Measuring what great looks like

Key facts

Most companies have a competency framework, but far fewer have tested whether it describes their best performers. This study, conducted with a US-headquartered B2B software company of around 2,500 employees, tested whether Core Drivers and Core Values relate to performance in the same organization, and whether they detect the company’s own seven competencies.

• The data: 511 manager-rated employees across 290 managers, drawn from a stratified sample of roughly one employee in five across four job families and five career levels. 473 had complete data across all three sources. Performance ratings came from HR records.
• The main takeaway: Driven, a Core Drivers disposition, was the clearest signal, at 1.64x the odds of a higher performance tier per standard deviation. Great performers rose from 21% among the lowest Driven scorers to 48% among the highest.
• What it is for: hiring, development, promotion and internal mobility, especially where nobody has yet seen the person do the job.
• The main limit: the study is concurrent, so it shows how the assessments relate to performance today, not how they forecast it. Behavior and performance ratings also came from the same manager, so a general impression of someone could account for part of the overlap.

How we did it

We built a manager survey of the organization’s seven competencies, written against its own benchmarks. Employees completed Core Drivers and Core Values, about 15 minutes in total. Their managers rated them on the competencies, with neither able to see the other’s answers. We then tested whether the assessments related to the existing annual performance rating, and whether they detected the competencies themselves.

What we found

• The instrument holds together. Internal consistency ran from 0.66 to 0.84 across the seven competencies, against a working standard of about 0.70. Employees with higher competency scores were roughly twice as likely to be rated high performers.
• Close to what managers see. The same split on manager-rated competencies gave 16%, 33% and 55%. The assessment gets close without anyone having observed the person.
• Driven runs through every confirmed link. It appears in all four competencies the assessments detected: Critical Thinking (.124), Accountability (.123), Tolerance for Ambiguity (.117) and Growth Mindset (.101). The reliability threshold at this sample size is about .09. For Interpersonal Influence, Team Work and Learning Agility the signal was inconsistent.
• Not just a proxy for what managers see. Driven still predicts performance with manager-observed behavior in the model.

What this means for other organizations

• Test your framework. Build a measure of it, test it against performance data the business already holds, then see what assessment adds on top.
• Use assessment when there’s no track record to observe, such as external candidates and internal moves into new work. It measures how someone is inclined to work, not what they have been seen doing.
• Match the method to the competency. Interpersonal Influence and Team Work need interviews, situational judgment content or work samples, and Learning Agility is better measured by a reasoning assessment.
• Test the next cycle. The scheduled next step is to test against the next performance cycle, and a multi-rater evaluation would settle the same-manager question.

Key takeaways