This is the second in a series of articles on Hiring Engineers by Jocelyn Goldfein, Angel, Advisor. Formerly: Engineer @ Facebook, VMware, Startups, Trilogy.
By Jocelyn Goldfein, Angel and Advisor
Software startups need two kinds of resources: capital and talent, and right now capital is the easy one.
Fundraising is difficult and demanding, but venture financing is more available now than it’s been in decades. That’s created huge demand for engineering talent from startups, who are themselves only bystanders in a larger Valley talent war between titans like Apple, Google, and Facebook. Competition for talent has spawned tactics ranging from the absurd to the illegal. It’s a little hard to explain to a country with a 6% unemployment rate and nearly 10 million job seekers, but in tech, the people who need help and advice are not the software engineers, but the companies desperate to hire them.
Help with hiring is the common theme across all the tech companies I advise (or have ever worked for.) So I’ll be writing a series on hiring engineers, including:
- Step 0: What to Look For
- Step 1: Sourcing (this post)
- Step 2: Filtering
- Step 3: Screening
- Step 4: Assessing & Deciding
- Step 5: Closing
- Possible sidebars: hiring bar, diversity, metrics & tools, recruiter/hiring manager teamwork
Step 1 is the place people reliably understand they need help, so I’ll start with sourcing and come back to step zero in the next post. This series will be framed around the hiring of software engineers, but this advice is probably generally applicable for any type of small company technical hiring (PM, Design) where there are more jobs than qualified candidates.
What is sourcing?
Sourcing a job candidate is really two distinct verbs: you have to find people you want to talk to, and you have to convince them to talk. In other words, we need to find people who are both talented and available (for some definition of availability). That represents your effective candidate pool.
The size and ratio of the circles doesn’t matter — they’ll vary with: the type of person you want to hire, your employment brand, your region, your product, etc, etc. With few exceptions, though, no matter what your company’s bubbles look like, the overlap is going to be a tiny percentage of each bubble.
The problem is that membership in these bubbles is often inversely correlated. The stronger an engineer is, and the more you want them, the more likely everybody else wants them too (starting with their current employer.) The personal cost of job switching is high. So most great engineers aren’t looking most of the time, and an email from you probably isn’t going to kick them into job hunting mode if they’re happy staying put.
In the late 2000’s, Google was such an aspirational employer that even engineers who were perfectly happy in their current jobs would drop what they were doing to interview if a Google recruiter came knocking on their door. Chances are, that isn’t you — even Google themselves don’t have it that easy anymore. If you have that cachet, make the most of it! For the rest of us, it’s much harder work.
Because of the small overlap, focusing on either bubble alone can be a terribly wasteful error if you’re small. The visible “market makers” (job boards and career fairs) are overflowing with job hunters who probably don’t possess the minimum qualifications — most of the people who have the qualifications are currently employed, and won’t publicly acknowledge that they might be open to a change. The classic technical recruiting techniques that evolved to deal with this involve a brute force search of the “qualified” bubble.
If you’re small, the key is to stay laser focused on the overlap. Let’s take a look at some tactics.
The referral (a mutual introduction of candidate and employer) is the single most effective sourcing technique. The match-maker vouching for both parties means there’s a strong chance this person is qualified to work for you, and a strong chance they’re willing to talk to you, a huge win/win. Even the weakest referrals (“His daughter goes to my kid’s school and I know he has developed a mobile app”) are more likely to land in the overlap zone than a random stranger on the internet. People are always going to be more willing to talk to you when there’s an introduction.
How do you get referrals? Start with your own network. Make a list of the great people you’ve worked with at past companies, went to school with, summer interned with. Think of people you know through industry associations, conference speakers, tech talks, and open source (it only works if they know you, too!) Facebook, LinkedIn, et al make it easy to scour your network of contacts to jog your memory. Even if you aren’t certain it’s someone you’d hire — that’s OK — that’s a decision to make after the interview. Right now you just want to build a pipeline of qualified people willing to interview with you.
Once you have the list of contacts — get in touch. If you know them well enough, you can cut to the chase. One low-pressure tactic that reduces the social awkwardness is to to describe your job and ask them if they have any referrals. Best-case scenario, they nominate themselves. Next best, you get more referrals.
Now get all your employees to do this same exercise. You don’t have to limit yourself to the engineers. If you’re a small tech company, everyone in every function works with engineers and has a point of view about the ones they’d like to work with again.
Now ask your investors and advisors to do this exercise. Keep going.
Research shows that “weak ties” are the most likely to help a job seeker land a job — not their closest friends. Same effect works in reverse for you — you’ll develop more new candidates from more peripheral contacts who have a network that doesn’t overlap with yours.
Is this nepotism? No. You’re only sourcing, not making job offers. All the candidates you find this way are still going to have pass your interview process. You’re just sourcing in a way that maximizes your chances of finding a mutual fit.
Suppose you already know someone amazing — say because you saw them speak at a conference, or you admire their open source contribution. You have no need of a referral to vouch for them. It’s still worth hunting up a mutual acquaintance to introduce you — because someone vouching for you makes the candidate a lot more willing to entertain a conversation.
Word of warning: The main risk of sourcing only via referrals is that if your network is too narrow, your employees may end up too homogenous. The top risk is to your culture — if you hire an entire team of people from Microsoft, you will inherit Microsoft culture, instead of inventing your own. Startups benefit from hybrid vigor. You also run the risk of not securing enough breadth of expertise (for example, if you hire all infrastructure engineers, they can probably work full stack and build your front-end, but you will be better served by a mix of engineers, some of whom passionately love building beautiful interfaces.)
My advice is just to be mindful of this when collecting referrals and making hiring decisions. If there’s a type of diversity you’re after, referrals are often the best tool for finding it — just guide your referrers with specifics of what you’re looking for.
If you only have 3–4 hires to make, you should focus all your efforts on referrals; you can pretty much ignore the other sourcing techniques in this post. As your hiring needs increase, you can supplement referrals with other techniques.
No matter how large your hiring target grows, referrals can be a mainstay of your hiring. Every employee you add to your company adds a new network of possible referrals. There are many great techniques to enhance your referral program at any size, including software to automatically scour the social networks of your employees; sending recruiters to interview new employees for referrals; making referring and hiring an admired part of your culture; and high touch communication with referrers.
Campus is a terrific place to find new engineers. Graduation is the one time in their career that engineers are certain to be open to job offers. Well — almost. Increasingly, the battleground is moving to the summer internship; big companies do their best to lock up their interns with full time offers before they start their senior year, so you may have to move upstream, too.
Be aware that the season for campus recruiting starts in August for students graduating the following spring. Making offers 10 months in advance isn’t an option for every company, but if you can, and if you have the knack for spotting and developing potential, you can find and hire some of the best engineers on the planet before they get taken off the market.
Campus recruiting at scale is a complex operation — big tech companies may hire from hundreds of schools around the globe, with dedicated recruiters for “high yield” campuses, and high touch programs including alumni events, career fairs, tech talks, on-campus interviewing, internship programs, sponsored club activities, even grad school fellowships.
Fortunately, you don’t have to do all of that to recruit a few new grads. You can be completely opportunistic and just post a few job listings. I recommend picking 1–2 campuses whose programs and faculty you can get familiar with and go a little deeper vs. making a shallower investment in more schools. And it always helps to have an alumni connection to bootstrap yourself with some referrals and knowledge of the school.
Stanford, CMU, and MIT are the three top-ranked CS schools and the favorite targets of the most aggressive tech employers. If you are prepared to go head-to-head with them, go ahead and target the most prestigious campuses. But realize that the reason those campuses are “high yield” isn’t because their curriculum magically produces the best software engineers. It’s that their admissions department is doing a lot of the heavy lifting of selecting high potential teenagers, and then the program itself becomes a self-fulfilling prophecy of creating more opportunities (internships and part-time jobs, access to mentors and role models, encouragement to build their own apps or contribute to OSS.)
You can find equally great raw potential in students at other schools (with a lot less competition from other employers), you just have to invest more time and effort in identifying them. Getting connected to professors who’ll send you their best students can be a massive competitive advantage. Hire a head TA one year and you’ll get a pipeline of top students who follow her for the next 2–3 years. Programming competitions like Topcoder are another way to surface highly productive coders at non-pedigree schools.
It’s tempting to stick with the schools in your backyard, but that’s what every other small tech company in the Bay Area (or Boston, or Seattle, or Austin) is thinking. Go farther afield, to regions less oversubscribed with tech employers. Look at smaller schools which don’t have the economies of scale for big companies to invest in visiting. Consider whether you can support the costs of immigration visas; comparatively few US startups are willing to deal with the hassle, so for those international students interested in working for a US startup, you’ll be a top destination.
Don’t be afraid of an internship program. Once you’re committed to making job offers at the outset of senior year, that’s pretty much the same lead time as making offers to rising seniors (or Masters’ students) at the end of their internships. Interns’ contributions generally pay for themselves, and after 12 weeks of working together, you know for sure you’re getting someone great.
When Facebook analyzed all of its recruiting sources to see if there were differences in ultimate on-the-job performance, returning interns were way out in front.
Referrals and campus are the two best sources for staying focused on the overlap. From here on out, you’re going to have to spend more effort on filtering; to be honest, until you need scale, I’d stick with referrals and campus and call it a day. Oh, and one more thing: publishing a job listing on your own company site.
You have to do that no matter what, and you have to make it good — because candidates you source via other methods are going to visit your website to check you out. This is your chance to perfect your pitch about why talented people should be interested in you — it should highlight your products, mission, and culture. It should also be slick and beautiful (especially if you build consumer products or want any hope of hiring a strong designer) and likewise, it had better be mobile friendly, especially if you are trying to hire mobile engineers.
It’s common wisdom that website applicants are low quality — that’s not necessarily true. Especially if you don’t have a well-known brand, visitors who come across your job listing are probably users of your software, and they may be particularly knowledgable or passionate about what you do. It’s not high effort to scan resumes for the minimum qualifications (and if volume is a problem, there’s software designed to help) — the fact that these applicants are guaranteed to be seriously interested in you tilts the ROI in their direction.
Consider 100 solid resumes with unknown interest in you, vs. 100 job applicants with unknown qualifications but definite interest in you. It’s a lot faster and easier to scan the applicants for qualifications than it is to try to make contact with the 100 resume-owners and convince them to talk to you.
The same math doesn’t really work for job board postings, because there you’ll see a lot higher volume from people who just want a job (any job) without any particular passion for you. I know entrepreneurs who’ve successfully hired via CraigsList, but I can’t recommend it based on personal experience.
What about running highly targeted ads for your job posting? My experiments in this direction are inconclusive, but I don’t think you’re going to get a lot of engineers to convert directly from seeing an ad to applying for a job. It might work in large enough volume when your ad is seen by candidates who are actively job hunting, but even then, expect clicks to browse your website, not direct applications. Overall, though, the cost of messing around with it is pretty low — if you have good ideas about ad targeting, give it a shot.
Produce High Quality Content and Engage With Your Audience
Great people are attracted to working with other great people. One of the best ways to get great people interested in talking to you is to publish your work in places where strong engineers hang out. Send speakers to high quality conferences, blog, and contribute open source software; publish papers if you are looking to hire PhDs. Then respond to the people who respond to you, and don’t be shy about dropping “by the way, we’re hiring” into the conversation.
This is an investment of time that you could be spending building and shipping product. You may not recoup it on recruiting benefits alone if you are only hiring a few people. But it has other benefits — for example, these activities are pretty intrinsically rewarding for the employees who do them, so it’s good for job satisfaction. Depending on your business, they may help you generate customer or partnership leads, not just recruiting leads. If OSS is a major part of your technology stack, being a contributing member of the community is essential and will repay itself tenfold.
This is where I think it gets interesting to spend ad dollars — if you have great content, use ads to build a larger audience for it. The people who engage in a serious way will emerge as the ones you want to build a relationship with.
If you can’t produce great content of your own, the opportunistic version of this is to hang out in the comments section of great content that relates to what you do, engage with other commenters who seem smart, and try to turn a dialog into a job conversation. (The in-person variation: attend a conference and hand out your business card; or host a booth if you’re big enough.) Pick your location carefully — depending on what you do, github is probably better than HackerNews which is probably better than VMworld.
The bottom line on this whole section: if this sounds like something you’d do for fun anyway, you might as well be recruiting while you’re in there.
Emailing strangers. Paying other people to email strangers for you.
“Passive sourcing,” as it’s known in the industry, forms a huge part of the recruiting machinery of large companies, because once you’ve run out of qualified people who visibly want to talk to you, the next step is to try meeting ALL the qualified people and convince some of them to talk to you.
The premise of passive sourcing is that recruiters can craft missives that convince those happily employed, non-job-hunting-engineers (hence, “passive”) to consider working for you. The reality is that these emails don’t convince anyone to reply. This is a brute force search algorithm to locate the engineers who are job hunting below the radar (they identify themselves by replying.)
Passive sourcing can come close to fulfilling its promise when you are in it for the long haul. Let’s say silicon valley engineers change jobs every 2.5 years, and take an average of 2 months to job hunt. Those are aggressive estimates, but it still only nets out to <10% of the engineering population open to a job change at a given point in time (and reality is probably much lower). Remember, the good ones are being selective, just like you are: hunting via their networks, not running around applying for jobs or listing themselves on job boards. An email that happens to land on one of those engineers in their job hunting window might be your one shot at catching someone you otherwise won’t hear from. If you wait 6–12 months, another 30% of the population you emailed is going to consider changing jobs — and some of them might remember your email and get back in touch. This is what recruiters mean when they tell you “it takes 6 months to develop a pipeline.” They are playing a volume game which needs both volume of email and volume of months to pay off.
The engineers you want can be pretty selective about their opportunities and it’s not a given that they’ll respond to those emails. After job hunting status (are they even looking?), the biggest factor in whether an engineer replies is your reputation. If your employment brand speaks for you, not much else will matter. Email quality has sadly little weight; email coming from an engineer/founder/CEO vs. a recruiter will increase the reply rate but not the actual “willing to interview rate.” If you’re too small to have a rep, then how she feels about startups in general and your domain may play a role, as will email quality and author. If the email is good enough to get her to click, then the quality of your site (and your social media presence, and your google search results) will play a factor.
Bottom line — it’s all a small percent of a small percent of a small percent. It’s low yield. This is a strategy for big companies who have high volume hiring targets that require digging in to the long tail, and the resources to do it (the key resources: an aspirational employment brand and a large recruiting team). Some VCs have the economies of scale to put together a recruiting team to generate leads for you this way — by all means take advantage if you have access to a program like that. But this is the last place I’d advise a small company to spend your own resources. If you find yourself using it as a last resort, I highly encourage you to read Meebo founder Elaine Wherry’s post on passive sourcing for startups.
Important note: I’m not running down recruiters here. Passive sourcing is a challenging field with dedicated and highly skilled professionals. You already know that when you hire engineers or execs from big companies, they have to adapt what they do to a small company environment. It’s exactly the same with recruiters. If you hire one from a large tech employer (and almost every tech recruiter cuts her teeth in those environments), then make sure she shares your vision of how you want to source, and that she’s putting her time into the sources that work for you, not just the sources she’s been trained to look at.
Moneyball for engineers?
As an engineer, my design sensibilities are offended by how messy and inefficient this whole system is. We have a market where the buyers and sellers can’t find each other because each party is interested in a comparatively small percentage of the total market population and has no good way to zero in on the most appropriate potential fits. Not to mention the transparency problem — job seekers need to keep their current employers from finding out that they are looking; in some cases companies have reasons to conceal certain openings until they are filled. As a technology optimist and entrepreneur, I think the whole thing is ripe for disruption.
One concept that fascinates me is using technology to predict whether a given human being is a high likelihood match for your job opportunity (and vice versa) in the same way that machine learning can be used to rank search results or suggest people you may want to connect with on social networks. What’s needed is a faithful digital representation of the human beings who represent the talent pool, and a bunch of experimentation and innovation in figuring out what “features” make them a likely match for a given job. We might not be able to reduce engineering talent to a set of metadata — but we thought the same thing about baseball players until sabermetrics came along 15 years ago. We might just need more innovation in our ability to describe and infer talent to human beings — exploiting metadata like graph structure and work output, not just resume keywords.
Another possibility is that what we need here is not technical innovation but market innovation; AirBnB and Uber are fascinating examples of how a new market-maker can shake up an established industry.
I’ve seen a number of startups attacking this in different directions (like Gild, Talent Bin, Milibo, or Hacker Rank.) You can bet the pure players at LinkedIn, Monster, and Dice and players in the ATS space are studying this deeply, too. I’m not ready to dub anyone a clear front-runner, but I’m watching the space closely. If you’re innovating in this space, or have had success adopting one of these tools — I’m interested in hearing about it!
The bottom line
Recruiting is worth a significant share of your time and effort, because building a better team is more than a competitive advantage; the quality of your team is life and death for a small company. Ultimately, my advice on sourcing boils down to this:
- Pick the sources that work for you and your team
- Don’t blindly apply big company techniques
- There’s room for innovation if you’re bold (or desperate)!
This post originally appeared on Medium.