Google Doubled Employee Bonuses for Referrals - and It Didnt Work

Google paid employees $2,000 for a successful referral. That wasn't working, so it doubled the amount to $4,000. Now, even for people making a lot of money by normal standards, $4,000 is not chump change. But that didn't work either.

Turns out, as the following article explains, that money really had nothing to do with it. Employees referred friends and acquaintances because they believed Google was a cool place to work and they wanted more great people like themselves around.

That was both the good and the bad news. Google asked more specific questions and followed up immediately, eventually increasing referred hires by a third. But they also got more people like the ones who already worked there. Which, if you are trying to sell stuff to people all over the world who are not exactly like those who already work for you, is a problem. A problem they are figuring out how to solve, probably by not using referrals. JL

Max Nisen reports in Quartz:

The company gather(s) 20 to 30 employees and asks them go through their social media contacts, with recruiters right there to follow up on potential candidates. Breaking down a huge question (“Do you know anyone we should hire?) into lots of small, manageable ones (“Do you know anyone who would be a good salesperson in New York?”) boosted referrals by more than a third.

Like just about every employer, Google is keen on referrals. They come with a built-in reference and, on occasion, some inside information, and provide a sense of familiarity to both the candidate and the interviewer. That tends to lead to a much shorter hiring time, a higher rate of acceptance, and a tendency for the referred hire to stick around (pdf) the company longer.

Over time, Google got better at referrals, as its hiring chief Laszlo Bock writes in his recent book. Financial incentives had nothing to do with it—when the company tried doubling its referral bonus from $2,000 to $4,000 first, it didn’t help at all.

People made referrals because they liked working at the company, not because they were hunting bonuses, Bock writes. So next the company simply tried making the process more pleasant. Instead of feeling as though their referrals were sent off into a black hole, employees who referred candidates got weekly updates on the status of the people whose named they had passed along. Referred candidates got a call within 48 hours—and, as the hiring process progressed, they would face fewer interviews than is standard for candidates without an employee referral.

Everyone ended up happier, Bock writes. But even then the numbers didn’t jump.

What worked was “aided recall,” a concept Bock’s team stole from marketers. People rarely remember an ad on their own, but with a tiny prompt—a brand name, an image, any kind of specific information—they are vastly more likely to recall it. From the book:

In the context of generating referrals, people tend to have a few people who are top of mind. But they rarely do an exhaustive review of all the people they know….. nor do they have perfect knowledge of all the open jobs available. We increased the volume of referrals by more than one-third by jogging people’s memories just as marketers do. For example, we asked Googlers whom they would recommend for specific roles: “Who is the best finance person you ever worked with?” “who is the best developer in the Ruby programming language?”

The company would also gather 20 to 30 employees and ask them go through all of their social media contacts, with recruiters right there to follow up on potential candidates.

“Breaking down a huge question (“Do you know anyone we should hire?) into lots of small, manageable ones (“Do you know anyone who would be a good salesperson in New York?”) garners us more, higher-quality referrals,” Bock writes.

This technique boosted referrals by more than a third.

But even this wasn’t nearly enough. According to Bock, even if Google were to hire referrals at 10 times the rate it hires applicants from other channels, the company would need 300,000 referrals a year to hire as many people as it hoped to. It never got more than a third of that figure. So it shifted its focus to building an in-house recruiting organization. In other words, referrals are still important but they're not everything. And for a company trying to diversify its workforce they shouldn't be - one consequence of referrals Bock doesnt discuss in his book is that people tend to refer friends or acquaintances who are similar to them, which can negatively impact diversity. Google experienced the effect first hand and now is working hard to reverse it.