Google's new ad model - and what it means for recruiting

Check out the New York Times article today on Google. Apparently, they're experimenting on a small scale with a new ad model - cost per action. In other words, advertisers are only paying for results. Don't make a sale? You don't pay. Simple as that.

Simple is an overstatement here, of course. The article goes on to say that publishers like Google have shied away from this model because it would add complexity to their business. Yet for the customer this is so much of a win that it would probably be worth significantly more money. This is the holy grail for marketers (da Vinci Code vision of this aside) - and I can see them being willing to pony up to participate in the same way that they offer referral fees to consultants and resellers, usually in the 10-20% range of sales. If this works, I can see a day where advertisers will have their choice of who advertises on a given site and will make their decision based on the size of the sale and probability of success rather than the reverse of this model that exists today, where advertisers who pay more get the impressions regardless of their probability of success.

By comparison, recruitment advertising is still in the Stone Age. We're looking under rocks on Monster.com, pounding away at leather hides with crude tools on Hotjobs, and praying to the sun-god on CareerBuilder. Vendors like Indeed have started to experiment with cost-per-click, but it's going to be hard to build a sizeable business with an audience that professes to want less applicants rather than more by charging them every time someone applies.

So what's the evolutionary leap here? Cost per hire. Yes, this adds complexity to any vendor's business model. For example:

• A hire is not a hire is not a hire - the cost for a customer service hire should not reasonably be the same as the cost for a security cleared engineering hire. 
• Tracking this is not as easy as tracking an immediate payment since it could take months before someone is hired.
• Some people may not be hired for the job they applied for, making tracking even more difficult.

Despite the above, I don't think a cost-per-hire model for a job board is outside of the realm of possibility at all, actually and companies would actually be willing to pay more for the hire than they would for a $400 posting - just look at the $38 billion search industry, which charges 20% of first year salary. Granted, companies are more price sensitive with recruiting costs than sales costs - the goal is not to spend 10-20% of salary - but many do, as evidenced by the sheer scale of the staffing firm industry.

Today there are technology companies like H3, BountyJobs, JobThread, Blue Chip and others that are charging based on hires and leverage more than just traffic by engaging recruiters and non-recruiters in the search process, potentially increasing the odds of success. None of them seem to really be pushing for a consumer presence and making the investment necessary to get there, however. I personally think this is a big opportunity if they can nail and scale their business models and make the successes customers they have had more repeatable (although one is usually a function of the other).   

There's a big market out there for the taking for a publisher. Just like the quest for the grail, the reward is great but the road is treacherous.

Is there anyone bold enough to embark on this adventure? 

Comments

[Steven Rothberg, CollegeRecruiter.com]

We've been successfully providing our clients with a pay-per-lead option under which they pay for each qualified and valid lead a/k/a resume as those are relatively easy to track both for the employer and our systems. But we've been unwilling to do pay-per-hire deals because those deals require our clients to be both honest (the vast majority are) and capable of providing accurate information (the vast majority fail this I.T. test). The fundamental problem is that their applicant tracking systems are not set up to automatically track a candidate from the point of application to the point of hire in such a way that the employer knows with absolute certainty from where that candidate was truly sourced. Most employers that we talk with will claim that they know but almost none actually know because they rely on the candidates to tell them. Study after study prove that candidate self-identification leads to improperly attributed sources in over 80 percent of the cases.

Until the ATS vendors are able to work through these issues with their clients, I can't see success for any pay-per-hire model.

[Dave Lefkow]

Based on experience (I've done several ROI projects for companies) I have a hard time believing that 80% number. But you're right - it's not completely accurate, which makes your job more difficult. There are ATSes that have "true source codes," but they're relatively rare.

The return might still be worth the risk - if you got an additional amount of revenue from a customer even within 20% inaccuracy, you could more than make up the difference. I believe that customers will pay a significant premium for a cost per hire model.