Lately, a good number of people have asked me for feedback on their tech recruiting startup ideas, and I’ve noticed that I tend to ask the same questions and give the same advice over and over. Below, I’ve reproduced some of these things. At the end of the day, a lot of my advice is based on the current market climate, where engineering demand severely outstrips supply. I don’t know how long this state of affairs will last, but this state of affairs is precisely why this space is heating up and why it’s so lucrative.
At this point, you might be asking yourself something along the lines of, “Who is this dipshit, and why is she qualified to give advice about starting companies?” I ask myself that every day, and as a disclaimer, I too am working on a product in this space (interviewing.io), so we’ll see if I take my own advice!
1. Engineering hiring isn’t a filtering problem. It’s a sourcing problem.
Many of the people approaching me are eng hiring managers who are considering leaving their job to build tools that fix the hiring problems they encounter in their day-to-day. As a hiring manager, one of the biggest problems you encounter is filtering, namely going through a bunch of candidates and figuring out who’s good. This is really hard. Resumes are pretty low-signal, and technical interviews are fraught with their own set of challenges. But sometimes hiring managers tend to forget that before candidates even get to their part of the process, they’ve chosen to be in the mix, and they have already passed through at least one or two filters. In other words, you’re taking for granted the fact that the people who are crossing your desk want to work at your company in the first place.
Hiring, like sales, is a funnel. At the top, you have your attempts to get people in the door, either through building the kind of brand where people want to apply or through spamming the world or any number of other things. The middle is filtering, where you try to figure out whether the people in your funnel are worth hiring. Unfortunately, filters don’t make people better, so you are constrained by the quality of your sourcing efforts. And, in this market, where engineering supply is severely out of whack with demand, where good people are rarely actively looking for jobs, and where contingency recruiters get at least $25,000 per hire, the biggest problem isn’t filtering through a bunch of engaged job seekers. The problem is engaging them in the first place.
To drive this point home, let’s do a thought experiment. Given the data available on the internets, you could reach out to every engineer at Facebook and/or Google who went to school at MIT and/or Stanford. That data is already out there, and between LinkedIn, Rapportive, Clearbit, Entelo, other candidate search aggregators, and/or a little bit of quick and dirty scripting, doing this isn’t that hard. What’s hard is finding the ones who want to talk to you. Therefore, any product in this space that focuses on filtering isn’t solving the fundamental problem.
2. If you’re building a two-sided marketplace, create value for great engineers up front.
As you saw above, in this market, great engineers are rarely looking for work. Therefore, if you’re building some variation on the theme of a two-sided marketplace that attempts to pair top talent with great companies, you will quickly realize that getting great engineers to sign up is hard. They simply don’t need it and don’t want to be inundated with more noise.
If you do want to engage with the best people (And you do, right? Because it’s easy to go broad if you start elite but really hard to go the other way… kind of like a hiring law of entropy.), think carefully about what value you’re providing up front to people who don’t need you — if you can get that segment interested, everyone else will follow. Hired did this really well by taking advantage of the fact that salary data is pretty opaque and convoluted. In other words, all of a sudden, great engineers had a reason to sign up because, with very little effort (clicking once to sign in with LinkedIn), they could get a pretty good first-pass approximation of their market value that they could then, at the very least, use as leverage and at best, to find a new job. This is a big deal.
How do you apply this to your idea? If there’s some hard-to-find info you can shine a light on (salary data, team structure, etc.), by all means do it. If you can’t do that, then think about convenience. As it stands, interviewing is a pain point. Great people often have many interview loops going on in parallel, different parts of which are more frustrating than others. If you can streamline part of the process, or let the candidates you work with skip some of the drudgery, you’re going to be in good shape.
3. Are you trying to force people into behaviors that run counter to their incentives?
Regardless of the space you’re in, you should be building something your users want, but I can make the pitfalls really concrete in the tech recruiting space. Specifically, many of the people that come to me asking for my opinion on their business idea oversimplify the incentives at play. They assume that everyone just wants to hire the best people and that’s it. However, the world of incentives is complicated, and you have to be mindful of the misalignment of individual incentives/those of the purchasing decision makers relative to the optimal outcomes for organizations. In other words, though everyone wants to hire the best people, unless your product is an unambiguous, direct line to doing that, shit gets hard. To illustrate these points, I’ll go through 2 examples. Both are drawn from reality. The first one is from my own experience, while the second is my analysis of why someone else’s product in this space didn’t pan out.
Incidentally, although in-house recruiters are not the only ones with complicated incentives, they are generally the gatekeepers for figuring out which hiring products to use, so I’ll focus on them here. If you’re selling to recruiters, you have to be able to put yourself in their shoes, and you have to be careful to not oversimplify what their incentives actually are. “Recruiters want to hire more good people” is often a gross oversimplification that causes you to lose a lot of time building the wrong things because you don’t deeply understand your audience. Your other option is to say that you’re going to go around recruiters and sell to a different org within the company, but that’s its own can of worms with its own set of tradeoffs that I’ll save for a future post.
Example 1 – Why presenting great candidates isn’t enough
So, with in-house recruiters, their whole job is to hire good people, right? However, what does “good” really mean? When I was focusing on running a 3rd party recruiting agency, I consistently ran into the same problem when working with in-house recruiters. I’d present a candidate who looked like shit on paper but that I had reason to believe was good (previous work, having run them through technical screen, etc). Recruiters at these companies, by and large, wouldn’t entertain a conversation with these candidates because it was too risky for them. Engineering time is precious, and if you present someone who looks suspect and they don’t work out, you incur a good amount of vitriol internally. If, however, you keep presenting safe candidates, some of whom make it through and some of whom don’t, no one can blame you.
To put it really concretely, I’d expect that as an in-house recruiter, if you presented 10 candidates from Google/Facebook/MIT/Stanford, 8 of whom didn’t get an offer, and 1 of whom did but was clearly never too interested in your company, and 0 of whom got hired, no one would bat an eye. On the other hand, if you presented 10 candidates, all of whom looked kind of weird on paper, 2 of whom got offers, and 1 of whom got hired, you’d probably going to get a stern talking to.1
It’s a bit silly, but that’s how things work. If you’re building a product that surfaces people for recruiters to talk to, think about whether those people are risky bets and how those risky bets affect recruiters’ metrics — recruiters are bound by all sorts of metrics, and how many people they hire is only a small part. So, getting them to make seemingly unsafe, ambiguous choices with no clear return probably isn’t the best way to go.
Example 2 – Why it’s important to understand internal engineer/recruiter dynamics
Here’s another one. Let’s say you’re building a sourcing tool that involves getting the engineers at a given company to hook in their social networks so that your recruiters can then reach out to the people that engineers at the company already know, as those people are likely to be 1) good at stuff because yay referrals and 2) warm-ish leads. Cool idea, right? And it makes sense because recruiters want to hire more good people, right? Unfortunately, as before, it’s not that simple.
While recruiters do want to hire good people, that’s not the only thing they want. There are tons of products coming at them claiming to make their lives easier all the time. Most of these products end up not doing that. Now, in addition to asking them to spend time on something with questionable return, you’re asking them to evangelize said questionable product to engineers, whose participation is necessary for things to work, before the recruiters even have a chance to confirm that the product is, in fact, useful. Add to that the tension/stratification that already tends to exist between a company’s engineers and its in-house recruiters, and you wind up with a tall order. Having to vouch for something before you know it’s great is uncomfortable and weird.
For the record, this product was a real thing, and it was called YesGraph. When I first heard about it, I thought it was an awesome idea, but then I learned that they ended up pivoting away from hiring. I believe that the misalignment of incentives I described is partially to blame.
Incidentally, if you’re curious to read more about the challenges around driving/productizing referrals, I wrote about it at some length on Quora.
4. If your goal is to scale your two-sided marketplace, don’t turn into a middleman by accident.
In any two-sided marketplace, the path of least resistance often involves sitting between the two sides and playing god to make sure that everything that is supposed to happen is happening. You want to make sure that the right people are talking to the right companies and that companies are interviewing your candidates and so forth. Sure, down the line, you’re going to build an automated system, but for now, you’ve gotta get in there and make some tweaks.
I am a wholehearted believer in doing things that don’t scale. But, you have to be mindful of whether what you’re doing is going to keep you from going in a scalable direction in the future. Matching users with companies by hand? Great, that makes sense while you learn what works and what doesn’t. But, if you aren’t mindful of it, before you know it, you have a recruiting agency on your hands, just one that happens to have a nice startup-y skin. And, by the time you realize it, you’re too vested in the unscalable status quo to change anything because money’s coming in. And then you have to hire a bunch of recruiters, and then you’re committed. And now, you’re growing linearly with the number of recruiters you hire.
So, sure, do things that don’t scale, but keep asking yourself whether what you’re doing will ever be possible to automate and be mindful of whether you’re creating an infrastructure of middlemen with a vested interest in the status quo. And if you realize that what you’re doing is really hard to automate with technology or solve through crowdsourcing/engaging with your users, then ask yourself whether there’s something else you should be trying.
1 As a footnote to this story, one of the first companies I worked with made a deal with me after seeing the nontraditional kinds of candidates I presented. They said that, though they’d be down to talk to the first 3 candidates I threw their way, if most of them didn’t make it to at least onsite, I was going to be fired. Two years later, this company is one of my favorite clients to do business with — I love their team and, even at the time, I completely empathized with their skepticism. Though it ended up working out well, this kind of deal is rare and getting over the hump was likely helped by the fact that I knew one of the founders.
Question about inhouse systems. A friend (business side, not engineering) has been going through the following process.
1) Sees position on job Web site. She is an almost perfect fit for this position.
2) Enters cv and other data at company’s in house recruiting Web page (this takes over an hour).
3) Receives email rejection (saying she’s not a good fit) in under 5 minutes.
4) The next week, she receives an automated email from the company, advising her of a position that is a possible fit. This position, however, is not even close to being a fit (e.g., requires a Ph.D in a field unrelated to hers).
This scenario seems common.
I was thinking of entering into this space, but I can’t figure out the motivations of people at the company who would find such a system to be satisfactory. Do they not test their own system? Do they not care? Any thoughts? Thanks.
Jim, in-house systems (ATS, or applicant tracking systems) are a common road-block for job seekers and often referred to as a “black hole”. Unfortunately, they are software purchases that have already been made and are heavily embedded systems that are often difficult to change (typical of HR software).
Any solution that would help surface the best candidates out of these ATS systems, for the internal recruiters that are using them, would be valuable. You wouldn’t want to create a new system, but one that ties into their existing system, perhaps applying some kind of data science to identify the low-hanging fruit the internal team could have missed.
Often times the auto email responses and ongoing communication one receives from these systems isn’t relevant, but in a scenario where there are potentially thousands of applicants entering the database each week, the internal team has to create some kind of process/structure to manage all of this activity.
The unfortunate flip side to this is that the best candidates that flow into the system this way are often overlooked.
My advice: find a company that is willing to let you help them solve this problem by giving you access to the back-end of their ATS, from which you could extract the data you need to add value here. Ideally you’d create a solution that leverages the job postings in the system, and applies a matching algorithm to surface the best candidates that are already in the database, but haven’t been identified yet.
Best of luck!
That sounds like a great idea. A back end add-on would be a much easier sell than replacing the current system. Lower political hurdles. It would be easier on the tech side as well, because the add-on could be tested and tweaked in parallel to the current system, without the need to go online until proven.
Happy to chat more about this. Let me know the best way to reach you.
Neat observation, made me think about http://corcodilos.com/blog/8064/how-hr-optimizes-rejection-of-millions-of-job-applicants
It’s a great article and I can totally relate to it, because I am myself running a tech hiring platform. Though I agree with most of your points, I would like to disagree with one – “If you’re building a two-sided marketplace, create value for great engineers up front”
Though this concept can work in other marketplaces I think it will fail in tech hiring marketplace. The reason simply being, the top of the funnel in tech industry can be really intimidating to the rest of the funnel. And even though other aspire to be there, they might not be able to relate to it because they feel they are simply not that good.
The other problem with this is, if one attracts the best engineers you can immediately start working with the likes of Facebook, Google, Uber and Square because these companies want nothing but the best and match the payouts, but if one is really solving a tech recruiting problem there are tons of other companies out there who might not be able to match exorbitant salaries but still want to hire.
If one starts with elitist, how does one attract the mediocre players on both the sides. At some level I feel that’s why Topcoder could never break out, because it was too elitist.
Would love to have your thoughts on this. Also I have a question, given the unwillingness of developers to be active job seekers do you feel there can be a two-sided market place where the core value proposition is only recruiting?
I’ll email you!
I thought this article was incredibly well-written. Anytime you are reading, and are thinking “man, that is exactly what I was thinking” or “whoa, I think I was about to go down that rabbit hole”, it is of tremendous value.
Kudos Aline, Kudos. For sharing this, for writing so well, and for helping those of us about to take this roller coaster ride with some words of wisdom.
As an entrepreneurial engineer that gets contacted by recruiters on a daily basis, who is about to start a business “kind of” in this space who is really worried about not becoming a recruiting agency, this resonated greatly, and was very true on many levels.
Once we get our product off of the ground, I hope to write a “part deux” to this article, with our experience in this space.
Until then…thank you for such a well written piece!
I’d like to hear your follow-up to this article.