North Dakota, you should be fully prepared to explain why you’d want to relocate your family from Austin, Texas.

It makes sense to have an easier process for an entry level role, but perhaps not so much for the company’s upper management. The complexity of the posted role should be mirrored in the process.  In many cases, the relative ease of the application does not match the initial due diligence that should come with that particular role.

Job board application flows often allow for initial due diligence questions to be asked prior to application submission; companies should leverage this opportunity to ask quantitative and qualitative questions that correlate to the job’s requirements, the company’s mission, and its overall values.

4.       The company knew who was going to be hired before the position was posted.

“It’s not what you know, it’s who you know.”

By the time that a job is first posted publicly, it may have already been viewed by internal employees, headhunters, or affiliates of the company. It’s entirely possible that the job was posted externally to benchmark the odds-on internal favorite for the role[7].

It’s also a myth that companies have to post all of their open positions publicly[8]. In fact, many sources suggest that only 20-30% of job openings are publicly posted[9][10].

In addition to building personal connections with strategic partners and key decision makers, networking enhances access to hiring opportunities[11]. This doesn’t mean a job seeker should ask “what jobs do you have available” to everyone they meet at a conference or event, and companies shouldn’t use these events to constantly poach talent from their competitors. Instead, these networking events create opportunities to stand out and be memorable through genuine conversations and exchanges of ideas and experience.

Both job seekers and companies have opportunities to further establish their reputations as trusted, valuable members of their networks and associations by donating time, energy, and resources to those organizations.

Summary

Without additional controls and manual review, automation of the hiring process has cultivated a system of inefficiency, frustration, and sub-optimal outcomes. It is a system rife with incomplete (or fabricated) information, and a system that appears to have lost sight of its original purpose - bringing together employers with high-quality candidates for their posted roles in a cost-effective manner. I’m concerned we are a hair’s breadth from deploying Tinder’s swipe left/swipe right approach to hiring matches.

The way back is not easy. It would require companies to be honest and open about available roles and the requirements and benefits associated with those positions. Job seekers would need to take more time to research both the open role and the company, and to determine if they meet the noted prerequisites and requirements. Job boards and social media platforms would need to be accountable and responsible stewards of the information provided, and to credibly challenge businesses and their job postings, especially when complaints are received.

I want to be proven wrong, but I don’t have a tremendous amount of faith in companies, job boards, social media platforms, and job seekers collectively coming together and systemically solving these inefficiencies, especially in the short term.

As a result, it is imperative that individual companies recognize the folly of posting ghost jobs and the virtue of managing their ATS as a critical model, for job seekers to embrace networking opportunities available to them, and perhaps ensuring job boards don’t fully merge with dating app technology to become CareerMatch, eWorkHarmony, or BumbleIn.

Follow NAQF and Geoff Kreller on LinkedIn for additional insights. For more information on how NAQF can help your organization with attracting and retaining talent, contact us at contact@naqf.org.

Article References

[1] https://www.resume-now.com/job-resources/resumes/what-is-an-ats-resume

[2] https://www.resumaipro.com/blog/why-your-resume-might-never-be-seen-the-dark-side-of-ats

[3] https://www.cbsnews.com/news/fake-job-listing-ghost-jobs-cbs-news-explains/

[4] https://www.nbcnews.com/news/us-news/job-scam-ziprecruiter-linkedin-work-postings-fake-listing-rcna238162

[5] https://audetlaw.com/lawsuit-updates/ghost-jobs-lawsuit-update/

[6] https://www.resumaipro.com/blog/why-your-resume-might-never-be-seen-the-dark-side-of-ats

[7] https://www.askamanager.org/2010/06/why-do-companies-post-positions-when.html

[8] https://legalclarity.org/do-companies-have-to-post-job-openings/#google_vignette

[9] https://gameofjobs.org/what-proportion-of-job-opportunities-are-posted.html

[10] https://www.cnbc.com/2019/12/27/how-to-get-a-job-often-comes-down-to-one-elite-personal-asset.html

[11] https://www.forbes.com/councils/forbesbusinesscouncil/2023/08/10/the-art-of-networking-five-ways-to-build-connections-that-matter/