Company Building: My Playbook for Startup Ops
A Smart Generalist's Guide
Welcome to an insider-only edition! This is the eighth installment of Silicon Valley Startups, following:
The Basics (VC 101, Startups 101)
Finding a Career (How to Found a Silicon Valley Startup, How to Break Into Venture Capital, How to Join a Rocketship Startup, How to Choose)
Crushing Your Job (Getting Shit Done at Startup Speed)
This week, we’re talking about a topic near and dear to my heart: how to build out the operations of a rapidly-scaling company.
💼 This is the playbook that I would run if hired as the VP of Ops at a 30-person Series A company that aspired to 10x within 18 months.
It’s largely what we did at Astranis while we grew that quickly — except that it avoids the mistakes I made the first time around.
The scope of this playbook includes many of the core business functions, and focuses where I think the primary value lies: in recruiting, onboarding, culture, and team management.
Many important parts of scaling up are not included here. This guide does not include engineering, sales, or fundraising; it also doesn’t include much finance, IT, or legal. These functions are likely to vary wildly depending on the industry and the specific technology of the startup. At Astranis, for instance, all of our contracts that are, individually, worth high tens of millions of dollars — when we raised our Series C, we had exactly one customer. Compare that to a normal SaaS product that sells for $5 per month, and that needs hundreds of thousands of customers to achieve that same scale. The IT/infra needs, legal requirements, financial regulations, sales cycles, team composition… it would all be wildly different.
But all startups need to recruit an amazing team, motivate them to do their life’s work, and build a managerial infrastructure to keep the team happy and productive at scale.
I joked at Astranis that the engineer’s job was to build the satellite, and it was my job to build the company. Here’s how I’d build a high-powered company, if I was going to do it again.
🏎 How to recruit hundreds of specialized engineers: structure and speed.
In the early days of a startup, the hiring bar is extremely high. Astranis asked early employees to do upwards of 40 hours of work on a highly-technical project before even bringing them onsite for a final interview. Such an intense effort is impossible to maintain if you hope to bring dozens and then hundreds of people onto your team, which poses the core challenge of recruiting at scale:
How do you reduce the amount of due diligence per candidate without lowering the bar and hiring less talented people?
The answer: structure and speed.
⚙️ Structure
Structure is a standardized, optimized way to test if people meet your hiring bar. This includes:
A clear definition of the bar for making a hire
A simple grading system that has an even number of possible grades (you can’t give people a midpoint, or they’ll pick it far too often)
The standard steps of the interview process, including lots of detail on the final round (the on-site)
The standardized communications that people receive at every step of the process
A guide to sourcing candidates for all hiring managers (the recruiting team will be the gold standard here, but you also want mere mortals helping; the “How to crush LinkedIn Recruiter” training)
The roles that people play throughout the process (e.g., hiring manager, culture interviewer, technical interviewer)
What a job requisition looks like, and how you get new ones approved
How to get recruiting resources assigned to a job requisition
Guidelines for how to engage with candidates, and pro tips on how to sell them on your company
A standard flow through the Applicant Tracking System (e.g., Lever, Greenhouse) — how to leave a grade, how to take notes, etc.
Training on all of the above
There’s obviously far too much detail to include all of the above in these pages, but here’s a taste of what this structure might look like.
The job of the recruiter is to fill the top of the funnel for hard-to-source roles by convincing talented engineers to apply. Once people apply and pass the written screen, the interview process begins. As long as the job requisition is open, the hiring manager and the assigned recruiter are best friends and meet multiple times weekly.
The goal of the interview process is to uncover evidence of exceptional ability. Candidates are rated at every step of the process from 1 (I would quit if they were hired) to 4 (I will fight to hire them). A 3 rating (good, not great) is enough to continue a candidate through the process, but a 4 rating is absolutely necessary for anyone to get hired. (All 3s = no hire.)
After a series of phone interviews, the candidate is brought on site, where they meet the team: the out-of-discipline leader responsible for evaluating cultural fit, the hiring manager, and their prospective teammates. Assuming they get an offer, the closing process begins: the candidate is given a verbal and written offer, and connects with an exec tasked with closing candidates. The process continues (i.e., candidates are still actively sourced and screened) until a job offer is signed in ink by both parties.
This is an involved process, and requires alignment across the organization about how it all works, where responsibilities lie, how the hiring bar is defined, etc. This is why training is the most important part of the process — and why speed through the pipeline is the #1 goal for the recruiting team.
⚡️ Speed
Recruiting is a numbers game. During our period of intense growth, Astranis hired one person for every ~100 that applied. We hired about 250 people during that time, which means that the recruiting machine we built got about 25,000 people to apply to join the company.
Processing that many people requires targeted logistical effort and ruthlessly difficult grading. The recruiting team will eventually hire dedicated recruiting coordinators just to keep all the logistical plates spinning; in lieu of those folks, and for the majority of roles in the beginning, the recruiting process will need to do two things: 1) set a clear standard for the turnaround time between stages of the process; and 2) set up communications tools and templates that drive the process forward as efficiently as possible. A good place to start is that no more than 48 working hours pass between stages without engaging the candidate — if a candidate passes their first technical phone screen, a date and time for the next interview is scheduled within two days. This isn’t that hard to achieve with an engaged candidate; calling them to set up a time the day after the interview will usually do the trick.
But even with the world’s best logistics engine, you still need to weed down the pipeline early. Your team of 30 people can’t do 10,000 first round phone screens, so ideally the percent of people advancing from stage to stage gradually increases: maybe 10% of applicants get interviewed, then 20% pass from interviews to an onsite, and 50% get an offer after the onsite. (That’s good for a 1% applicant-to-offer conversion.)
To hold that high of a bar, you’ll need to convince your team that false positives (people you shouldn’t have hired, but did) are infinitely worse than false negatives (people you should have hired, but didn’t). Bad hires are a huge waste of time and money. Worst of all, they suck productivity from managers and teammates alike. Avoiding bad hires is absolutely worth the pain of continuing an interview process for a job requisition that might already be dragging on — settling is not an option, no matter how tempting it might seem at the time.
If you can run a tight process, hold a high bar, and keep candidates flowing through the funnel quickly — you’ll recruit 200 engineers in no time.
One quick failure mode to avoid: overselling
As an executive in charge of hiring, it’s tempting to try to close every candidate, no matter what it takes — and therefore to lose sight of the fact that what you’re ultimately looking for is a fit between the candidate and the company.
Selling a candidate on a version of reality that they are unlikely to experience at the company (e.g., saying that the work is more stable than it really is) will come back to bite you, hard. You’re hiring a smart, discerning person. If they really don’t want what your company has to offer, they’ll join and then leave when it doesn’t meet their expectations — which is unfair to them, because you’ve wasted their time, and unfair to your team, because you wasted theirs as well.
🍷 How to get new teammates up to speed quickly: establish and teach cult(ure) tactics
Within my first few weeks of joining Astranis, I took on a significant task — lead an hour-long session at the upcoming company retreat to help define our company culture.
This was a tall order. I didn’t know anyone at the company yet; I had never worked at a startup before (other than tiny ones I had started); and I had never worked alongside hardware engineers before. And yet, this was a task I fought hard to get assigned, so I dove in headfirst and did my best to get shit done.
Over a two-week span, I made it my goal to interview every single one of the first thirty employees at Astranis.
I asked them a series of questions designed to help me figure out what made Astranis “Astranis”:
How do you answer people who ask you what you do for a living?
Why did you choose to work at Astranis?
Are you worried about anything that might change at Astranis as it grows?
What's the most likely thing to stop Astranis from achieving its full potential?
Is there an Astranis “type”? As in, some criteria that you would apply to say someone will or will not succeed here.
…and more
These conversations were incredible for what they both contained and omitted. I learned that Astranians loved to learn from each other and teach each other, but that they didn’t have a set of company values that everyone knew. I learned that almost exactly half of the first 30 Astranians were referred into the company by someone they knew, but that most of them thought there was no such thing as an Astranis “type.” Many folks mentioned the mission as a core reason they joined the company, but their answers for what the mission actually was were non-standardized (but, granted, they were fairly close to each other).
The conversations also gave me an endless supply of hilarious and heartwarming quotes to kick off my presentation to the company — I turned it into a game, where I’d show a quote (“Astranis is a place for people who are happy doing hard things and can do hard things”) and then make people guess who said it (Hall, who went on to found a YC-backed startup).
The best quote of all, though, came from my boss at the time, our VP of Operations and Strategy, Miki Heller. She said:
“Cults do a lot of things right. We shouldn't poison ourselves at the end, but everything else up until that.”
And she is right. I shouldn’t have had to talk to thirty early employees before I got a sense of what the culture was like at Astranis — any cult leader would be ashamed to find it was that hard for a new initiate to learn what their cult was all about. Startup company cultures could learn a lot from cult leaders!
Every employee at every fast-scaling startup should be great at answering the question “who are we?”
Ideally, everyone should have the same answers to the following questions:
What are our core beliefs about the market, the world, and the right way to do business?
How hard do we work? Why do we work so hard?
What are our rituals? Why do we do them?
Who are our teammates? What makes someone a great fit to work here?
Where did we come from? What stories from our founding exemplify who we are today?
These might seem like challenging questions to answer completely honestly, but, luckily, no company mythology is completely honest. eBay was not founded because the founder’s wife wanted to sell her Pez dispensers; Elon Musk was not present at the founding of Tesla. You shouldn’t make anything up, but you should feel totally comfortable being artistic in your storytelling. Because that’s really what culture is: a shared story that defines who you are as a company.
If your culture is strong enough…
Employees will know the kinds of behaviors that are rewarded and punished at your company. (Aside: this is why I prefer “Core Behaviors” to “Core Values.”)
New employees will feel like they are part of the company, and will have no trouble answering to their family/friends what it is they do
New employees will become productive faster. (Some core cultural artifacts might include a “company dictionary” that explains your slang, a “for dummies” guide that explains how the product is built at a high-level, and a “company directory” that explains who is who.)
Employees will have a standardized onboarding experience that includes the core trainings they need on their first week, and a standardized orientation experience that introduces them to company leadership within their first month of joining.
And if your company culture is solid internally, it can also become your external brand. The best example of this is Anduril, a startup worth $8.5 billion that builds hardware and software for the military.
After years of their founder explaining why he was right and the defense primes were wrong, Anduril published their mission to the world.
It’s an epic piece. It includes non-obvious guiding principles, potential sources of failure, and ways they intend to make major progress.
It exemplifies everything you want in a cult startup mission:
Not everyone will like it, but the right people will love it (meaning: some people who you don’t want at the company will choose not to apply when they read it — an awesome success, because it wastes nobody’s time!)
It positions the company as a central figure in an epic, still-unfolding history
It makes clear what the company stands for, and why
And it’s written down for the whole world to see
You might not be able to go this far, but as a leader at a fast-growing startup, you absolutely must write down as much as you can to define your company culture.
Small companies create, share, and solidify culture easily. Folks certainly know everyone else at the company, and probably know what they are doing; the whole team can fit around a reasonably-sized lunch table to two; early employees have already known each other for years (perhaps even from before the company’s history); and, let’s be honest, early startup employees are a special kind of crazy for agreeing to join your fledgling startup in the first place.
But none of those small-company cultural enablers hold at scale.
You will not know what everyone does anymore. You won’t even know everyone’s name. You will not talk to everybody daily, weekly, or even monthly. Messages have to filter through managers to truly be heard, which dilutes your direct access to people as a founder. Your success will attract employees that are seeking you out for the wrong reasons. Most new hires will not know anyone at the company when they join.
So, if you only rely on the informal culture definition that worked when you were 30 people, you will not have a strong company culture at 300 people.
You just won’t. It’s impossible. It will feel weird to start defining the culture when you’re still just 30 people — it’ll feel like explaining the punchline to a joke — but it’s far, far better than the alternative: hiring the wrong people, struggling to explain company decisions, and losing employees when they learn that your company isn’t for them.
👩💼 How to scale up team management: defining success, and coaching your leaders
Small companies love talking about their flat org charts. It’s a sign of pride, and a selling point for early employees. “You won’t be buried underneath layers and layers of managers, you’ll have direct access to the executive team!”
As a company scales, however, it has to reckon with the facts that 1) having an effective people manager is a crucial part of getting the most out of each employee, and 2) most managers hit a hard limit at about 8 direct reports as the maximum they can effectively lead. That fact — not a love of bureaucracy or a Big Company Conspiracy — is why big companies have lots of middle managers.
If your company is less than 74 people, you easily have only a single management layer (individual contributors [64] report to managers [8], managers report to the CEO [1]). But any more than that, and you’ll need to introduce new layers of people managers; and if any particular discipline (e.g., software engineers) grows to have more than 64 employees, yet another layer will have to emerge! In practice, companies resist this for a little bit too long, having some managers stretch and take on more than 8 direct reports — but it’s a losing battle.
Layers of management are a necessary evil that comes with scale. But there’s also good news: management training is perhaps the highest-leverage activity you can complete as an exec at a fast-scaling startup.
You need to teach your managers what it means to be a manager, ideally before they sign up for the job:
A manager’s job is to keep their individual team members happy and productive, while being a loyal representative of the company.
Every employee is different in their particular circumstances, which is why managers have to be humans, not robots — but all employees want the same three things:
They want their work to matter to them — if they work hard, they want to know that they will be rewarded
They want their work to matter to their company — if they work hard, they want to know that the company will be better off
They want their work to matter to the world — if their hard work makes the company better off, they want to make some kind of positive impact on the world
A manager’s job, then, is to make sure that each individual employee understands how each of the above are true. If they succeed, their direct reports will be maximally productive and happy to be working so hard.
This will mean very different things for different people. Some employees will want to know a lot about their compensation (e.g., recruiters or salespeople who have performance-related comp), others about training opportunities (e.g., early-career folks who want mentorship and opportunities to learn new skills), others about career growth (e.g., people who someday want to become startup founders).
You’ll need to train your managers to have both inspirational and tough conversations about each of these topics — and many more! —with their direct reports. Because, sadly, you won’t be able to be the single point of contact for all important things anymore.
As the company scales, each employee will lose direct access to the senior leadership of the company, so their managers inevitably become “the company.”
This reality will be met with resistance, particularly by new managers who don’t want to be “the man,” and especially those who have the brilliant, unique idea to try to manage their teams more as peers/friends than as a big, mean, overly formal boss.
But a critical message that must land when training up your managers is this — as a manager, you are a representative of the company. You can disagree with company decisions to your Director and in the meetings where things are being decided, but team leads cannot prioritize their individual relationships with their direct reports over the best interests of the company. Doing so is destructive for the company and for the manager.
Say, for example, that an employee is unhappy about a benefit that used to exist, but is being taken away. It is easiest and most natural for the manager to empathize with the employee, saying “yeah, dang, I also really liked that benefit!” But doing so is a trap. By empathizing, the manager is conveying a number of messages that they don’t intend: most notably, that the company is wrong/mean, and — counterintuitively — that the manager is ineffective at advocating for their direct reports on matters of company policy. The correct way to handle such a conversation is to explain the logic of the change and why it’ll be better for the company and therefore the employee, but that’s not always obvious to a new manager.
Making sure that managers are ready for such conversations requires you to build new communication channels to distribute messages to your managers. You’ll spend much more time on “messaging” for major changes. As inefficient as that seems, it’s better than the alternatives of different teams receiving different messages, or of nobody getting any official message, which allows the rumor mill to churn.
If you nail manager training, you’ll get better information about the company and your employees will be happier without needing your direct involvement.
There’s a lot of work to do.
As you scale a startup from 30 to 300 employees, your job is difficult, but simple. You need to build a machine that adds talented, motivated people to the team; that integrates those people into your culture; and that keeps them happy and productive.
To do so, you’ll need to put together a lot more structure than you’re used to needing. At a minimum, you need:
An ironclad recruiting policy
A clearly documented mission and core behaviors
A one-day orientation program that allows senior leadership to introduce new hires to the company culture
An onboarding program that gets new hires contributing quickly
A training program that teaches people managers how to keep their employees happy and productive
And, you know, a product that people love. And the support functions around it. And an executive team (a topic for another time).
You can do it. Start with the cultural stuff first. Let me know if you need help.