Early Stage Recruiting

08/04/2016 08:35 am ET Updated Aug 05, 2017

Startup recruiting is notoriously hard. After doing a few full-throttle startups in my career, I finally understood what makes it so hard, and why many (especially, first time) founders stumble on this particular step in a big way. Founders find themselves in something akin to a "steel vise" that creates big pressure to do the wrong thing -- from both ends. Pick strategy A, or strategy B, and it still will be the wrong choice. Let's review why that toxic "vise" forms and how to escape it.

The Art of Start
The main reason recruiting became so hard is because everything else became so easy. Naval Ravikant, founder of AngelList, summed it up best a number of years ago:

Raising the first $25K for product development is easy -- join an incubator. Raising the next $100K is easy -- investors are following the incubators with automatic notes. Building a product and launching a product are easy -- develop on Open Source Stacks, host on Amazon, launch on Facebook, Android or iOS

The typical path of a startup nowadays is straightforward:

  1. Ideate new concept
  2. Hack prototype together
  3. Raise seed and/or join incubator
  4. Hire team

Steps 1-3 are all but trivial for technical founders. Founders progress happily along this path ("hey, why do they keep saying startups are hard -- this is awesome!"), quite enjoying their startup adventure, and then, suddenly (for the founder), step 4 hits them like a brick wall. In contrast to everything else they've experienced so far, it is incredibly hard.

Consistency bias plays a bad trick at this point: if everything the founders had experienced so far was relatively easy -- it took a few weeks, maybe a month at most, to achieve significant and meaningful progress with product, prototype, and fundraising. They are mentally unprepared for what seems like an incremental next step ("let's add few top notch people to the team!") is actually hundreds of times harder. The "Vise" is starting its ponderous crush with the founders right in the middle of it.

The Vise
After the first few post-funding honeymoon weeks, the founding team realizes that the clock is ticking. The first month is up with precious little real news to share with investors. The company is still a small outfit with all of the same problems they had last month. They need to hire everybody, in every vertical, right this moment. The anxiety of "non-hiring" and glacial progress goes through the roof and creates a huge pressure on the founders.

What is largely invisible to founders, is that from the outside their startup is not an Uber yet. It looks like all the others seed-stage startups that incubators nowadays churn out by the thousands. The old barriers are gone, creating a new seed startup is trivial, and the generic, undifferentiated startup is just a new labor pool. Job ads attracts all sorts of talent: far more from the middle than from the top. "First responders" who are in a constant job hunt all their lives will all but overwhelm the submission queue. Top talent may eventually show up, yet counter-intuitively it won't be in the first batches to apply, and the signal will be drowned out by an incredible amount of noise. Common rookie mistake, at this point, is to not do a real search at all, and just hire friends or roommates.

Of course, our founders are smarter than that, so they roll up their sleeves, invest a lot of time in the hiring process, interview a lot, quickly weed out the weak and then they are left with an undecidable pile of "strong maybes". There is objectively not much more to do here -- by interviews alone the external attributes of strong maybe and top tier are exactly the same. To make things worse, "we interviewed a lot" is subjective: it might mean 30-50 interviews, while rockstars might not even show up once in the first few hundred candidates.

Nonetheless, under huge pressure to increase the company's velocity, the founders start either hiring aggressively or at least doing tryouts. "Let's see how this person works out in practice, and then we'll decide if it is worth keeping them full-time."

Confirmation Bias
It is exactly here, on that final step that is supposed to be the sharp selector, where the system catastrophically fails.

It is pure psychology: the founders really want their new talent to work out. They are not looking to fire candidates they spent so much energy to hire; they are only looking to find more reasons to confirm the correctness of their own choices. Founders want to feel that they are the leaders of a big, sizable workforce and see those commits/designs pile up in the company's repositories. Founders want to feel that they are doing well in "scaling the company". They want another bullet item how they are "crushing it" in the next month's investor update. And they want to get out of that recruiting hamster wheel and get back to building the dream product: that's what they started the company for, not to do 40 interviews a week. That creates an equally strong pressure on the founders to just "lock in" early, and uneducated, hiring decisions.

All this taken together creates a self-reinforcing daily loop on the founders to overlook/excuse the exact type of issues that are actually critical key differentiators between "above average" and "top tier" talent. These issues might, in fact, look subtle, perhaps just a 10 percent difference! Those of us, with battle scars, know full well how much difference that last 10 percent will make over the lifetime of a startup -- it is all compound interest in a non-linear system. It is massive. Let that 10-20 percent suboptimal result compound for 3-5 years, and it will be all the difference between next the Uber and some irrelevant roadkill.

Inexperienced founders find it all too easy to lull themselves in acceptance: "this person will learn this over time", "I will mentor them", "it will be a growth opportunity for them". They come up with countless other excuses to avoid going back to square one and, especially, the social unpleasantness of firing (i.e., not to mince words -- being a total asshole to) somebody you just met. Founders decide to stack rank candidates on a 1-10 scale during tryouts... and then do their utmost to talk themselves into believing that a 6 is actually almost as good as a 10! It is not an accident that Ben's book has a whole chapter on "how to fire friends". The early stage is worse: even assuming an impossibly high early stage startup candidate pipeline success ratio of, say, 20 percent (we've seen as low as 5 percent), which means firing 8 out of every 10 just to get the first 2 solid hires. In other words, get ready from the start to fire everybody.

The Ruthless Efficiency

Once you understand the full weight of Steve Jobs's quote that "the first 10 people in the company fully determine the outcome", you will realize that the early stage recruiting funnel must have ruthless efficiency:

  • Inject: Any candidate is only tested by practical co-working lasting a few weeks or even a month. Interviews are just a low band filter.
  • Eject: Candidates who show any strategic shortcomings should be terminated from these tryouts or early positions immediately and without second thought. Position, budget and mental awareness of an empty slot are freed up as soon as possible. Due to the overwhelming presence of "first responders" in the initial flow, not doing that runs the risk of actually missing out on really terrific candidates.

Recruiting is the first brutally hard challenge for first time founders, and the first filter that will separate thousands of "labor pool" startups from the real stars of the next decade. It is a ruthless process that will not win CEO any friends or popularity contests. Efficient recruiting is, perhaps, the most important lesson that you can learn as first time founders. In the long term toxicity of suboptimal early hiring has doomed more startups than I can count.

One piece of advise about this process: 'ruthless and efficient', does not mean, and should not mean 'inhuman'. Another classic founder mistake is to mistake inhuman workplace practices (young Jobs) with cold efficiency (post-Pixar Jobs). The people that the CEO is going to fire are not bad people, nor they are bad professionals. They are simply the wrong people for the current strategic needs of the company -- that is all there is to it.

One way to humanize that objectively unpleasant process is to be transparent about it. Do not be underhanded in any way. Share openly with all candidates that you are terminating close to 90 percent of tryouts, and that you are a high-speed, high-demand, high issue-ownership startup. Guess what: top tier talent is actually pretty darn confident in their skills. They like the challenge. They like working with other top notch people who cannot stand any mediocrity.

When giving people the axe, be short, polite and to the point. "Hey, Jim, we reviewed your tryout commits from last week. It took you much longer than expected. Also there were no comments in either the code or commits and we had to spend extra time to understand what you did. I'm thankful that you gave us the opportunity to try working with you, yet we have decided not to continue this contract. We're happy to help you find your next gig; let me know if you want me to float your resume on our incubator's maillist" (oh, did i mention that your initial queue will be overwhelmed by 'first responders'? That is why.)

The good news is, once you know the hiring process, and the hard future costs of failing at that process, doing the right thing becomes much easier. Our exec team had done a few startups in the past, so we just played it this time by the numbers. Countless interviews led to dozens of tryouts across engineering and design. Interview, contract, terminate. Interview, contract, terminate. It was not pleasant to onboard people expecting to work with them for years only to terminate them a few weeks or a month later. Yet with a clean, transparent process it wasn't impossibly hard either. For the first few months it took most of our time. Since we were too busy with interviews, the company velocity actually dropped a lot: we weren't doing much of the "real" work, just interviewing and working on incremental parts with all candidates.

Yet it was all worth it in the end. After a few months, we connected with two leadership candidates in design and engineering that made our collective jaws drop from the stellar level of their craftsmanship. They weren't just x10 better than rest of the queue we seen so far: they were in a different dimension of quality. They brought radically new ideas for the product and helped bring totally new perspective on how to popularize our incredibly hard cryptographic concepts for non-technical users. We, of course, pounced hard on them (and now they have been with us over a year). The key fact is that we had that opportunity to pounce hard in the first place only because we kept the door in both verticals wide open: dozens of candidates who preceded them in the previous months since the company's' founding were tested, measured, and then, politely yet firmly -- terminated.