Metrics matter, and the best metrics matter more.

Back in 1982, one of my search clients (a Fortune 200 company) revealed the results of a two-year study regarding the source of their best employees. Less than 5% came from people applying to job postings. One-third came from their university hiring program, and the balance were people who were referred by an employee or direct sourced by a highly-networked recruiter. More recently, I heard about a similar multi-year study at another Fortune 100 company. The results were the same.

Conclusion: Referrals (A) always drive quality of hire, and if hiring managers see more referrals they won’t need to interview as many candidates (B).

Now consider the following analytics rule:

If A and B predict C, then track A and B to ensure C.

Substitute sourcing mix and sendouts per hire for A and B and Quality of Hire for C.

If that’s the case then sourcing mix and sendouts per hire need to be monitored consistently. One way is to calculate the ratio of referred or direct sourced candidates presented to the hiring manager on any assignment to those who applied directly to a job posting. A 2:1 ratio is a good start. This means two out of every three candidates are either referred or direct sourced.

Quality of hire problems arise whenever the ratio falls below 50/50 or 1:1. This means the recruiter is spending too much time sourcing active candidates. This is revealed by another metric: candidates per hire. Whenever a hiring manager needs to see more than four candidates to make a hiring decision, it means the recruiter is on fishing expedition – sending out too many candidates in the hope one fits.

The cause here is less clear. It could be the hiring manager doesn’t know the job, can’t attract good people or is overemphasizing skills and experiences as a prerequisite. Alternatively, it could mean the recruiter is working too many requisitions, is not spending enough time networking or is not strong at recruiting passive candidates. It could also mean there’s a problem with the company or the job. Regardless of the cause of the problem it’s revealed by an excessive number candidates needed to be seen to make one hire. So rather than send out more candidates, stop the process and figure out the cause.

Here’s some troubleshooting advice so that you can have fewer sendouts per hire:

1. Examine the sourcing mix of the slate of candidates presented.

If it’s not at least 50% this is likely the problem. In this case you’ll need to shift your emphasis to obtaining more high quality employee referrals and more direct sourced passive candidates.

2. Is the recruiter requisition load appropriate?

It’s impossible to obtain a sourcing mix of 2:1 when working more than 15 unique jobs alone at any one time. In this case, recruiters will need to concentrate their passive candidate recruiting efforts on only the more important jobs.

3. Determine if the job is worthy.

The best people are looking for career moves, not lateral transfers. You can attribute much of your quality of hire problems to managers who over-specify skills and experiences as barriers to entry. The solution: Shift to a performance qualified assessment approach.

4. Assess the quality of your recruitment advertising.

Most recruitment advertising is designed to weed out the weak rather than attract the best. If your job descriptions are full of must-haves and generic hyperbole, you can rest assured this is part of the problem.

5. Determine if the hiring manager is the problem or the solution.

If the hiring manager won’t or can’t prepare a performance-based job description, you’ll never attract a top person. Another clue the problem is the hiring manager: No track record of hiring and developing top people. Use this advanced troubleshooting guide to help hiring managers overcome these critical barriers.

6. Figure out if your hiring process is too restrictive or unwieldy.

Track opt-rates of those who see the job post but don’t apply, those who refuse to take pre-hire questionnaires, the ratio of good candidates to weak candidates who do apply, and the number of candidates at the top of the funnel needed to make one hire. When these numbers are too high it means more time is spent weeding out the weak rather than attracting the best.

Prediction is not the same as achievement. If doing A and B consistently results in C, then you need to consistently track A and B. That’s the difference between predictive analytics and process control. While candidates per hire and sourcing mix predict quality of hire and time to fill, you need to track both for every recruiter on every assignment. This is how to ensure you’ll maximize quality of hire and reduce time to fill, rather than having to explain why you didn’t.