Article originally published on LinkedIn, 5 Sept. 2016
On this US Labor Day, here’s a quick look at some aspects of recruiting and job seeking that don’t seem to get much attention – and certainly not in one place. Having been on the hiring and applicant side of the job market at various points, I have paid attention to the process and numbers. Some numbers defy belief, but those are only tip of the proverbial iceberg of a system that is changing in the wake of technology. Taken together, what do these numbers and dynamics say about the health of the system and its future?
- Disparity between time put into preparing a resume and the time it is reviewed: A recruiter might look at a resume in 6 or 8 seconds (or maybe 30 on a slow day), if we are to believe commonly cited articles. Estimates on resume preparation vary (and are complicated by the fact people re-use or revise existing resumes), but one estimate is that a professional resume writer may take 6 hours on one resume. That would mean a recruiter spends somewhere between 0.0003% and 0.0014% of the time it took to write a resume to dispose of it. Even reducing the amount of time going into a resume or accounting for reuse of content there is still a large disparity. This is a systemic problem, not the fault of recruitment.
- Large numbers of job applications submitted/received per open position: One estimate is that there are on average 118 applications per job opening, though one assumes that the number would tend to be higher for more junior positions and lower for more senior ones. Considering the process at any level results in rejection of all but one, in most cases, this represents a significant waste of human time with a real economic cost on all levels (although there may be less applicants and consequently rejections for senior positions, the hourly cost of lost time at that level would be higher). One job seeker boasted of 1200 rejections. If we allow a minimal 1 hour per application (which would total over 6 months’ work) and a modest professional rate of say $25 that would amount to something like $30,000 in labor – not even counting the increment of cost to the employers – and nothing to show for it on either side. That may be extreme, but the aggregate time that goes into the arc of unsuccessful job applications across an economy must be huge. With no nostalgia for the days of typewriters, shoeleather, and rotary phones, has digitization actually increased the unit cost of finding/filling positions? (Analogous perhaps to how wordprocessors have paradoxically increased use of paper?)
- Increasing technological imbalance in the overall application process: Job seekers these days have to put more time into each application through manual completion of online forms and sometimes skills tests, while recruiters have less time per application, but more automation to compensate (in dealing with the quantity). Digital technology has facilitated the job seeker’s task in many ways, but only to a point, while there is now talk of applying artificial intelligence (AI) technology to recruiting and selecting candidates.
- Problems with automation & the “keyword arms race”: The large number of job applications received has required recruiters to automate their work, which at this stage of technology still revolves largely around keyword searching databases of applicant information (applicant tracking system or talent-management software). The same source that carried the estimate of 118 applicants per position also said that ~50% of applications are deselected by this method. But a lot depends on the keywords. The oft-cited case where one company screened 29,000 applications for one engineering position without a single applicant being found qualified is an extreme example of what can go wrong. The main dynamic however involves advice to job seekers to add keywords into their resume – nay, “load” them with keywords. To which the natural response on the recruitment side is to “raise the bar” with software to screen in more sophisticated ways. It is hard to see how such an “arms race” has a positive outcome.
- Job site massification: Newspaper classifieds have long since given way to online listings, and once those started it was maybe inevitable to get specialized websites for employment. These follow a logic that leads to staggering numbers (SimplyHired.com‘s tag line, “Explore millions of jobs…” is typical): more jobs, more hits, higher SEO rating, and more employers want to list. That means money. It also means the dysfunctions inherent in a numbers game – jobs listed more than once, or after they have been filled. Apparently there is more to this industry, as one can learn from a quick tour of JobBoardDoctor.com, but it seems to serve first the employer.
- Sites that offer to serve up lists of appropriate openings: Some companies offer services to search for the job seeker, based mainly on terms one provides. Actually this concept is not new, having been introduced as search agents on the human resources pages of various organizations. Now however specialized sites work across organizations and get into big numbers. Recruiter.com proclaims to job seekers: “One platform. 6+ million jobs” (an email publicity puts it “Let Recruiter sift through 6 million jobs daily and tell you about new opportunities in your area”). And to employers: “Put your job in front of 35,000 recruiters” (who presumably have access to the job seekers). StartWire.com claims to have almost 8 million jobs and an “employers network” of over 17,000. Also, it is possible for a job seeker to get daily emails each with scores or hundreds of listings – again not so new as DevNetJobs.org has for years been mailing ~2000 listings every few weeks to subscribers, which one must read through like classifieds of old (and maybe … a keyword screen search).
- Free resume listing services – one in a million: Online resume listing services, sometimes in connection with the mass job sites, receive vast numbers of submissions (all presumably loaded with keywords), and advertise to those submitting resumes that some huge number of employers can access them (in the tens or hundreds of thousands) – guaranteeing automation. Although the number of resumes collected for employers to view isn’t states, these are certainly very big haystacks in which to place one’s needle trusting it will be found. Also, here like #6, huge storage and churning of information at what cost/benefit?
- Services that broadcast resumes to large numbers of potential employers: These companies offer for a fee to send your resume or CV to some large number of employers on a curated list (as opposed to just listing the resumes for free). For example, DevelopmentAid.org, an otherwise helpful specialized job listing service, offers “CV broadcast to 29949 organizations.” This may seem like revenge of the resume writer, but at the risk to the sender and receiver of turning the resume into spam. In addition to what the job seeker pays for the service, the company must invest in compiling and updating the list of addresses, and the recipients must spend time dealing one way or another with the extra email. No idea on total cost to the economy.
- Niche services – the good, the bad, and the expensive: Among the various intermediaries in the job market are independent recruiters (doing searches on contract), job coaches, and placement agencies. At their best these help various parties deal with imperfect information and improving strategies, but this is a mixed bag. Some work on the basis of fees from the company seeking to fill positions, and others on fees from those seeking employment. As in any industry there are better and worse buys, but the latter group has a harder case to prove – as one employer financed recruiter once put it, “I don’t feel it is fair to take money from consultants while they are without jobs or in between assignments.” On the higher end, fees asked of job seekers can easily rise into the thousands of dollars, with one service asking five figures – all with no guarantee of results. There is significant money involved in this sector, with possible ethical issues as well.
Taken together, are these indications of a paradigm at the verge of a change? What to? Presumably more technology – and the oft cited AI – will play a part, but how, and for whom?
(The material in this post comes from personal observations and some research for a forthcoming blog post on “AI and the self-driving resume.”)