“…and the children are all above average.” Wouldn’t be great if you could say that about your candidates? Garrison Keillor’s famous ending to his Lake Wobegon monologues offers us an opportunity to ponder how that idea can be applied to our mental model for HR analytics and practices for staffing process improvement.
Let’s begin with the concept of average. One dictionary offers this definition: The result obtained by adding several quantities together and then dividing this total by the number of quantities. In the context of recruiting, I prefer to describe it as the result of mistakes that occur while trying to only hire top talent. Your worst hires pull your average down.
William Scherkenbach noted author and former Director of Statistical Methods for Ford Motor Company in his book The Deming Route to Quality reminds us that “If you believe in the law of averages you will always have above average and below average, and there is not a darn thing you can do about it.” In the simplest of terms: half of your candidates are below average. All the time, every time. Think about that.
So what is an above average candidate?
The law of averages surfaces as variation in your process. You hired your best, and you hired your worst, using the same candidate evaluation methods. An average is a place on the continuum between the two hires that bookmark the ends.
When we apply this to a candidate population, we have to answer two questions:
- What is the average?
- How does the average in the candidate population compare to the average in the current employee population?
At Lake Wobegon, the children are all above average. But above average on what? And Scherkenbach would balance that with an assertion that the children in the sister city, Lake Woe-is-me, are all below average.
Quality of hire carries with it a notion that there is one or more metrics against which a candidate can be evaluated. Enter productivity metrics, competency models, KSAOs, success criteria, performance frameworks, etc. These methodologies for describing and evaluating on-the-job behaviors of current employees become the metrics for evaluating candidates as well.
Scherkenbach goes on to assert the real opportunity is to know what your average is and continuously raise it. This is where the notion of above average really comes into play. It is possible for the best candidate to be below average when compared to the average of current employees. And hiring the best candidate from a below average pool lowers the average of the current employees. Think about that.
It is possible to conduct an analysis that determines average or actually a variety of averages within existing employees. Data about their performance exists or can be created. Objective measures of performance can be used to collect data on a group and calculate an average level of productivity. Supervisor ratings of observed performance can be collected using behaviorally anchored rating scales. This data can be used to calculate the average of each competency as well. These two data sets comprise the metrics for quality of hire.
Quality of hire metrics become the standard for differentiating among candidates. How then, does one gain insight into a candidate’s ability to achieve certain objective measures of performance? And, how does one ascertain a candidate’s capacity to deliver behaviors similar to those defined in a competency model?
Over time, new hire performance can be evaluated using the same criteria, objective results, and competency ratings. When that data is collected, the quality of hire evaluation can be documented and reported.
Proxy Measures and Evaluating Average
Most forms of candidate evaluation are proxy measures. A proxy measure is a substitute or surrogate measure. The most obvious and common proxy measure is level of education. A diploma or degree is often set as a threshold measure of some functional level of literacy. Assumptions are made about basic reading writing and reasoning levels associated with each level of academic attainment. I wonder if Garrison Keillor is referring to above average high school achievement in Lake Wobegon?
No doubt you have seen two individuals with similar academic credentials but with very different levels of literacy. This is an example of where proxy measures fail to do its job. At issue here is the variation that exists in the proxy measure, or how abstract the proxy measure is in relation to the quality of hire metric.
Behavioral interviewing is a form of proxy measure that assumes storytelling about past behaviors relate to what we might expect in terms of behaviors in the future. And there is a large degree of truth to that.
There are two inherent challenges in achieving an effective evaluation with interviewing. One is interviewer skillfulness at probing and documenting responses in a useful manner. The second is a candidate’s ability to articulate how they accomplish results.
Beginning with thoughtfully constructed questions that elicit job-relevant examples of past efforts is considered a best practice. Recruiters who stay on-script with competency-based, behavioral interviews do get a pretty thorough evaluation of candidate-job fit. Candidate responses can be rated against evaluation criteria and ratings can be used to determine if the candidate is below, at, or above average. Unfortunately, only about 35% of recruiters stated they apply this level of rigor to interviewing practices.
There are more reliable and direct ways to evaluate and identify above average.
Direct Evaluation of Average
The closer the evaluation exercise is to actual job demands, the more accurate it can be in assessing capabilities in relation to an average. That is precisely the role of simulations for pre-employment assessment. Job specific simulations present the candidate with work scenarios and job-relevant work exercises to capture data on how an individual actually handles a range of job demands.
Having a large group of current employees complete a simulation captures the data to document in-house average. With these two data sets, it is easy to differentiate among candidates. Candidate results on the simulation produce a score that can be compared to other candidates and to existing employees. The score on simulations can help identify candidates with above average capabilities.
Having all applicants for a job complete a simulation provides the data to document candidate pool average, as well as the variation from low to high.
Handsome, Good Looking, and Above Average
Garrison Keilor sends the listener on their way with a closing comment about characteristics of the entire population at Lake Wobegon. I’d like to send you off with a few closing comments about characteristics of above average recruiting practices.
You can invest your interview time and effort with above-average candidates. But you must first invest in the rigor of defining the metrics, creating an objective evaluation method and calibrating it with your existing population. Using off-the-shelf evaluation resources may be a good step in the right direction. However by default that implements a ‘me too’ approach to candidate evaluation when in fact you may be working hard to create a workforce that is distinctive. And may not contribute to a competitive advantage.
On page 10 of The Differentiated Workforce, authors Beatty, Becker, and Huselid provide a visual model for considering the need for and strategic impact of job specific, company-specific workforce practices. One might draw a sound conclusion that certain jobs in your organization demand an extremely rigorous candidate evaluation experience.
Calibration is the business term for validation analysis. Validation analysis documents the relationship between competency ratings, objective performance metrics, and simulation results. Validation is the measurement rigor that links candidate evaluation to your business drivers.
Recruiting departments that use HR Analytics and conduct in-house validation analysis go beyond above average, they ensure their efforts create a workforce that delivers superior results. That by itself makes them pretty good looking to the CEO. If you want to do that, we can help.