The Candidate Volume Fallacy – A Mathematical Debunking
Everyone can be forgiven for thinking candidate volume from a paid job posting matters. It sounds like it should matter. More candidates mean a higher chance of success, right? Just like if you’ve lost 10 times in a row on the roulette wheel, you’re more likely to win the next spin because you can’t lose forever…
Just one problem, neither is true. Here’s the math to prove it.
Let’s keep on the roulette example. Each spin of a roulette wheel is called an “independent event”. That’s nerd for: one spin has absolutely no impact on any other spin. It doesn’t matter if you won or lost the last spin. Each spin is its own event.
You can probably see where I’m going with this…
Each candidate is also an independent event because each person is a separate person. Having me apply to a job makes no one else more qualified to get the job. Think about it. If I tried out to play professional football, my tryout doesn’t make Tom Brady a better or worse quarterback.
Where this gets confusing is that it is also true that you can’t lose on every spin of a roulette wheel forever. Each spin of the roulette wheel does have a probability of success that you can calculate. Just like each randomized candidate has some chance of getting the job. If you’ve studied statistics, you’re familiar with the term “standard deviation” which basically describes the idea that there is a probability that the “odds” will do what we expect them to over a given period of time. For example, flipping “heads” on a coin has a probability of 50%. With a high probability like 50%, we get a standard deviation of 3.5 on 50 coin flips. So, 68% of the time we’ll see 21.5 to 28.5 heads. 90% of the time we’ll see 18-32 heads.
For things that have a lower probability, the standard deviation has a much bigger impact. A single number on a roulette wheel, let’s say 23 (for Michael Jordan), has a 2.6% chance of hitting with a standard deviation of 1.13 on 50 spins. So, 68% of the time we’ll see 0.17 to 2.43 hits of 23 over 50 spins. 90% of the time, zero hits will happen commonly.
For a paid job posting on a job board, anyone can apply so each candidate application should be considered “randomized”. And the odds of a randomized candidate getting a highly skilled job is very low. I would argue far worse odds than betting 23 on a roulette wheel. So it’s not only easy to get 50 candidates and hire none of them, but it’s actually a normal outcome. That’s why this has probably happened to you if you’ve done enough hiring. You need hundreds of candidates before you can say, “I’ll be surprised if a hire isn’t within this huge candidate pool”.
That’s why candidate referrals are so important and yet still undervalued. The odds of a candidate referral getting hired are much higher because they aren’t randomized candidates. They have a selection bias that favors getting hired, but I’ll save that for another post.
So how do you then evaluate potential vendors for job postings? After all, job postings are still a low-cost option for sourcing candidates when compared to dozens of labor hours sourcing candidates.
If you’re already using a set of job posting vendors, look at the number of candidates that have made it to at least the second round of your interview process, then divide by the cost of that job posting. That’s your per unit price for a relevant candidate. For example, if you had 1 candidate get a second interview, and 2 others made it to the offer stage, that’s 3 candidates that made the second interview round or later. If you paid $300 for the job posting, you paid $100 per relevant candidate. ($300 / 3 candidates = $100 per relevant candidate)
If you’re looking at new job posting vendors, what you should be looking for are page views and job shares. Aim for higher page views and more job sharing. Why? Remember I said you would need hundreds of randomized candidates for a hire? Well every page view is kind of like a potential candidate and page views are priced far cheaper than candidate applications. This allows you to more cost effectively select your vendors and drive better results. As an example, a posting vendor that gives you 50 page views and 50 candidate applications is very bad. You want something closer to 500-1000 page views, regardless of candidate applications. If the vendor targets your target employee audience that’s even better.
Job sharing is also hugely important because firstly, it leads to more page views and secondly, it also leads to more page views in your target employee audience. People in tech are more likely to share jobs in tech and less likely to share jobs in hospitality (and vice versa). Increased job sharing will drastically improve your average candidate quality and help you reach passive candidates not accessible through most job boards. So, look for job posting vendors that can help you increase your job sharing totals and increase your page views.
Don’t bet your hiring plan on vendors that have given you a lot of candidates you weren’t interested in, but seemed “due” to pay off.