Latest / The Science of Leading / The Best Ways to Predict Job Success

Transcript

The Best Ways to Predict Job Success · The Science of Leading

• 0:00
  Hey everyone, welcome back to The Science of Leading.
• 0:02
  I'm Claire Monroe-and I'm here with my co-host and resident organizational
• 0:06
  wizard, Edwin Carrington.
• 0:06
  Today we're digging into a question that honestly
• 0:11
  just... never dies.
• 0:11
  Like-what actually works when you're hiring?
• 0:11
  Because I've
• 0:17
  definitely thought I knew the answer before-and,
• 0:20
  uh, turns out... I didn't.
• 0:20
  You're not alone,
• 0:23
  Claire.
• 0:24
  It's one of those questions that sparks a lot of strong opinions-most
• 0:28
  of them built on gut feel, not data.
• 0:28
  And that's where the trouble usually
• 0:32
  starts.
• 0:32
  Yeah.
• 0:34
  I mean, early in my career, I followed the "default script."Stacks of
• 0:38
  resumes, a few glowing reference calls,
• 0:40
  and then-ta-da-the infamous unstructured interview.
• 0:40
  And if someone made
• 0:44
  eye contact and had a firm handshake?
• 0:46
  I was sold.
• 0:46
  Then fast-forward a few months...
• 0:50
  and I'd be like, "Wait-how is this person totally lost in the role?"The
• 0:56
  halo effect is brutal.
• 0:56
  It is.
• 0:58
  But it's also still the norm.
• 0:58
  Most companies default to resumes,
• 1:04
  references, and open-ended interviews-because they feel comfortable.
• 1:08
  Familiar.
• 1:08
  But when you dig into the research... they just don't hold up.
• 1:08
  Unstructured
• 1:15
  interviews?
• 1:16
  Only predict performance about 14 percent of the time.
• 1:16
  Reference checks-closer
• 1:22
  to 7.
• 1:22
  And resumes?
• 1:25
  A shocking number-up to 75 percent-have misleading or padded info.
• 1:25
  But
• 1:31
  we keep using them because they feel like we're making smart decisions.
• 1:35
  Even when we're not.
• 1:35
  And there's that weird assumption,
• 1:38
  right?
• 1:38
  Like-if someone writes a slick resume or totally crushes the interview...
• 1:43
  they must be great at the actual job.
• 1:43
  But I've seen people nail every
• 1:47
  step of the hiring process-then completely stall out once they're hired.
• 1:47
  Exactly.
• 1:52
  What we're really measuring is presentation,
• 1:55
  not performance.
• 1:55
  And that opens the door to bias.
• 1:59
  You end up hiring people who look the part-on paper or in person-but
• 2:04
  not necessarily those who can do the part.
• 2:04
  It filters out anyone who doesn't
• 2:09
  fit the expected mold... even if they'd excel with the right opportunity.
• 2:09
  So
• 2:15
  you're not just gambling on the wrong signals-you're shutting out some
• 2:18
  amazing people in the process.
• 2:18
  That's a lose-lose situation.
• 2:18
  It is.
• 2:23
  It feels efficient-but the cost shows up later.
• 2:23
  Turnover,
• 2:28
  misalignment, underperformance...It all adds up.
• 2:28
  That's why the comfort
• 2:34
  of traditional hiring can be so misleading.
• 2:34
  It's fast upfront-but expensive
• 2:39
  in the long run.
• 2:39
  Alright-let's shift gears.
• 2:42
  Because this is getting a little bleak.
• 2:42
  What does work?
• 2:46
  I know there's a whole body of research around better predictors.
• 2:46
  There
• 2:50
  is-and the contrast is pretty dramatic.
• 2:50
  Take work sample tests-where you
• 2:56
  have candidates actually do a task from the job.
• 2:59
  That can predict performance with up to 85 percent accuracy.
• 2:59
  Structured
• 3:04
  interviews-where questions are consistent and responses are rated against
• 3:08
  a rubric?
• 3:09
  Around 57 percent.
• 3:09
  And cognitive or skills assessments?
• 3:14
  You're looking at about 65 percent.
• 3:14
  It's not even close.
• 3:14
  Okay but-when
• 3:20
  you say "work sample," are we talking like,
• 3:22
  a giant test?
• 3:22
  Because I think some people hear that and imagine... 200 questions
• 3:26
  and a scantron.
• 3:26
  Ha-no, not at all.
• 3:26
  A good work sample is just a mini version
• 3:31
  of the job.
• 3:31
  Say you're hiring a marketing analyst-you might give them a
• 3:35
  data set and ask for a short insights summary.
• 3:35
  If it's a frontline support
• 3:40
  role, maybe it's a mock customer scenario.
• 3:40
  I worked with a manufacturing
• 3:46
  company that switched to hands-on tests for machine operators-simple,
• 3:50
  role-specific tasks.
• 3:51
  Within a year, their turnover dropped by 20 percent.
• 3:51
  Whoa-20 percent?
• 3:57
  That's huge.
• 3:57
  And structured interviews-those are the ones where it's not
• 4:01
  just "so tell me about yourself," right?
• 4:02
  It's more like, "Walk me through a time you handled XYZ."Exactly.
• 4:02
  It's
• 4:08
  about consistency-everyone gets the same questions,
• 4:12
  and you judge answers against the same standard.
• 4:12
  It reduces bias,
• 4:16
  improves fairness, and gives you something defensible if decisions are
• 4:20
  ever challenged.
• 4:20
  And cognitive assessments?
• 4:23
  They're especially useful when you're hiring for potential-not just experience.
• 4:28
  They help you see how someone thinks and learns.
• 4:28
  Okay,
• 4:32
  okay-but we have to talk about AI.
• 4:32
  Last episode we went deep on automation-so
• 4:37
  I'm curious... is AI screening actually helpful?
• 4:41
  Or is it just another barrier between candidates and humans?
• 4:41
  It can help-if
• 4:46
  it's done right.
• 4:46
  AI screening can speed up shortlisting by 30 percent,
• 4:50
  and it can reduce certain types of bias-especially when it focuses on
• 4:55
  skills and behaviors instead of keywords or gut feel.
• 4:55
  But here's the risk:
• 5:00
  if your AI is trained on past hiring data full of bias... it just automates
• 5:05
  that bias faster.
• 5:05
  So you have to monitor it.
• 5:09
  Audit it.
• 5:09
  AI should assist-not replace-human judgment.
• 5:09
  So no magic wand.
• 5:09
  Just
• 5:17
  a really fast assistant who still needs oversight.
• 5:17
  Exactly.
• 5:17
  Used well,
• 5:21
  it saves time and adds structure.
• 5:21
  But the real gains come when you pair
• 5:26
  AI with solid assessments and structured interviews.
• 5:26
  That's when the whole
• 5:30
  system starts to work together.
• 5:30
  Okay-let's say someone's listening right
• 5:34
  now and they're like, "Yep.
• 5:35
  That's me.
• 5:36
  Time to rethink how we hire."Where do they even start?
• 5:36
  Start with what's
• 5:41
  most predictive.
• 5:41
  Build in assessments-work samples,
• 5:45
  role-relevant cognitive tools,
• 5:47
  even behavioral profiling.
• 5:47
  Then layer in structured interviews with clear
• 5:52
  scoring guides.
• 5:52
  Use automation to reduce admin burden,
• 5:56
  but keep humans in the loop.
• 5:56
  And finally,
• 5:59
  track your metrics: time to hire,
• 6:01
  retention, role fit.
• 6:03
  That's your learning loop.
• 6:03
  I love that idea of a loop.
• 6:03
  But be honest-how
• 6:09
  do you get an old-school hiring manager to even try this?
• 6:09
  Like,
• 6:13
  someone who swears they can "spot talent" in five minutes?
• 6:13
  You don't start
• 6:18
  by arguing-you start by testing.
• 6:18
  Run a pilot with one team.
• 6:23
  Show them the outcomes.
• 6:23
  Better retention,
• 6:26
  smoother onboarding, fewer regrets.
• 6:26
  When people see results-real ones-they
• 6:32
  become your biggest advocates.
• 6:32
  Science beats opinion when it's tied to
• 6:37
  ROI.
• 6:37
  That's the move.
• 6:37
  Don't fight the whole system.
• 6:41
  Just prove it works-one step at a time.
• 6:41
  So if you're listening and this
• 6:46
  all feels... big?
• 6:46
  Don't worry about a full overhaul.
• 6:46
  Pick one place to start.
• 6:52
  Maybe it's a work sample.
• 6:53
  Maybe it's just standardizing your next round of interview questions.
• 6:53
  Track
• 6:57
  what happens.
• 6:58
  Learn from it.
• 6:58
  And build from there.
• 6:58
  Exactly.
• 6:58
  You don't have to get it perfect-you
• 7:04
  just have to get it going.
• 7:04
  Over time,
• 7:07
  your hiring becomes sharper, fairer,
• 7:09
  and a lot less stressful for everyone involved.
• 7:09
  And hey-if you're wondering
• 7:14
  how to actually do this stuff?
• 7:14
  You can try OAD's behavioral assessments
• 7:18
  for free at o-a-d-dot-a-i.
• 7:18
  They're simple,
• 7:21
  fast, and give you a better sense of fit before someone even walks in
• 7:25
  the door.
• 7:25
  Which, like-why wouldn't you want that?
• 7:25
  Couldn't agree more.
• 7:25
  This
• 7:31
  is how you build teams that thrive-not just survive.
• 7:31
  And you don't have
• 7:35
  to guess your way there.
• 7:35
  Alright,
• 7:37
  that's a wrap on this episode.
• 7:37
  Edwin,
• 7:39
  thanks as always for keeping us grounded in the science.
• 7:39
  And to all our
• 7:42
  listeners-stick with us.
• 7:44
  We've got more smart people stuff coming.
• 7:44
  Edwin...Always a pleasure,
• 7:47
  Claire.
• 7:47
  And to our listeners-take care.
• 7:50
  Keep leading with intention.
• 7:51
  And with data.
• 7:51
  We'll see you next time.