Hiring in an Unfair Game: The Moneyball of Recruiting
June 20, 2012
Paul Basile, CEO Matchpoint Careers, Inc [email protected]
Introductions
POLL 1: who are we?
• In-house talent acquisition specialist • Talent management specialist • HR generalist • Professional recruiter • None of the above
POLL 2: Moneyball and me
• I know the Moneyball story very well • I know Moneyball, but what’s the link to recruitment? • I have heard, vaguely, of Moneyball • I hate baseball • I know who Brad Pitt is…
Why Moneyball?
Why Moneyball?
The Moneyball story
The Lessons of Moneyball
The Lessons of Moneyball
The Lessons of Moneyball
What’s the Moneyball of Recruiting? 1. Identify the problem to solve 2. Find the solution to that problem 3. Act on what we know: implement the solution 4. Eliminate everything else
1. What’s the problem?
What we want What we can afford
The difference top performers make
Top 15% of performers
Bottom 15% of performers
$48,000 $80,000 $112,000
↑↓ 40+%
How do we recruit now?
How do we recruit now?
How do we recruit now?
Our results
Our results
Our results
Our results
So, what’s going on?
Sele
ctio
n ap
proa
ch
Job performance
False positives
True negatives
True positives
False negatives
What we get
Sele
ctio
n ap
proa
ch
Job performance
False positives
True negatives
True positives
False negatives
What we want
Job performance
False negatives
True positives
False positives
True negatives Se
lect
ion
appr
oach
2. The solution: predict performance
0
-0.1
0.1
0.2
0.3
0.4
0.5
0.6
0.7 Correlation coefficient
Weak predictors
0
-0.1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Unstructured interviews (0.18)
Years of education (0.10) Years of job experience (0.18)
Graphology (0.02)
Age (-0.1)
Weakly predictive
Correlation coefficient
Medium predictors
0
-0.1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Knowledge of the job (0.48)
References (0.36)
Unstructured interviews (0.18)
Years of education (0.10) Years of job experience (0.18)
Graphology (0.02)
Age (-0.1)
Weakly predictive
Somewhat predictive
Personality tests (0.40)
Correlation coefficient
Strong predictors
0
-0.1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Cognitive ability tests (0.51)
Cognitive ability tests with behavioral assessment (0.67)
Knowledge of the job (0.48)
References (0.36)
Unstructured interviews (0.18)
Years of education (0.10) Years of job experience (0.18)
Graphology (0.02)
Age (-0.1)
Weakly predictive
Somewhat predictive
Powerfully predictive
Personality tests (0.40)
Correlation coefficient
Structured interviews (0.51)
Strong predictors
0
-0.1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Cognitive ability tests (0.51)
Cognitive ability tests with behavioral assessment (0.67)
Knowledge of the job (0.48)
References (0.36)
Unstructured interviews (0.18)
Years of education (0.10) Years of job experience (0.18)
Graphology (0.02)
Age (-0.1)
Weakly predictive
Somewhat predictive
Powerfully predictive
Personality tests (0.40)
Correlation coefficient
Adapted from I. Robinson and M. Smith, Personnel Selection (2001) British
Psychological Society
Structured interviews (0.51)
Predictive selection
Employee performance
• Cognitive ability • Behavior • Preferences
Differentiators
Predictive selection
Employee performance
• Cognitive ability • Behavior • Preferences • Skills
• Knowledge Differentiators
Baselines
Example: Project Oxygen
Example: Project Oxygen
Technical ability the least important success factor
3. Act: Implement the solution
Measure the job demands
Measure candidate skills & knowledge
Measure candidate cognitive ability
Measure candidate cognitive ability
Measure candidate behaviors
Measure candidate behaviors
Measure candidate preferences
Bring the data together
Job
Bring the data together
Job
Candidates
Bring the data together
Job
Candidates
Rank & shortlist
Bring the data together
Job
Candidates
Hire
Rank & shortlist
Timing
Timing
Application forms / résumés
Interviews / other assessments
Psychometric assessments
Traditional recruitment pipeline
Application forms / résumés
Interviews / other assessments
Psychometric assessments
Self-selection, employer-specific assessments
Interviews / other assessments
Psychometric assessments
Performance-predicting recruitment pipeline
Traditional recruitment pipeline
Timing
4. Eliminate all else
What to eliminate
But is it practical?
Candidate time
But is it practical?
< 1h 30 mins
Candidate time
But is it practical?
< 1h 30 mins
Candidate time Employer time
But is it practical?
< 1h 30 mins < 30 minutes
Candidate time Employer time
But is it practical?
< 1h 30 mins < 30 minutes
Candidate time Employer time Data processing time
But is it practical?
< 1h 30 mins < 30 minutes < 1 minute
Candidate time Employer time Data processing time
But is it practical?
< 1h 30 mins < 30 minutes < 1 minute
Candidate time Employer time Data processing time
$$$ Current cost-per-shortlist
But is it practical?
< 1h 30 mins < 30 minutes < 1 minute
Candidate time Employer time Data processing time
$$$ Future cost-per-shortlist
Results
Results
Results
Results
Recap: Winning an Unfair Game 1. Identify the problem to solve:
- Get performance and get it reliably and affordably, while competing with richer, bigger players.
Recap: Winning an Unfair Game 1. Identify the problem to solve:
- Get performance and get it reliably and affordably, while competing with richer, bigger players.
2. Find the solution to that problem - Predict performance; we know how to do it; the best
companies do it – but don’t try to compete on their terms
Recap: Winning an Unfair Game 1. Identify the problem to solve:
- Get performance and get it reliably and affordably, while competing with richer, bigger players.
2. Find the solution to that problem - Predict performance; we know how to do it; the best
companies do it – but don’t try to compete on their terms
3. Act on what we know: implement the solution - Assess people, assess jobs; use technology to do this quicker
and cheaper than traditional means
Recap: Winning an Unfair Game 1. Identify the problem to solve:
- Get performance and get it reliably and affordably, while competing with richer, bigger players.
2. Find the solution to that problem - Predict performance; we know how to do it; the best
companies do it – but don’t try to compete on their terms
3. Act on what we know: implement the solution - Assess people, assess jobs; use technology to do this quicker
and cheaper than traditional means
4. Eliminate everything else - This is the hardest part; not done, you will lose an unfair game