I was reading an essay on LinkedIn: Is HAL looking at my Resume? and the connections started flying in my mind. After going through several movie-and-book analogies, including one from Terminator 2: Judgment Day, in which the Terminator cyborg played by Arnold Swartzenegger is relating the rise of the machines prior to the nuclear war triggered by SkyNet: “Human decisions are removed from strategic defense” – just like companies are using computers to filter candidates and turning to electronic games as screening tools. But I settled on something from Jurassic Park as the main thread.
As the dinosaurs run amok, there is mounting evidence that – despite a population of dinosaurs intentionally made all-female to prevent breeding – there is breeding taking place, the people in charge of the park scoff. They point to their advanced, super-duper computer’s vision system monitoring program, and do another run to prove that there is no breeding of their “all female” population. They expect N of a specific species; the computer does an analysis run from the cameras and identifies N. They crow: see, no breeding going on.
One of the visitors to the island then asks them to run a search for more than N. The computer runs an analysis again… and finds more than N. To the chagrin of the people in charge, they find that there is, in fact, breeding in their should-not-be-able-to-breed population.
This can be summed up as the computer “thinking” per its programming: In looking for X, I expect to find X; anything not exactly X, therefore, does not exist.
Applicant Tracking Systems – Only As Good as the Criteria
They have to rethink the way they go about hiring. At the moment, especially in larger employers, the hiring process has become very mechanical, without a lot of space for judgment. One employer told me that 25,000 people had applied for a reasonably standard engineering job in their company and that the hiring systems indicated that none met the requirements. The HR people who used to decide, “Are all these job requirements really necessary?” and, “Even though this candidate has never done precisely this job, I’m sure they can do it,” have been replaced by software. If employers thought about their supply of talent as carefully as they thought about their office supplies, they could solve these problems easily.
And a comment on a blog (whose link doesn’t work any more!) echoes this situation (emphasis added):
[A]n ATS is only as good or bad as the person setting it up; Taleo doesn’t decide whether or not a person is qualified, it evaluates based only on the criteria it is given. You can easily set up an ATS to do only the function it is named for: tracking applicants, so you have all applicant’s information in the same format in the same place for easy comparison.
In Search of a Fantasy
WANTED: Single / divorced woman, age 25-29; must have long blonde hair past their shoulder but no longer than mid-back, with tight natural ringlets; blue eyes, figure 36-24-36 (plus or minus one inch only on each); 5’3” – 5’6” and 100 – 125 pounds, politically liberal but loves target shooting and the outdoors. IQ above 130. No children but willing to have at least three; demure outside but wildcat in bed; certified disease-free; no criminal history. Must be practicing Eastern Orthodox. Must enjoy reading, museums, and the theater. Financially stable with high credit score from a good family with at least a Bachelor’s degree – and only had relationships longer than two years. No long-term unattached people.
Imagine if these were truly my requirements – hard-and-fast-must-haves, and I insisted they be met – I’d be single forever. But “Sixty-seven percent of hiring managers don’t feel like they have to settle for a candidate without the perfect qualifications for the job” (per a DeVry university survey). For example, I recently saw an ad for a position at a target company. This ad has nine “must have” requirements. Two all but demand that the person have had the job in order to be qualified to apply for the job; i.e., “5 years minimum experience with the design of <snip>”, and “Solid understanding of the fundamental physical principles involved in <snip>, including knowledge in one or more of the following areas: fluid mechanics/advanced flow modeling, filtration, air- or water-borne particle dynamics, tribology, electrostatics, pumps, vacuums, agitation, ionization.” Gee, could you get more specific? More importantly, where – aside from already working there – could I get such specific experience?
Across the years, two of my – *cough* – “favorite” job descriptions listed in openings read (going from memory, but pretty close as they made an impression on me):
- Wanted, urinary catheter design engineer; must have at least five years’ experience designing urinary catheters.
- Design engineer for support structures of high-energy military radars. Must have at least ten years designing high-energy military radar support structures.
Such is the fine-mesh nature of the filters being programmed into ATS software to pass candidates through for human examination.
(A side note: In class the other day we were discussing the employment market. One person said that his company has a slew of open requisitions, with internal people encouraged to apply – through the ATS portal. Of the people who were already employed at the company, and from my understanding many of whom were already involved to some degree in the project requiring the ramping-up, none were passed through the ATS.)
ATS in the Air
Discussions of ATS programs seem to be on an uptick. Two notable people, both of whom I follow on twitter and with whom I communicate personally, have recently posted essays on the topic.
First up is Marcia LaReau, who cautions people against trying to work around ATS systems in her essay The ATS Work-Around…Maybe Not Such a Good Idea. It’s a good read; it certainly shows how companies have not only become reliant on such software, but how powerful HR has become in the hiring process (in cross-reference see Ask the Headhunter Nick Corcodilos’ blog essay Why HR should get out of the hiring business).
She raises a good point… ATS software exists for a reason. And job seekers who attempt to go around them entirely do so at their peril.
Providing some additional insights, Neil Patrick of 40pluscareerguru posted this piece: Applicant tracking systems – the hidden peril for job applicants.
First, he points out research that shows the software can reject even hypothetical “perfect” resumes used as tests (emphasis added):
In a test last year, Bersin & Associates created a resume for an ideal candidate for a clinical scientist position. The research firm perfectly matched the resume to the job description and submitted the resume to an applicant tracking system from Taleo, the leading maker of these systems.
When the researchers then studied how the resume appeared in the applicant tracking system, they found that one of the candidate’s job positions was ignored completely simply because the resume had the dates of employment typed in before the name of the employer.
The applicant tracking system also failed to pick up several key educational qualifications the candidate held, giving a recruiter the impression that the candidate lacked the educational experience required for the job.
This perfect resume only scored a 43% relevance ranking to the job because the applicant tracking system misread it.
This echoes an anecdote I heard – protecting the confidentiality of where I heard it – where a company tested its ATS with an artificially-concocted “perfect fit” resume and it didn’t go through the ATS for review.
Neil then goes on, echoing Marcia, with some recommendations on how to structure a resume to increase the likelihood a resume will get through to a human reviewer.
Questioning the Timing
I would like to pose, as an exercise for your consideration, the following idea: did the “skills gap” and “shortage of qualified candidates” coincide with the introduction and increasing prevalence of ATS software – and growing insistence that hiring managers not go around HR through networking (married to hiring managers waiting for the fantasy date)? If I were a betting man, I would say “Yes”.
What Can Be Done?
In my essay Force and Counterforce: Equilibrium I proposed several new metrics for consideration to balance the search for the fantasy date. Let me summarize one, specifically, in reference to ATS software:
New metric: The ratio of resumes submitted to a company’s ATS to the number passed through for consideration. This will give an indication of the fineness of the filter; while some people may apply for “stretch” applications where they are clearly not qualified, for the most part I believe that people will not apply for something where there is not at least an 80% fit. If a company finds more than, say, 50% of the submitted resumes being filtered out, that is an indication the filter is too fine and that the “must have” requirements need to be scaled back.
Granted this will increase the work load of HR people who will have to review more resumes, but… if one of the prime functions of HR is to get people into open requisitions, then this is a necessary action.
A second thing is networking. Just as job seekers attempt to contact hiring managers, the reverse should be true. If there are open requisitions – meaning that the company needs people to get things done – part of every manager’s job should be to attend networking groups (e.g., Acton Networkers and WIND) which are a rich source of skilled professionals seeking work, and being freed from some morning activities (in a time counterbalance) to attend trade group meetings once or twice a month in the evenings. This has the added benefit of a lower-stress environment to examine how people behave to identify “fit”.
Lastly, if the expected flood of “perfect fit” applicants is not happening, HR needs to sit down with hiring managers to take two actions:
First, a review of each must-have requirement; is each absolutely necessary? If yes, how can they be generalized to broaden the mesh? (E.g., instead of “Must have five years of experience with SolidWorks”, try “Must have five years of 3D solid modeling experience; SolidWorks preferred.”) This identifies the skill behind the software. Another example: instead of “Must have eight years of experience in medical devices” consider the real requirement: working in a highly-regulated industry, so broaden the criteria to “Eight-plus years working in regulated industries; medical devices preferred.”
Second, for each must-have requirement remaining, identify a specific course or training program that could fill in the gap for that item should someone who is otherwise a solid 80-90% fit be identified. Get approval for the training budget from upper management.
The Definition of Insanity
Albert Einstein is purported to have defined insanity as doing the same thing repeatedly expecting a different result. If candidates are not pouring in to fill positions, what will employers do to change how they find people?