Hiring principles
Part 2 of a 3-part series on learning how to hire as an early startup employee: the principles & philosophies I adhere to when hiring.
Quite possibly the most valuable skill I've gained in the past three years as an early employee of a rapidly growing startup is hiring: hiring well (getting good candidates) and efficiently (with minimal pain). Hiring went from one of my least favorite parts of working at a startup to something I looked forward to, love to volunteer myself for, and miss now that it’s not core to my new job! In this post, I'll share the core principles that unlocked this transition, and which guide every action I take during a hiring process.
Before I figured out how to hire, my hiring processes took forever (many months) and were incredibly emotionally draining. But now! I love hiring! I also have very strong opinions on how to hire, and love telling them to anybody who will ask. So join me on my soapbox, as I share the core principles underlying every single decision I make during a hiring process.
This is the second in a three-part series on hiring. The first post covered the basic strategic decisions for hiring, this post focuses on my general hiring principles, and in the third I'll share the nuts and bolts of how I like to set up my interview processes.
Hiring principles
The following four principles drive every decision that I make when I develop a hiring process, talk to candidates, and work with my interview panels. They’ve become kind of a mantra, and I’ve found that anytime I’m stuck or a process isn’t going well, coming back to these principles can help me figure out how to get unstuck. These principles also led to an excellent success rate: both at finding great candidates, and also at having them accept our offers. Basically everyone that I hired on my previous team loved their hiring process, and shared after the fact how excited they were to join our team after their positive experience during the interviews.
Principle 1: Focus on signal
I come back to this one over and over again in a hiring process: it always comes back to the signal that you are or aren’t getting and making sure that you are enriching for and getting targeted signal.
When you're designing a panel or take-home, ask yourself: "will this give me the signal I need?" When you’re designing your panels: “what kind of signal do I want to get from the candidates?” When you're conducting an interview: "have I gotten the signal I need? Where else do I need signal?" When an interview process is struggling: "are we missing some key signal in order to make this decision?" If you're not getting signal from something, then don't do it. If you've gotten the signal you need, then skip everything else that gets at that same signal.
Importantly, the lack of signal is also signal itself: if an interview process isn't getting you the signal you need to confidently make decisions, then that's signal that you need to revisit your process. If your panel is consistently having trouble evaluating a specific skillset for candidates, that's signal that you might need a different collection of people on your panel or that you may need to revisit the questions you're asking.
Principle 2: Respect the candidate
Candidates are people, and more importantly they are people you are not paying. If you ask them to do an unreasonable amount of work or your process is unorganized and confusing, then that's pretty unfair to ask reasonable people to put up with for free. Excellent candidates won't stick around for a bad hiring process; they'll just go to another company who has their shit more together than you.
Candidates are also professionals who you are hiring because they are presumably good at something that you need. Respect their skillsets: don't put senior people through an insultingly basic technical interview, and don't put junior people through an impossibly difficult task.
Finally, candidates are people - respect who they are, don't expect work to be their entire life, and recognize that they're coming from a totally different place than you. Take the extra moment to self-reflect and make sure you're not dinging someone because they have a different accent than you. Ask yourself if weird communication during the interview could be due to nerves, and if so, try to call them in a lower stress environment to get the signal that you need.
Recognize that things which may be super simple to you are actually simple because you're an expert in your field, and in fact feel totally new and jargony to folks from other backgrounds. With technical exercises, I’ve found that the answer I consider easiest, simplest, and most obvious to get is usually the answer that only the best candidates get. They don’t know your tech stack or your schemas, so if they can get anywhere close to the “right” answer then you know they’re very good.
Think about what's going on in the world and adjust your expectations. For example, we had a data science candidate who'd had a baby right at the beginning of Covid, and one of our interviewers thought it was strange that she hadn't played with Covid data at all. I reminded my colleague that she had a baby during a pandemic, and so she was probably busy doing that instead of playing with data. (My colleague took my point well, disregarded that "signal", and we hired her!)
Have empathy for things going on in their lives. Candidates are often juggling their current job alongside the hiring process, which can be very stressful. Or they’re just living their lives, like you are outside of work. I once had a candidate totally ghost me for a phone screen, only to scramble and apologize because she was on vacation and had totally lost track of time. I didn’t ding them for that, because I would have done the same thing. Just remember that they’re humans, they don’t owe you anything, and they’re going through a stressful process for free.
Bad processes that disrespect the candidates' time, skills, and humanity will weed out excellent individuals who can just go to another company that is better at hiring. If your process doesn’t respect the candidates, then at best you'll be left with people who are really excited about your company and are willing to wait and take a bet on you. But at worst, and I imagine most often, you'll be left with mediocre candidates who can afford to wait because they're not getting hired anywhere else.
Principle 3: Keep it real
I try to keep every part of an interview process as close to what the candidate will do for their actual job as possible.
This is especially critical for any take-home exercise you have: you're asking the candidates to spend time on something for you, so it's only fair that they get something out of it as well. If you put in the extra 20% of time and thought into developing a take-home that gets at the signal you need and is also super relevant to what they'd be doing day-to-day, then at least they also get signal on what they're getting themselves into. It doesn't even have to take that much effort to do - for example, instead of asking someone to interpret a generic architecture diagram, draw up a toy example of something in your stack, give them all the context they need to interpret it, and ask the exact same questions you would have asked in the generic example. Or, instead of giving them a generic data science coding exercise, give them some of your toy data! Explain what the data that you usually work with looks like, give them a much smaller subset of it, and let them have fun with it. That way, you get to see how they code, and they get to see if they'd even like working with whatever type of data you have.
Keeping it real also applies in developing the panels: make sure that the people they will directly work with are on the panel, and be very explicit about it. When I'm hiring for a senior or manager role, I introduce the group presentation with "these are the people on your team. Ask them anything you want, they are literally the people you will work with/manage." I also love going through "hypothetical" scenarios which are not hypothetical at all - this gives you great signal on how candidates approach problems that they will actually need to solve, and it gives them great signal on the types of problems they'll be asked to solve.
When you keep it real, it all combines into a really fun experience: you give the candidate a watered-down version of a problem you are actually facing, and they get to work with you to find a solution with the people they would actually be working with. They are getting great signal on what the job is, and you are getting great signal at how they’d be at the job: truly a win-win.
Principle 4: Respect the panel
Finally, it's critical to respect the interviewers on your panel and set them up for success. When you're leading a hiring process, you're asking for a lot of time from a lot of people. It's critical that you make good use of that time.
As the hiring manager, it’s your job to set your panel up for success: be very clear about what the role is, where it fits within the company, how individuals on your panel would work with this person, what you're looking for in a candidate, and how the interview process will go.
Give the panel clear expectations and guidance: when you're talking about evaluation criteria, give your panel examples of a bad answer, a good answer, and an exceptional answer. I’ve often find that my panels are looking for the “exceptional” answer, especially when I’m working with less experienced interviewers. Being very clear that candidates don’t need to come up with the perfect answer (because how on earth could they?) to be hirable really helps move the process along to successful outcomes. And sometimes, candidates do give the perfect answer and you know that you should hire them immediately.
If you have multiple interviews in the process, make your intention for each one clear with everyone involved: what signal is each interviewer trying to get at? That helps interviewers know what they should be looking for, and avoids different panels gathering the same type of signal. Tell your panel how you'd like to receive their feedback, and give them the scale you're expecting them to use. I like to do a 1-4 rating, because it forces people to choose a rating not in the middle.
Respecting the panel also means keeping them informed about how the process is going. I like to make a slack channel for the panel (make it private, so that the candidate you eventually hire can't see any chatter about them) and share updates periodically. If you have a million phone screens coming up, let them know! If you're not finding anyone good, share that. And most importantly, if you've decided to extend an offer or have heard back from a candidate, let them know - your panel has put in a lot of time to this effort, and they deserve to know what's going on.
I was once on a panel where I interviewed someone who was not right for the role but a truly amazing catch for the company at-large. My co-interviewer and I submitted extremely positive feedback to the tune of "not hiring this candidate for any role would be a huge mistake". We didn't hear anything about this candidate until one day it was announced that she was starting in a few weeks for a role we hadn't heard about. I was thrilled that we'd decided to hire her, but felt upset that my time and enthusiasm weren't respected enough to warrant an FYI from the hiring manager as the negotiations had moved along.
It's also your job to keep the panel on track - if you're feeling that the panel is straying from your vision of the role, then it's your job to get everyone together to clarify expectations and build consensus on what everyone should be looking for. I interviewed a candidate once who wasn't quite the right fit, but did have some excellent pandas skills. One of my co-interviewers was so excited about this because they we our primary pandas expert on the team. But we had to remind them: "we're hiring for a data scientist comma automation, not comma pandas." They were bummed, but agreed that given the role we're trying to fill, their favorite candidate was actually not a great fit.
Your panel may also need to be periodically reminded about the expected leveling for a role. If the panel's feedback makes it clear that they're holding out for a unicorn or if their feedback is unfairly harsh given the level you're hiring for, it's useful to level-set when reviewing feedback.
Finally, respecting the panel means making sure that they feel heard and have buy-in to whatever decision you're making. This can be accomplished super easily, with synchronous sessions to review feedback after each batch of candidates who make it through the whole process. Even if the answer is a clear yes or no, I find it helpful to still get the panel together - if it's a no, the discussion can help elucidate how to improve the process next time or what signal to focus on more. If it's a yes, then meeting can build excitement and conviction before extending an offer!
That said, hiring someone is not a democracy. Not everyone's vote gets the same weight, and the hiring manager has the final say. If you're going to up- or down-weight someone's vote, make sure to communicate about that to the panel. For example, if I'm hiring for a people manager, I usually give their future direct reports veto power over hiring someone. (Of course, making sure that if they do want to veto a hire, it's not driven by bias, unreasonable expectations, or misconceptions about the role).
And that’s all there is to it! I’ve found that by adhering strongly to these principles, I can reliably lead successful hiring processes, develop useful technical exercises, and rapidly build consensus within my hiring team.
Focus on signal, respect the candidate, keep it real, and respect your panel: if you do these four things, then maybe you too will start to love hiring as much as I do!