Hiring and Retaining Talent
Let’s start with hiring. Unless otherwise specified, “talent” in this piece refers to engineers of various kinds — UI/UX engineers, software engineers, and so on — because startups are most starved for them… The advice may or may not apply to other positions; judge for yourself.
Hiring
How to Break Out of the Hiring Predicament
Against the backdrop of a booming internet industry, even the BAT giants routinely roll out all manner of “talent programs” to poach the elite, so finding the right talent is now harder than ever. For a startup, paying money to post listings on the big job boards and hoping that a brilliant, good-looking, personable soul with modest demands will willingly wade into your mud pit — that’s about as likely as a “short, ugly, poor” guy winning over a “fair, rich, beautiful” girl. Of course, there will always be the idealistic, ambitious soul unwilling to bear the golden handcuffs and dead set on the startup road — but such people are as rare as phoenix feathers and qilin horns, to be chanced upon but never sought. So broadly speaking, for startups today may be the best of times for raising money, but the worst of times for hiring. The path of experienced-hire recruiting through the traditional job boards just doesn’t work well anymore.
Two years ago, we managed to attract three or four promising young talents through Innovation Works’ campus recruitment, hailing from the Chinese Academy of Sciences, Peking University, Beihang, and elsewhere. We finally overcame our psychological reservations and offered what I considered a very good package for a fresh graduate — only for them to be snapped up overnight and carved away with hefty pay by deep-pocketed giants like NetEase, Tencent, and Microsoft. Even leaning on Innovation Works as a great sheltering tree and drawing on its influence among students, this was the outcome for us. How much harder, then, for those startups with no name, no capital, no connections — only a dream — to win anyone’s favor. So the campus-recruitment road is a hard one to walk, too.
So startups need to explore other routes.
The first route is to mobilize the team’s personal networks to find talent. This is slow, painstaking work that pays off over time — it can’t be rushed. So the technical co-founder should be constantly expanding his network in tech circles, preparing for lean times while things are calm — like the fable of the grasshopper and the ant that everyone read as a child. Don’t wait until winter sets in (you need to hire) before remembering to store up provisions (expand your network).
Of course, expanding a tech network isn’t as simple as meeting someone, swapping business cards, and exchanging pleasantries — otherwise you could just let the BD folks handle it. Why bother the technical co-founder, someone who isn’t particularly silver-tongued and finds it hard to warm up to strangers quickly?
Engineers are a proud, unruly breed. They bow only to technical skill, technical vision, and the ability to ship projects — not to a glib tongue. Toward empty talk they’ll pay lip service on the surface, but inwardly they make their own cool judgments (ah, the engineering men and women of the world~), and it’s hard to win their genuine buy-in. Only tech-related discussion — or better yet, live, real code — can hit their sweet spot. Engineers tend to approve of the person first, and only then approve of the project. And that approval of the person builds up gradually, through repeated interactions.
The second route is a roundabout one: hiring interns as a means of acquiring talent. Post notices on university forums like SMTH to attract students in the second half of their junior year or their second year of grad school (how to get a forum ID we’ll set aside here), then seduce them — no, tempt them with perks. Big companies generally have standardized, company-wide intern pay. If they offer 120 a day, you offer 150; if they offer 150, you offer 200. For a few thousand yuan a month you land yourself an uncut gem — happy now? Students approach internships and full-time jobs with different mindsets. Going to a big company for an internship means being one radish in one hole — no room to show your talent, not much to learn. But going to a startup, so long as the technical boss is solid, basically guarantees a lot of growth. When hunting for a full-time job, though, they want stability and a high salary; considerations like growth get set aside, because soon enough they’ll be playing a very practical game of houses and cars with their future mother-in-law. This intern-hunting strategy is actually a lot like Tian Ji’s horse-racing gambit — so long as you’re willing to put in the effort, you can land some very fine seedlings.
The scary thing between people is having no point of connection; once you have one, many things become much easier. Once an intern starts working, if the technical co-founder mentors them well — teaching them hand in hand and building a bond, letting them recognize the company’s grand vision through the projects they do — then when the internship ends, the odds of keeping them with a reasonable salary are far greater than the odds of landing a campus recruit with a lavish one.
There are other routes too: trying newly emerged recruiting platforms, using LinkedIn to flirt with people you’ve got your eye on, and so on. In short, the technical co-founder has to deploy the full range of his technical and personal charisma before the hiring predicament can gradually be cracked.
How to Interview
When hiring, interviewing is a matter of technique. Getting the most complete picture of a person in the shortest time is a real test of the interviewer’s ability. While the team is still small in its early days, the technical co-founder must step up as the gatekeeper.
The purpose of an interview is to win talent, not to show off a sense of superiority — “I’m better than you.” I’ve seen interviewers who take pride in stumping a candidate, even provoking them, and dress it up with the euphemism “stress test.” I don’t like this approach — I want the interview to let the person show their greatest ability as much as possible. So I start by lightening the atmosphere, chatting about everyday things, finding common ground. Then I ask some foundational questions and dig into the points on the résumé that interest me, little by little, drilling deeper step by step. Foundational questions aren’t easily covered by faking or memorizing answers; even if they happen to have memorized one, give a bit of context, probe a little deeper, and the truth will show. Many candidates struggle to clearly explain the concepts and use cases of lock-free and wait-free. Even if they manage to stumble through those, ask them to describe how to build a lock-free queue and they’ll most likely get stuck (they often won’t even think to ask me about the use case — and even after I hint that they only need a single-producer, single-consumer lock-free queue, they still can’t continue).
Fundamentals, fundamentals, fundamentals. In my view, solid fundamentals beat work experience. Work experience can be coasted through — what does ten years of it really prove?
Ask plenty of open-ended questions in an interview, have the candidate describe scenarios, then ask follow-up questions based on that. For example: What’s the hardest problem you’ve solved? If I asked you to do xxx, how would you think about it?
Don’t forget to give the candidate a chance to show their real strengths during the interview — because sometimes your questions may land in an area that’s alien to them. Once you’ve asked enough and formed a preliminary judgment, hand the floor over to them: “I’ve asked quite a few questions. Is there anything you consider a strength of yours that I haven’t touched on? Care to share?”
Finally, don’t forget the “written test.” After all, you’re hiring someone to get things done, not to talk a smooth game. I usually give out some simple problems (including a project) and ask the candidate to take them home and finish as fast as possible. The finished code must be robust and come with test cases; the project must also be deployable as a working version on Heroku. What? Never touched Heroku? Well, that’s exactly a chance to test their learning ability…
I don’t like written tests done on paper. On a whiteboard, I’ll only look at pseudocode and architecture diagrams. Real code has to be written on a computer, made to run, with inputs and outputs, with tests. Once I have the code, I first check whether the test cases are thorough and whether boundary conditions, exceptions, and edge cases are covered. Then I manually feed in a few sets of test data to see if it crashes, and only after that do I read the code: whether the logic is correct, whether it’s clean, tidy, robust, commented, PEP8-compliant (for Python), and so on.
The written test is a place where you can play plenty of tricks — so long as the interviewer is creative enough. Moreover, I generally don’t have candidates do the test on the spot; I send it home with them. Three considerations here: 1) On-site test time is limited; you can’t produce anything meaningful in an hour or two. 2) I deliberately pick some libraries the candidate may never have used and require them in the project, to test their learning ability. 3) If they never submit the code, either they’re not interested in the company or their ability is limited; if they do a great job and can clearly explain the details of their own code, they’ll make an excellent team member.
If you end up with only one headcount but several strong candidates and you’re torn, pick the one whose thinking is clearest and most rigorous. How do you judge that?
-
Their expressiveness in the interview — the ability to explain something clearly and understandably
-
Their writing ability (many people keep blogs these days, and the essays there reflect their writing skill)
Retaining Talent
Hiring is only the first step in acquiring talent. Next comes retaining it — for as long as possible — in the service of the company. The first priority for retaining people is of course that the company be attractive enough: for instance, the business booming and revenue rising with the tide; or, even if not yet profitable, being in a phase of opening new frontiers and capturing new territory; at the very least, it must offer all sorts of stories, expectations, and every possible kind of beautiful future. If none of these apply, and every possible upside has already been exhausted, then the only thing that can keep people around is the founder’s character.
Character
Putting “character” here is a little absurd. But it’s the truth. Startup teams are different from ordinary corporate teams; in the context of a startup, the founder’s character can have an extraordinarily strong rallying power.
There’s a rather crude old joke about the “four ironclad bonds between men”: “schooled together, soldiered together, whored together, looted together.” Vulgar, but it has a point. When I left Juniper to start a company in 2011, even if I’d raised my arm and rallied my team — who would have come with me? Absolutely, positively no one. Colleagues at a corporation treat each other with the courtesy of guests; however pleasant the collaboration, there’s always a veil between you — even writing an email means agonizing over whether an exclamation mark is too much, whether you should add a little garnish, tack on a smiley. So however good a manager’s character, their rallying power is only so-so. A startup team is different: coding together into the small hours, arguing over requirements until flushed and red-faced, playing Ultimate Frisbee until teeth are gritted — this is a bit like “schooled together, soldiered together.” Perhaps one day the “four ironclad bonds between men” will add one more: “started a company together.”
This is also why some founders of good character still have people who won’t abandon them even when they’ve hit rock bottom.
Character stands for a lot of things. My former boss Frank says that in NetScreen’s success, “spreading the wealth so the people gather” was one of the key reasons. Here, “spreading the wealth so the people gather” reflects one aspect of character: not being greedy. If a founder forsakes greed, casts off anger, is free of delusion, keeps a low profile, isn’t arrogant or dismissive toward others (no pride), and acts with resolve and decisiveness (no doubt), then that’s good character — and people will be willing to follow.
Providing Room to Perform and Grow
Assuming reasonable compensation, why does an employee stay on a team? It comes down to room for personal performance and growth. For engineers hired laterally, room to perform is critical; for fresh graduates or interns, room to grow fast is critical. Only by understanding these needs and responding to them correctly can you lay the foundation for retaining people.
If you’ve worked at a big company, you’ll notice there are quite a few capable young talents there who are frustrated and unfulfilled. Give them a stage, draw out their potential, and they’re top-notch at what they do. But alas, wearing golden handcuffs, their edge worn away, shuttling every day among email threads and piles of meetings, they lose themselves. If you happen to pick up such a person for a steal and they join your startup, it’s crucial to let them fully exercise their abilities and find the right stage for them. Many companies talk about providing “a stage with no ceiling” for employees; few actually deliver.
Fresh graduates and interns are usually uncut jade waiting to be carved, requiring serious investment in training. Many startup teams think “time waits for no one” and have no time to train — but that’s not necessarily so. My firsthand experience is that their eagerness to learn is very high; to finish a task, they’ll spare no effort. I’ve mentored several outstanding interns; given sufficient training and help, their combat effectiveness is no worse than full-time employees’, and the energy and positivity they bring greatly ignite the whole team.
When training fresh graduates or interns, don’t hold back on anything work-related. There’s a well-put saying: If you don’t provide enough context and help, an employee’s failure is the boss’s failure (the translation is a bit clumsy). I usually start with the overall architecture — what components we use, why we made those choices (or compromises) — then gradually move on to the code. In daily work, answer their questions tirelessly, even if they seem very naive — what is youth for, if not a little naivety~
Many people worry that interns will get their hands on too much of the company’s core code — what if they leave after a few months? This is the man of Qi worrying that the sky will fall. Never mind that everyone signs an NDA when they join; even without one, a startup still struggling on the edge of survival may not even have a clear idea of what “core” is. Worrying about such off-key things is truly unnecessary. Trust is a two-way street; in personnel matters you must trust the people you hire — interns included. Of course, necessary permission controls are needed — but those should be role-based, not aimed at any particular person.
Once interns hit their stride, offer them a reasonable salary, and the odds they’ll stay after graduation are very high. After all, you’ve had several months to show your personal charisma, the company’s vision, and the value of the product you’re building. If that still isn’t enough to keep them, then dangling higher pay to force them to stay is pointless.
Invest Emotionally
Emotion is a subtle thing, and often the decisive factor. Grab an employee for a cup of coffee now and then, chat idly, talk about life, about ideals, about books — that’s making a fresh deposit into the emotional bank account. As a boss, it’s best to have a fairly detailed understanding of the family situation, emotional state, and life situation of everyone on the team (or at least the key employees). They say a relationship is half built by spending time together and half by talking. Though you spend every day together, coding together, debugging together, hackathon-ing together, a bond that’s never had a deep conversation always lacks that final kick — it never solidifies. ^_^
Emotional investment is a long, steady stream; it’s like dollar-cost averaging into a fund. And that investment must be heartfelt, with no thought of return. When major changes hit, there’s no need to play the emotional card — if the people are right and the investment has been enough, what should be there will be there; otherwise, even playing the emotional card is useless and may only wound both sides.
