Hiring is among the highest-stakes decisions organisations make and is frequently made using methods that research shows predict poorly.
What research measures
How well a selection method predicts subsequent job performance.
Which has been studied extensively across decades.
Meta-analyses comparing methods are published and have been revised as data improved.
Unstructured interviews
Conversations without a consistent format.
Which are the most widely used and among the weaker predictors.
Interviewers form impressions quickly and then seek confirming information.
Structured interviews
Consistent questions, scored against defined criteria.
Which predict substantially better than unstructured ones.
The improvement comes from consistency and from scoring rather than from the questions themselves.
Work sample tests
Candidates performing tasks resembling the actual job.
Which are among the better predictors.
They also give candidates a realistic view of the work.
Cognitive ability
Consistently among the strongest predictors in research, and legally constrained in some jurisdictions.
Which produces adverse impact concerns requiring careful validation.
Job-relatedness and validation are the legal requirements where these are used.
Cultural fit
A widely used criterion with poor definition.
Which frequently operates as similarity to existing employees.
Defining specific behaviours rather than fit addresses this.
Reference checking
Weak predictive validity as commonly practised.
Which improves with structured questions about specific behaviours.
Legal caution has made many employers unwilling to say anything substantive.
What improves outcomes
Define the job specifically, use structured methods, include a work sample, score independently before discussing, and be honest about what the role actually involves.
Job descriptions
Long lists of requirements that few candidates meet.
Which reduce applications from qualified people, particularly from underrepresented groups.
Distinguishing essential from desirable substantially widens the pool.
Candidate experience
Long processes, poor communication and unexplained rejection.
Which affects reputation and future applications.
Candidates talk to each other and to their networks.
Take-home exercises
Work samples completed independently.
Which are informative and disadvantage candidates with limited time.
Keeping them short and paying for substantial ones addresses this.
Panel and scoring practice
Independent scoring before discussion prevents anchoring.
Which is a small process change with measurable effect.
Onboarding
Early experience predicts retention substantially.
Which makes it part of hiring rather than separate from it.
Internal mobility
Filling roles from existing employees.
Which is faster, cheaper and carries known information about performance.
Organisations frequently overlook internal candidates for roles they could do.
Diversity of the pipeline
Where candidates are sourced determines who applies.
Which is upstream of any selection decision.
Referral-heavy hiring reproduces existing composition.
Probation and early management
Addressing performance concerns early.
Which is uncomfortable and considerably better than the alternative.
Most managers describe waiting too long as their consistent error.
Documented decisions
Recording the basis for hiring decisions.
Which supports consistency and is relevant if a decision is challenged.
The practical improvements
Structure the interview, add a work sample, score independently, and be honest about the role.
Unstructured interviews
Conversations that vary by interviewer and candidate.
Which the research consistently finds weakly predictive.
They mainly measure how comfortable the interviewer felt.
Structured alternatives
The same questions, in the same order, scored against defined criteria.
Which predicts substantially better and is less pleasant to conduct.
Work sample tests perform comparably well.
Consensus effects
Group discussion converging on the first opinion expressed.
Which loses the value of independent assessment.
Collecting written scores before discussion preserves it.
The cost of a bad hire
Recruitment, onboarding, lost output and team effects.
Which is commonly estimated at several months of salary and is probably understated.
Job descriptions
Vague or inflated requirements.
Which deter qualified candidates and attract unsuitable ones.
Long required-skills lists reduce applications from underrepresented groups particularly.
Reference checks
Contacting previous employers.
Which produces limited information given legal caution.
Specific questions produce more than general ones.
Speed
Long processes lose good candidates.
Which is a real cost that thorough processes must balance.
Structured does not have to mean slow.
Bias
Similarity, appearance and background influencing assessment.
Which is well documented and not eliminated by awareness alone.
Structure and defined criteria reduce it more than training does.
Where to start
Define the criteria before seeing candidates and score against them independently.
Which costs nothing and is the single largest improvement available.
Everything else is refinement on top of that.
A closing caution
None of this is prescriptive. Businesses differ by sector, by scale and by stage, and practices that work well in one context fail in another for reasons that are not always visible from outside.
What is consistent is that the businesses handling these questions well tend to have written something down, measured it in a defined way, and reviewed it on a schedule rather than when a problem forces the issue.
Where a decision carries legal, tax or employment consequences, professional advice specific to your jurisdiction is worth the cost, and this article is general description rather than advice.