Fight Recruiting Spam Through Measurement and Modeling

Features Recruiter.com

Recently, Ninh Tran and I had a discussion on the topic of recruiting spam. The struggle is real, folks – not just in our industry, but for the whole business landscape.

Continue Reading Below

We cannot fix the problem with other industries, but we have to try in ours. The conversation Tran and I had brought us to the conclusion that we could use big data to take on the challenge of spam. If you're interested in this mission as well, please join us.

Let's start by first defining what spam is. We are not talking about the tasty, salty treat that you ate with eggs as a kid, but the email clogging up your inbox – and, if you are like me, draining your phone's battery life. The definition of spamming, according to Wikipedia, is "the use of electronic messaging systems to send an unsolicited message (spam), especially advertising, as well as sending messages repeatedly on the same site."

The key component of that statement is "unsolicited message." The major problem of spam is that we did not ask to be emailed but still are. Sometimes, people sign up for a hosted webinar about a particular topic, and by doing so, they are saying that they would be okay with being emailed product announcements from the vendor that hosted the conversation. I just wanted to clear that up, as earned interest is one of the requirements that makes a message solicited.

Spam accounts for 14.5 billion messages globally per day, and it makes up roughly 45 percent of all emails. If you are surprised by these numbers, you don't check your spam box very much.

Ironically, we're still producing spam emails every day, intentionally or not. We don't have the statistics on how much spam is generated by recruiters specifically, but we do know that recruiting spam has been a serious problem in our community. Just look at the#FightSpam group on Facebook run by Steve Levy.

Continue Reading Below

As recruiters, we should look at the data in a new way so that we can learn from our own best practices.

But enough from me – let us hear from the COO and cofounder of Hiretual, Ninh Tran. This is going to be cerebral, so Aaron Lintz, Glen Cathey, Stacey Zapar, etc. get to geek out. Me, I am taking a nap.

[caption id="attachment_89696" align="aligncenter" width="975"] Fig. 1: Global spam volume as percentage of total email traffic from January 2014 to June 2016, by month (Source: Statista)[/caption]

From Ninh Tran:

Before we can come up with any solutions, we should dig deeply into the reasons why recruiting spam happens. Derek Zeller summarized the hard truth about recruiting spam in this article. The most insightful part to me is the statement, "If someone doesn't realize they have a problem, they will never even try to fix it."

The problems of spam are often attributed to having a wrong mindset like "spam works" or a mixed bag of "lack of training, bad leadership, or people just being lazy," as presented by Allison Kruse and Glen Cathey at SourceCon.

When I dig a bit deeper, I find that those who spam often don't have to deal with the repercussions of spamming, like a damaged employer brand or an angry mob of people (Google "recruiters are" to see how people feel about the profession). Imagine if every spam message we sent would cost us a couple minutes of our lives – would we spend a little time personalizing our messages? I would.

Fortunately and unfortunately, that's just my imagination.

What I have found is that people are not fighting spam because they are not asking the right questions or not questioning their methods at all. I get it – spam is a complex problem, and complex problems don't have simple and clear answers. But that shouldn't stop us from trying to ask the right questions – which just might do the trick and point us to many viable solutions.

Most of the time, we all totally agree that we should fight spam. However, we rarely know if our own recruiting has been spammy or not. We have no standards by which to measure our email recruiting performance beyond the uninsightful response, acceptance, open, and click rates. How do you measure your spamminess? Do you actually mark down how many unsubscribes or angry responses you get? Does that lead to visible changes in our recruiting processes?

A fundamental approach to solving our spam problem involves addressing the lack of insightful measurements and effective models to guide us. What follows is a deep analysis of email recruiting strategies. The model I propose is general and practical to measure the performance of email recruiting. My hope is it could be helpful to our recruiting community.

Measuring the Impact of Email Recruiting

Here are some variable we should define first:

X: the number of emails you send.

Y: the output or benefits you get. This includes people who apply and engagements you have, etc.

Z: the costs you pay for spam. This includes the reputation/brand image downgrade, the higher probability of being marked as spam, etc.

W: the total profit of email recruiting. If the value is negative, the email recruiting fails; otherwise, the email recruiting is successful. The larger the number, the better.

The goal of our email recruiting is to maximize profit (W). We know that profit is the difference between benefits and costs. Therefore,

Equation 1: W = Y - Z

To simplify the model, let's assume two things:

First, that the benefits (Y) linearly depend on the number of emails sent (X).

Equation 2: Y = A * X, A is a constant (0 A 1)

Second, that the costs of spamming (Z) linearly depend on the number of emails sent (X).

Equation 3: Z = B * X, B is a constant (0 B 1)

If we put these 3 equations together, the profit (W) linearly correlates to the number of emails sent X as well.

Equation 4: W = (A - B) * X

Let's take a concrete example:

If A is 0.01, B is 0.1, and X is 500, then W will be -45. Your email recruiting strategy fails.

If A is 0.1, B is 0.05, and X is 500, then W will be 25. Your email recruiting strategy succeeds.

Whether the W is positive or not really depends on the difference between A and B. Let's define A and B in the following way:

A: coefficient of great email recruiting strategy.

B: coefficient of bad email recruiting strategy, which is dependent on A and C.

C: bounce rate ~ the sum of unsubscribe, failed delivery, and spam mark rate.

A and B vary for different strategy designs. They are correlated as well. If the email is personalized, strongly relevant, and full of context, the value of A will increase and the value of B will decrease. If the email is automated, seen by recipients in bulk, and lacks relevance or context, then the value of B will increase and the value of A will decrease.

Every recruiting team that is going to implement an email strategy should measure the coefficients of A and B. The most straightforward way to measure value for A is to calculate the good response rate (A) from favorable responses. For spamming rate (B), use this formula as an example: B = (1 - A) / 3 + C, where C is the bounce rate, the sum of spam mark rate, unsubscribe rate, and failed delivery. Solving for A using the formula for B and substituting everything into equation 4, we get:

Equation 5: W = ( 4A - 3C - 1 ) / 3 * X

My response rate A is, on average, 40 percent, and I use Hiretual to get an average bounce rate (C) of 10 percent. Suppose that for X, I send out 100 emails per day. Therefore, my email recruiting profit (W) will be 10 per day. In this case, the more emails I send, the more profit I get.

In case you're wondering, here's why InMails are generally considered spammy: A ~ 20 percent, C ~ 20 percent, X = 100, then W ~ -27.

A lot of us will see a 40 percent response rate as really high, and many will wonder how to achieve such feats. That being said, it is nowhere near the response rates of the legends like Mike Chuidian, who averages 84.41 percent. Mike and recruiters like him embody our collective hope to fight spam.

There is a method to this madness – just ask the right questions.

[caption id="attachment_89697" align="aligncenter" width="975"] Fig. 2: General reference of response rate and bounce rate for email recruiting[/caption]

Parting Thoughts

Spam hurts everyone. Spam works against us, our companies, the people receiving it, and our whole industry. As sourcers and recruiters, we should be very careful with our email recruiting strategies. Coefficients of good strategy and bad strategy are the key performance indicators. Monitor your email recruiting performance by referring to the model in equation 5. Make sure that you send good emails that engage people, and you will see more and better results in your work.

Derek Zeller draws from more than 18 years in the recruiting industry. For the last 13 years, Derek has been involved with federal government recruiting, specializing within the cleared IT space. Currently, he is senior recruiting lead at ComScore and advising the team at Hiretual. You can reach Derek here.

Ninh Tran is the COO and cofounder of Hiretual, a sourcing platform that combines recruiting with science to make the internet to recruiter-friendly. You can reach Ninh here or here.