The case for using SI (smart intelligence) instead of AI (artificial intelligence) is demonstrated in the chart. It summarizes how people get jobs. The first thing that’s clear is that not many people who apply directly to job postings get hired – far less than 1%. The second point that’s less clear, but as important, is that acquaintances get hired more frequently than strangers if they’re reasonably strong, even if they don’t have all the skills required, because there’s less risk and it’s more convenient.
Here are some of the big reasons acquaintances get the better deal:
- They’re seen more often if referred by a trusted person and get to the top of the resume list when referred by anyone in the company.
- They’re hired more frequently. One survey from Lever ATS indicated that once a referred candidate is interviewed the chance the person will be hired is 20% vs. only 10% for a stranger. This double-whammy increases the chance a referred person will be hired from one in 125 to one in 12!
- Their on-the-job performance is more predictable. This doesn’t mean the person is a better candidate though. They’re generally good candidates but frequently there are better candidates who applied but there too many other risk factors – temperament, style, attitude, fit and the like – preventing these other people from being hired.
- It’s a lot faster to hire referrals and acquaintances. For one thing they’re often contacted before the job requisition is formally approved. Getting someone on board more quickly is a huge advantage.
Over the years in numerous books, a new Lynda.com course, blog posts and presentations to business groups I have made the point that we should hire strangers the same way we hire acquaintances. I developed Performance-based Hiring with this idea in mind. The problem I noticed throughout my executive search career was that top people I personally knew were typically assessed improperly by people who didn’t know them at all. As a result some remarkable people were judged as incompetent and some decent people were judged as remarkable.
Here are some ideas on how to bridge the gap and evaluate strangers and acquaintances exactly the same way:
First, define the job as a series of 5-6 performance objectives. I refer to this as a performance-based job description. A typical performance objective describes the task, the action required and a measurable result. For example, it’s better to say, “Collaborate with the engineering team to develop the product specs for the XYZ product within 60 days,” rather than, “Be responsible for product marketing” or ”Must have 6-8 years of experience, an MBA in marketing and an undergraduate engineering degree”.
Second, eliminate resumes as a prerequisite to engage in a conversation. Instead have interested candidates submit a two paragraph write-up of something they’ve done that’s comparable. Let them throw a video or a sample into the mix. Only those who are qualified will actually do this extra work. This is the key to converting to a SI commonsense system from one based on AI.
Third, eliminate pre-interview screening tests. DISC and PI-like personality tests are not predictive. For one thing they only assess preferences not competencies. While they are somewhat confirming if used too early in the process they screen out passive candidates and all superior candidates who want to explore a situation before getting serious. This puts a lid on quality of hire.
Fourth, proactively control interviewer bias. Lack of job knowledge opens the door to bias, perceptions, bad judgement and intuition to become the deciding factor when interviewing candidates. One of 12 ways to reduce bias is to script the first 30 minutes of the interview. Another way is to use a well-organized panel interview with 2-3 people.
Fourth, assess accomplishments, not skills. Ask all candidates – strangers and acquaintances – to describe in detail an accomplishment that best compares to each of the performance objectives in the performance-based job description. This post (now read by over 1.5 million LinkedIn members) describes this technique.
Fifth, organize the interview and debriefing session. Here’s a simple form you can use to organize the interview around the factors that best predict on-the-job performance. By having each interviewer justify his or her ranking using evidence for each factor you’ll eliminate yes/no gladiator voting which inadvertently glorifies bias and intuition.
AI was introduced into the hiring process to solve a problem that was self-created: Letting people apply to jobs they’re not qualified to handle. Rather than eliminating the problem at the source companies have invested unnecessary resources to weed out the unqualified rather than figure out better ways to attract and assess the most qualified whether they’re strangers or acquaintances. Regardless, the above five steps are a useful short-term workaround to a problem that doesn’t even need to exist.