A yield model is a great business planning tool used throughout modern organizations. It relies on basic math to calculate the net output (outputs minus rejects) from a process. For example, if you are a baker and find that 1:10 muffins produced fails to meet desired quality standards, you have a net yield of 90%. The yield model helps you as a baker realize that in order to deliver 100 high quality muffins to a customer, you’ll need to bake 12 batches.
Within the recruiting function, yield models are commonly used for two activities, projecting the number of initial applications needed to close projected requisitions, and calculating the resource requirements (labor, time, tools, etc.) that will be consumed in the process, i.e. budgeting. While not all organizations leverage a traditional yield model per se, nearly all organizations rely on some derivative including companies like Wal-Mart, Microsoft, Tenet, Humana, Wipro, Life Technologies, Textron, and Southwest Airlines.
While many recruiting leaders are aware of the role yield modeling plays in planning activities, most are unaware of the third, and perhaps the most important use, identifying recruiting process failure points. It turns out that if you look at each of the individual steps within the hiring process, you can use any decrease in the “yield” of that step as an “alert” to warn you that problems maybe occurring.
How a Yield Model Can Help You Identify Recruiting “Failure Points”
Recruiting managers and recruiters are constantly complaining about their recruiting problems, but few have any systematic way of identifying precisely at which step in the recruiting process the most impactful problems are occurring. Fortunately, a yield model can also be a helpful tool in a Failure Point Identification process.
Benchmark Pass-through Rates At Each Step
I recommend that you begin the failure point identification process by calculating the average “pass” or pass-through rate at each step in your process. You can start with benchmark data from other firms. For example, the following data covers over 130,000 applications from nine companies in five different industries.
| Process Step | Average % that “pass through” to the next step |
|---|---|
| Resume Screen | 55% |
| Phone Screen | 56% |
| Interview | 53% |
| Offer made | 73% |
| Accepted | 83% |
You can see from this chart that 56% of the candidates who undergo a phone screen successfully move on to the next step, and that 83% of the candidates extended an offer accept it. While you can certainly use these benchmark numbers as a guide, the best approach for identifying your firm’s recruiting failure points is to gather your own yield data over time. After gathering the data, you should then set a range of minimum and maximum passing percentages. Once you establish a benchmark “low passing percentage” at each step, you can then consider any passing percentage data that falls below that benchmark minimum level as an “alert” that something is wrong at that step.
For example, if your average offer acceptance rate has been 83% and it drops to below 70%, it would be wise to look further into that stage of the process to identify the root cause of the problem. Obviously a drop in offer acceptance rates could be affected by competitor actions or a drop in the unemployment rate, but it could also mean that your recruiters and hiring managers need to be retrained in how to effectively sell and close candidates.
In contrast, if you are exceeding the maximum passing percentage, it’s an opportunity to examine that stage of the process to determine if someone has discovered a new “best practice.” Of course, it’s also possible that there is a data collection or calculation problem. That possibility should also be examined. The key is to use any negative change in the pass rate at each step as a “heads-up” that a large problem could soon negatively impact the overall yield of the recruiting process. Unfortunately, the yield model cannot tell you the cause of the problem, for that, you need to use “root cause analysis,” which usually requires a combination of surveys, interviews, or focus groups.
Using a Yield Model to Calculate Applicant Volume Needed
As mentioned earlier, many organizations use yield models to help develop recruiting budgets. Assuming no changes are planned to your recruiting process, a yield model can be used to forecast the volume of applicants that will be needed to produce the volume of hires the organizations anticipates needing. When a yield model is used in this way, it is more commonly referred to as the recruiting funnel or pipeline. This representation is borrowed from sales functions which have yield modeled sales functions for centuries.
While yield models measure volume, they do not measure quality. Yield modeling assumes that effective quality control measures are in place to ensure that poor quality outputs do not advance from one step to the next. For example, if your organization has a weak recruiting process coupled with a poor employer branding effort, you may need 100 applicants merely to get a single hire, because a large number of the initial applicants will be unqualified. To ensure that quality is a component in your modeling, I recommend that you look at both raw applicant and “qualified applicant” (those who meet at least 90% of stated qualifications) volumes as the first level of your modeling.
A Yield Can Also Help You Set Realistic Expectations
Anyone who has ever managed a process can attest that a great deal of achieving execution relies upon setting realistic expectations among all process participants. I learned the value of yield models in recruiting by working with unrealistic hiring managers. One particularly optimistic hiring manager was starting up a new business unit within a Fortune 500 company and proclaimed that he would hire 100 people in a month and begin operations within 60 days. Using a yield model, I pointed out that based on his past recruiting performance, his expectations were unrealistic. This particular manager had a history of interviewing a minimum slate of 10 candidates per position and almost always required three separate one-hour interviews. Assuming no changes to his process, that meant he would need 1,000 qualified applicants in order to “yield” 100 hires. Interviewing 1,000 applicants at three hours each would require 3,000 hours of the manager’s time alone. Achieving both goals might seem possible if our “yield model” didn’t warn us that even if the manager worked full-time on interviewing (50 hours per week), it would require four months simply to cover the 3,000 hours of interviews. After being presented with this best case scenario yield model, he moved his projected opening date back from 60 days to a full year.
Final Thoughts
In my experience, a large percentage of the metrics that are used in the recruiting function have little value. Unfortunately, one that has great value but that is often underused is the yield model. At the very least, it should be used by sourcing to identify the number of qualified applicants who are needed at the beginning of the pipeline. It can also be used to ensure that realistic timelines are set for filling positions. With those foundation practices in place, you can then use yield models for failure point identification. In this capacity, it can quickly direct you to a specific point or points within your recruiting process where your problems are originating. Once you understand the broad strategic value of yield models, it’s no longer acceptable to blame the whole process or to say that you can’t find out where a problem is occurring.