This report is one of the “deep-dive” analyses from the new AIM Group Recruitment Intelligence report. For more information, click here; to sign up, click here. If you sign up for six months — just $199 a month — you receive our 2021 Recruitment Marketplaces Annual (a $1,495 value) as part of the deal!
- U.S. market maturity is driving focus on quality over quantity
- Companies are taking various approaches to determining quality
- Cost per qualified applicant is next step in pricing evolution
In the still nascent world of performance-based recruitment marketing, job ad clicks count less and less every day.
It’s not that they don’t matter. They are still the transactional currency of programmatic advertising. They speak to brand power, ad writing skill, placement effectiveness and other factors that are part of a complex equation. But to put it simply, the clicks that count the most are the ones that convert to bona-fide applicants, and to quality applicants and ultimately quality hires.
Applications and hires, not clicks
As recruitment advertising becomes ever more sophisticated, the programmatic firms and their employer-clients are focusing on conversion rates and less on the volume of clicks job boards and recruitment marketplaces send. The analytics tools used by the agencies and many, if mostly large, employers are capable of tracking clicks to completed applications and even all the way to a hire.
“What these lead you to is a question of whether the cost per click actually matters. And whether the click and its validity actually matter,” explains Tom Chevalier, chief product officer with Appcast, the U.S.-based programmatic specialist that has Axel Springer as majority stakeholder.
“You don’t really want clicks anyway. You want applications and hires.”
Ultimately, said Terry Baker, CEO of programmatic advertising specialist PandoLogic, it means delivering applicants who become quality hires.
That’s the leading edge today in programmatic marketing. Nearly all the vendors are experimenting with ways to analyze candidate applications to decide who’s a qualified applicant and likely to turn into a hire.
“We’re moving down funnel in the recruitment marketing process to help drive quality applicants. If you only rely on clicks to do that, you can’t make a determination,” Baker says.
No argument there. Recruiters have struggled for years to find the diamonds in a pile of resumes. Early ATS vendors addressed the problem by keyword matching resumes to job descriptions, ranking highest those with the most matches. Sophistication improved with the introduction of machine learning, but it’s far from flawless as so many systems rely on what amounts to beefed-up keyword matching.
Focus on quality applicants driving evolution
Now programmatic vendors are getting into the act. It’s the next step in the evolution of an industry sometimes blamed for delivering large numbers of low-quality candidates.
Blame for that has to be shared with the introduction of “click to apply” buttons utilized by recruitment marketplaces. Even with knockout questions, they make it easy for a job-seeker to apply to dozens of employers in a single session.
So now, just as conversion rates — the rate at which a click from a career site converts to an application — replaced click volume as the critical metric, programmatic agencies are looking at ways of separating the application wheat from the chaff and delivering — and charging for — only qualified applicants or quality applicants to an employer.
The distinction is significant. A qualified candidate can be one matching all or some significant percentage of the requirements listed in a job description. Deciding quality, like beauty, is more difficult as it lies with the employer.
Appcast is using an employer triggering event as a proxy.
“There’s generally something that’s happening after the application for employers that are hiring at some scale that they’re doing to quantify the quality of the candidate,” said Chevalier.
That could be scheduling an interview, having the candidate take an assessment, a background check or even making an offer.
“The approach that we’ve been moving towards is one that the customer’s qualification steps are already probably okay for them. And they believe in them because they’ve implemented them. And so we try to align to a qualification that they are already in alignment with.”
Diversifying options for employers
Recruitology is experimenting with a more familiar approach to deciding who’s a qualified applicant.
The U.S.-based company, which serves small and mid-sized employers, parses incoming resumes to compare them to the job description and then offers to bill the customer only for completed applications in the 60thpercentile of relevancy, explained Roberto Angulo, founder and CEO of Recruitology.
“We’re trying to lead that trend [to pay per applicant and pay per qualified applicant], but we’re also mindful of what the employer wants,” he told us. “So if the employee wants to pay for posts, they pay per post. If they want to pay per click, they pay per click. And if they want to pay per app, then we give them the option of paying for the app.”
As do all the programmatic vendors, Recruitology breaks down the results of a campaign, showing clicks and applies and the sources of each.
Angulo shared an anonymized billing report showing clicks and applications by date and source and cost. In the sample, two applications came from Craigslist. One was charged for and the other was free because the application didn’t meet Recruitology’s quality standard.
“We apply a quality threshold, where we automatically compare application resumes against the job (using NLP and matching technology). If the match quality is not high enough for a particular job from a particular application, we won’t bill that application against the campaign,” Angulo said.
Determining quality through AI chat data
Like Recruitology, PandoLogic compares an application to the job description, but goes deeper, incorporating job taxonomy to supplement the employer’s requirements and an NLP bot’s interaction with the candidate.
“You have to have a taxonomy that does the natural language processing on the job description, builds a data model as to what the requirements are, and then measures against that data,” Baker explained.
With the recent acquisition of Wade&Wendy and it’s conversational chatbot, Baker said, “We can determine quality now based upon our direct conversational AI data and engagement with a candidate.”
“Having more data that helps, that you can quantify and that you can gather from the candidate, as well as from the team and the recruiter, to establish whether or not a particular applicant raises the bar.”
The more data the AI process has will only improve the application filtering process, Baker says.
“The actual determinant is going to be based upon that three-month, six-month, one-year review. How well did this candidate perform? The performance metric is going to be the actual determinant” for deciding a qualified applicant.
That could be a gateway to charging employers for their hires, in essence a technological mimicking of the executive search model. And perhaps going executive search one better with a cost-per-quality-hire model.
As attractive as that sounds, hiring is fraught with unpredictable human factors beyond the ability of any AI program: a hiring manager’s decision based on the strength of a handshake or a decision to promote from within.
So, although Chevalier said, “Everything is moving towards deeper and deeper analytics to maybe even past the hire, to quality hire,” a cost per hire model of any kind is, today, purely aspirational.
CPC is the norm, CPQA is the next step
For now, the trend is a more modest, yet significant supplementing of cost-per-click (CPC) pay models with cost per applicant (CPA) and, more significantly, the emerging cost per qualified applicant (CPQA).
All the leading recruitment marketplaces in the U.S. and many niche sites, as well as all programmatic companies, have CPC pricing. Some career marketplaces, like Indeed, are CPC exclusive. Programmatic companies all have some form of CPA program and, like Appcast, Pandologic and Recruitology, they’re experimenting with CPQA.
CPA as a product is a fairly recent development, dependent on the vendor having hooks into the client’s ATS or using some other method of knowing that an application has been submitted.
Cost per applicant, however, has long been a part of the complex calculus programmatic agencies and sophisticated employers use to determine the effectiveness of ad placement.
Consider this simplified example: marketplace A logs 1,000 clicks yielding two applications. Marketplace B also logs 1,000 clicks, yet produces five applicants. If each site is paid 30 cents a click, the cost per applicant on the first site is $150. On the second site, the cost is $60.
On a purely CPC basis, the spend would shift from the first employment site to the second site. If that was the only factor, programmatic performance-based ad placement would be simple.
Cheap doesn’t mean effective
But as Steven Rothberg, founder and chief visionary officer of U.S.-based CollegeRecruiter.com, explained, it’s a mistake to base a buying decision entirely on the cost of each application.
“You can’t just buy from the cheapest sources. You also have to buy from the sources that can drive volume,” he said. A career site might have a 10% conversion rate, but based on only 100 clicks, where another job marketplace might have only 2% conversion, but deliver 1,000 clicks.
“Complicating that even further is if an employer says, ‘We also want the candidates to be diverse. We also want the candidates to come from universities, from the southeast, U.S.’ or whatever other kind of segmentation. Now, you have to take that into account when you’re deciding which job boards to cut off, which to increase their CPC and which to reduce their CPC. It gets really messy really fast.”
Think back to the example. What if the applications from marketplace A were both diversity candidates and none from the other marketplace were? A placement based purely on cost per click would shift the campaign spending to the cheaper, higher converting career site, reducing or even eliminating diversity candidates.
Rothberg sees CPA and CPC as intimately woven together, which makes it difficult for SMB employers to wrap their head around the importance and value of both.
“If you talk to an employer, and you say, ‘Would you prefer to pay per click or per application?’ Virtually every employer, correctly so, is going to say, ‘I’d rather pay per application.’
“Then you say to that employer, ‘Okay. But you’re only converting one out of 1,000 clicks into applications, and your competitor for the same talent is converting 100 of the 1,000 into applications, are you willing to pay 100 times what your competitor is for that same candidate traffic?’ And the employer would probably say that would be terrible. That would be so unfair.”
Yet conversion has more to do with the employer brand, the way a job description is written, the benefits and pay and other factors than it does with the site where the click originated, said Rothberg.
Those are matters beyond the control of either a recruitment marketplace or a programmatic agency. Still, some career sites do a better job of providing clicks that convert to applications than others, he said. And those sites get a higher percentage of the click price than others.
CPC and CPA models have both proven effective as pricing models. Indeed is a multi-billion-dollar company purely on offering a pay-per-click service. CPQA is the next step in the evolution of recruitment marketing.
As PandoLogic’s Baker told us, “It’s our job at the top of the funnel to drive as much data as we possibly can to influence better conversion, better decisions, better outcomes down funnel. So the further you go down the funnel, to engage candidates, to do the natural language processing on the job description, to build that data model to compare it to what’s in a resume, to compare it to the candidate data you’ve collected, the more likely you’re going to influence the outcome for the recruiter, the hiring manager.
“That’s adding value,” he said. “It’s not just passing a click and hope that that click converts. There’s no value added to that.”