By Paul Basile, CEO of Matchpoint Careers
To hire is to predict. Hiring technologies or practices that don’t help predict simply don’t help.
And yet, more and more technological innovations applied to recruitment and talent management seem blind to prediction or, more frequently and more damagingly, antithetical to the ability to predict.
Take applicant tracking systems (ATS’s). Please. We can’t seem to live without them but we aren’t living well (yet) with them. Well over 90% of large companies use ATS’s, in part to screen out applicants because the volume of submissions overwhelms human sifting. Data from last year in the US confirm: there were 89 applicants per job posted; 144 individuals, on average, applied for each entry-level job offered; Starbucks received 7.6 million applicants for 65,000 jobs, and 1 million applicants vied for 2,000 Proctor & Gamble jobs.
Employers call on the aid of technologies against this tsunami of candidates – who themselves are aided by technology, of course, in their ability to apply rapidly for many jobs. As a result, it’s estimated that 85% of applicants don’t possess the basic requirements for the job to be filled.
ATS’s filter the maybes from the no-thanks using keywords fixating on employer-specified factors like years of experience or work at (desirable) former employers or quality of school attended. These are all poor predictors of performance as a great many scientific studies have demonstrated.
The science of psychometrics has revealed the key predictors of performance. For the large majority of jobs, these include (a) cognitive ability, (b) behaviors, (c) preferences and, to a lesser extent, (d) skills. Not experience, not school pedigree. To screen out or rate candidates on anything except that which predicts performance serves no one’s purpose.
Further, candidates attempt to game the system by loading selected keywords onto their resume. They are in fact advised to do this. This won’t increase the likelihood that the employer hires the right person nor that the candidate finds the right job. Keyword screening and candidates’ attempts to bend that system to their own ends further exacerbate The Recruiting Problem: current hiring systems too infrequently lead to the selection of the right person for the job.
As evidence of The Recruiting Problem, note that nearly half of new hires leave within 18 months. Fewer than half of new hires – after all, the technology parsing and interviewing and social networking – can manage a “very good” or even “good” performance rating six months into the new job. And about a third of company failures, each year can be attributed to poor hiring decisions. Yikes.
All of this rests (not so securely, as it turns out) on the resume. The resume remains the primary means of selecting among candidates. And that, dear Eliza, explains a lot. Resumes lead neither employer nor candidate to the right job-person match because:
1. Resumes don’t tell you what you need to know.
• Resumes say next to nothing about the strongest predictors of future performance – a candidate’s behavior patterns and general mental ability. Nor do they reveal much about a candidate’s engagement, the other major driver of performance.
•They don’t tell you anything about how a candidate will fit in this specific job in your organization. The business pages are full of lurid accounts of bright stars in one company who sank without trace in another.
2. Resumes say lots about things that don’t matter.
•Length of experience, skills, and knowledge are pretty weak predictors of future performance, but you learn lots about them from a resume.
•Group performance is only tangentially related to individual potential, yet most facts on a resume (revenue figures, cost savings, etc) relate to work done in teams.
•Resumes are records of the past. What you care about is a candidate’s future.
•Hobbies, organizational affiliations and choice of font and screening-friendly keywords tell you nothing about someone’s ability to do the job, but might (and usually do – especially if you use resume screening software) influence your decision.
3. Resumes are plain unreliable.
•About 25% of new hires admit they gave less than accurate information during recruitment.
•Most of the other 75% did the same, only they call it “tailoring” their resume to the job – leaving out irrelevant information (like that time they got fired) and highlighting anything that might be precisely what the employer is seeking.
•The resume fails to disclose the setting of someone’s work – the economic conditions in the industry at the time, the company culture, the style and behavior of the individual’s boss. These matter.
Keyword searching, automated screening, and other whizzy technologies can’t fix the core problem because the resume does not tell us what we need to know. (Nor, to be fair, does the job post tell the candidate what s/he needs to know.)
Applicant screening technologies were created to handle huge numbers of applicants. But today we don’t need to reduce the number of applicants: technology enables us to carefully assess as many people as we want. We don’t need to “screen out” or “reduce the pile”; we need simply to do the best job of predicting performance that we possibly can.
The fundamental needs are truth and relevance: relevance in the information we collect (performance predictors) and truth in that information (valid assessments not keyword-clever tailoring of the not-particularly-relevant facts).
For employers overwhelmed with applicants, desperate to find a way to sort the best from the rest: stop reading resumes and start using assessments – valid, proven science – to predict performance. Do it now.
Valid assessments are assessments that have been demonstrated, through rigorous validation studies (always ask for validation data for any assessment under consideration) to measure what they say they are measuring. Valid assessments can and do predict performance with accuracy that is higher and usually much higher than other methods. Valid assessments give relevant information and they uncover the truth. Valid assessments do these tasks reliably and defensibly. Valid assessments are science, they are analyses. Recruitment practices not using them will remain ineffective, costly and disappointing to both employer and potential employee.
We can predict performance. Best-in-class companies know that and practice it. Many more organizations can adopt this best practice especially now that advances in science (psychometrics) and technology (the reach of the internet and the analytical power of servers) make assessments affordable and easier to implement.
What else would any employer need to know to make a hiring decision?