Which Personality Types Predict Sales Performance?
The broader historical study covered 10,765 insurance agents at one carrier. The model comparison used the evaluable n=229 subset with personality data. In that small subset, personality reached an AUC of 0.647, against 0.558 for keyword screening. The result does not support hiring or rejecting on a type alone.
Source: "Decision Traces," Saad Bin Shafiq, Nodes, 2026. Personality analysis on the evaluable n=229 subset with Predictive Index data at one carrier. Read it on arXiv.
What the data showed
Production rates varied across personality types in the evaluable n=229 research subset. The individual type groups were smaller still, so the public result is limited to the approved model comparison: personality reached an AUC of 0.647, while ATS keywords reached 0.558.
Why personality beat keywords
Keywords describe what a candidate has done. A personality assessment measures a different signal. In the evaluable n=229 subset at one carrier, personality reached an AUC of 0.647, above keyword screening at 0.558. See the full model comparison.
What this does not say
The personality analysis came from an evaluable n=229 subset, and the individual type groups were smaller still. Personality was not assigned at random because the assessment ran through selected locations and channels. The result needs validation on larger samples before anyone uses a type to hire or reject. Single carrier, single role.
Frequently asked questions
Does personality predict sales performance? In the evaluable n=229 subset at one carrier, personality reached an AUC of 0.647, against 0.558 for resume keywords.
Can the study rank personality types for hiring? The individual type groups were too small to support a public screening rule. The approved result concerns the model comparison, not a ranking of named types.
Should we hire only certain personality types? No. The samples per type are small, and screening hard on a single trait repeats the keyword mistake. Personality is one strong signal to fuse with others, not a filter to apply on its own.
What assessment was used? The Predictive Index, fused with behavioral scoring and applicant tracking data.
Related reading
- From AUC 0.647 to 0.735: how data fusion improves prediction
- Why too much leadership drive can hurt sales performance
- What 10,765 agents hired revealed about resume keywords
See which traits predict production in your own sales force. Request a Decision Replay.