Our Managing Director, Noah Roth, was recently featured in Money Inc discussing the effects of using the wrong recruiting metrics. Below is that post in its entirety:
Malcolm Gladwell has written extensively about what he calls “the mismatch problem.” We design tests to predict performance. The tests are based on common sense inferences about a correlation between a test and an outcome. But we usually fail to check if the tests are good predictors of real life performance. And they often are not.
- The LSAT does not predict success in law school, but law schools still require it.
- The Wonderlic test does not predict quarterback success, but NFL scouts still use it.
- Teacher certification does not predict the ability to motivate students or convey information to them, yet all 50 states have licensing standards.
Metrics are important. Unless you are measuring both your performance, and the performance of your vendors, you can’t track progress or improvement.
But what happens when you measure the wrong things?
Like the examples above, you probably wind up improving on what you measure, but not necessarily on your core objective.
It seems obvious that if you retain an executive search firm, you don’t want to waste your time interviewing irrelevant candidates. You want your vendor to deliver a small number of highly targeted, highly motivated candidates. That’s why, on the surface, the ratio of candidate submissions to accepted offers seems like a meaningful metric.
But are you actually making your search more effective? There are several reasons why it may not.
- Efficiency and effectiveness are often inversely proportional. A perfectly efficient search would result in hiring the first candidate submitted. A perfectly effective search involves hiring the most qualified available candidate. The best candidate is not always- or even often- the first candidate submitted. Would you rather hire the best candidate or the first one?
- The best candidate is probably currently employed. If you judge your recruiting partner based on submission-to-hire-ratio, s/he has incentive to only share candidates who are actively looking- candidates who are currently unemployed, or underemployed. Wouldn’t you prefer to encourage your recruiting partner to swing for the fences. They may strike out on a candidate who does not ultimately accept your offer. They might also hit a homerun.
- A good recruiter will give you what asked for. A great recruiter will show you what you need. Many hiring managers want to hire someone who has already done the job they are hiring for. If a candidate has already done this job and has career ambitions, the odds are that he or she wants to grow into another role, not do the same job again. Other hiring managers believe that specific experiences are a proxy for the ability to grow into the role for which they are hiring. As an example, having a series 7 is required for a broker. Does it mean that a candidate can build an excel model? Or manage a client relationship? Are there enough candidates who have a series 7, modeling, and client management experience in Sioux Falls, ID, or are you asking for a profile that doesn’t exist? Many hiring managers are working off of a hunch. Great recruiters are leveraging data and experience.
I’m not suggesting getting inundated with irrelevant resumes. That is neither efficient, nor very likely to be effective.
A recruiter who is being evaluated by submission-to-placement-ratio is like a waitress at a diner. If you order something off of the menu, he’ll bring it to you, if they haven’t run out. Don’t expect a gourmet meal, or garnishing tips. If you want to order off of the menu, ask someone in your HR department to inmail every candidate on LinkedIn matching a specific profile, and then cross your fingers.
A great recruiter will challenge your assumptions, back up recommendations with data, bring you passive candidates, and deliver on the experience of helping people make quality hires every single day. You may not make a hire for every 3 submissions. But you’ll get the best candidate out there.