Defining the role

Amateur managers think defining a role is writing a job description. Their first move is to go find one: a competitor’s careers page, another startup’s job board, or asking Claude or ChatGPT to write one. And they think the purpose of the exercise is so they can post the role online, or send it to investors and friends saying “I’m hiring an account executive,” so other people can send them candidates. That’s collecting a document, not defining a role.

Defining the role is determining what needs doing, what needs accomplishing, and then finding a person who has spent their career developing the competencies required to deliver on those outcomes.

You’ll know you’ve done it when you can describe the role in one or two sentences. That’s the mission: the M in the MOC, the Mission, Outcomes, Competencies document this whole skill produces. The novice version is “we need a head of sales to run the sales team.” The expert version is “build the sales team from two individual-contributor reps who report to the CEO into its own organization capable of driving $XM in revenue, complete with operations, predictable forecasting, standardized compensation and quota, one or two reps consistently exceeding quota, and a hiring pipeline so we can scale the team to 20 within eighteen months.” Same hire. Wildly different clarity.

The amateur alternative is thinking purely in functions: I need a salesperson, so what does a salesperson do? They have to cold call. They have to prospect their own leads. That’s a list of activities, and you can hire someone capable of every activity on it who still never gets you the outcome you need.

Think in outcomes instead. An account executive expected to close $1.5M in year one off pipeline handed to them by SDRs, and an account executive expected to self-source $4.5M of pipeline and close $1.5M of it, share a title and almost nothing else. Once the outcomes are clear, you can work out what competencies someone needs to achieve them, and then your interviews become a search for proof of those competencies. That ladder, outcomes to competencies to interview evidence, is how you de-risk the whole hire.

If you can’t articulate the outcomes, you don’t know enough to hire. Here’s what happens instead: you hire the person, you never tell them what’s expected of them, they don’t achieve what you expected, you get frustrated, and you fire them. They feel it was unfair, because it was. They never got a fair shot at a job nobody described. Everyone ends up pissed, you’ve spent a pile of money learning nothing, now you’re not sleeping through the night, and you’re asking your therapist why you’re so stressed.

Defining the role is also where you find out whether this should even be one hire. Especially the first time you hire for a role, what you’re looking for often turns out to be two jobs. The first hire a great head of sales makes is usually a sales operations person, not another rep, which shocks founders every time. Sometimes the answer is a contractor, sometimes it’s AI, sometimes it’s two people. The EOS folks call this “right people, right seats” [1]: a lot of role definition is seat definition, and only then do you go find the person for the seat.

And the written definition keeps working after the hire. The best hiring process on Earth still misses about three times out of ten for individual contributors, and about half the time for executives. I heard that from a talent partner at Andreessen Horowitz years ago, and a decade of hiring and coaching has convinced me it’s true. It should give you grace when you feel like you’re bad at hiring, and it should put real weight on being willing to start over.

Your role definition is your defense: with the outcomes and the 90-day expectations written down, you can make the keep-or-cut call against a standard instead of vibes.

The hardest part of being great at role definition is patience. By the time you’re hiring, your plate is already over full. The role exists because you have more work than you or your team can handle, so investing twenty or thirty hours (yes, 🤯, I know) in defining it feels like a distraction. Novices treat it like a task to knock out. Experts treat it like a research process, because they know the cost of the wrong person in the seat: the second interview process, the cultural damage of a rejected hire, the year of wasted resources, is so much greater than the cost of the time.

It’s a J-curve. For a while you invest time and energy with no visible payoff: drafting the mission, running calibration calls, arguing about outcomes. The payoff comes later, when the right person is in the right seat quickly because you didn’t mis-hire and didn’t have to replace them. The real failure isn’t even the firing: most people don’t fire quickly. They knew by day 30, 60, or 90 that it was the wrong person, didn’t have the conviction to act, kept them a year, and burned a year of resources because they didn’t want to spend thirty hours up front. Do that ten times and you may as well have told your board member to put their money in that homeless guy’s burn barrel.

People who are great at this know not only what they need, but also what they want. The outcomes are what you need. What you want is someone you actually like spending time with. New managers carry a twisted guilt about this, as if wanting to like your team is unprofessional. The airport test is the old screen: stuck for ten hours in an airport with this person, do you groan? You don’t need a best friend, but tolerating someone who makes your skin crawl is never an acceptable hire, and it’s amazing how many founders tolerate it twice or even more.

Here’s the rubric. Find yourself on it honestly.

The first step is realizing you need to, and you’re doing that right now. You’re here! That counts.

Second, price the consequences of staying the same. This comes from two guys who studied change for a living, Lewin and Schein. They had a theory of change [2]: motivation to change has to outweigh the comfort of the current pattern, so make the current pattern uncomfortable. If you keep defining roles the way you define them now, you’ll keep hiring the wrong people and keep firing them. Put a dollar amount on that. A $150K hire you replace at month twelve didn’t cost you $150K; it cost you the salary, the recruiter, the second interview loop, the team’s morale, and a year of the outcomes (and upside) nobody delivered.

Third, read one good treatment of the craft: the a16z hiring process doc [3], or the MOC tool right here.

Fourth, on your very next hire, build an MOC before the first interview. Full stop. And use an AI to help you. Don’t ask it to write a job description, which is just outsourcing the thinking. Have it interview you: paste the MOC tool link into Claude or ChatGPT and tell it to ask you the questions one at a time. Most people find reacting and editing far easier than writing from a blank page, and the thinking stays yours.

Fifth, if you don’t have an open role right now, practice on one you already filled. Pick one person on your team and write their MOC: the mission of their seat, the outcomes you expect over the next twelve to eighteen months, the competencies the seat requires. Don’t try to do all eight of your directs; pick one. You’ll be surprised how much clarity it gives you about managing them, and it’s the rep that builds the muscle between hires.

Sixth, wire it into your hiring process so the skill survives contact with urgency. At five people you probably don’t have a hiring process. At twenty you might. At fifty you’d better, and role definition needs to be the first station in it, so that the next time someone says “we need to hire,” the system reaches for the definition before the job board.

Seventh, calibrate in proportion to your ignorance. The first time you hire for a role you’ve never held, never managed, and maybe never seen done well, you can’t define it alone, and you don’t have to. CEOs of fifty-person, $10M companies ask me “Andy, how do I know what good looks like for this role?” You run the same fucking process you’ve always run: five or ten calibration calls with people who’ve done the role, hired for it, or worked beside someone great at it. If you’ve hired fifty people into this role across your career, two calls might do. The calibration call is its own tool; the point here is that the number of calls scales with how new the role is to you, and this is something you’ll do over and over for the rest of your career.

Eighth, if those steps sound unappetizing and you don’t want to do them, I suggest you find another job.

Start with the money, because the contrast is genuinely funny. Just about any hire is going to cost $100K and some as high as $400K+ a year. Plan on keeping them four years and you’re staring at a $400K to $1.6M decision before you count the value they’re supposed to create.

Now: if you were buying a $2M house, you’d research school districts, fight about which direction the windows face, and pay a professional inspector before signing. For the multi-million-dollar hire, the same founder spends two hours, grabs a competitor’s job description, and posts it. That’s what a couple of Penn researchers named cognitive surrender [4]: handing your thinking to an external artifact. Surrendering it to a job description you found is no different from surrendering it to a chatbot. Either way, nobody in your company did the thinking on how this person generates a return.

The flip matters just as much: refusing to use AI at all is its own failure mode. A lot of people aren’t great writers. They open a doc to define the role with the cursor blinking and decide to answer an “urgent” email instead. A scaffold you can react to and edit beats a blank page [5]. The line to hold is simple. Use the tools to draw the thinking out of you, never to do the thinking for you. Great craftsmen use great tools.

The other failure modes, in roughly the order I see them:

1. Not doing it at all. Cowboy hiring. See the rubric; it’s the bottom for a reason.
2. Building the role by consensus. Take input from everyone who’ll touch the role, absolutely. But if you assemble the definition by committee and the hire fails, the outcomes were still yours, and now you’ve failed with a role you never believed in. I’ve seen far more leaders regret not listening to themselves than regret trusting their gut.
3. Hiring who you think you should hire instead of who you deeply believe you need. If this feels redundant to 2, it is, but it is repetition for emphasis because so many people screw this up. Early founders talk themselves out of their own judgment: “I’ve only been a founder a year, what do I know?” You’ve been a person a lot longer than that. You know what works for you in other people, and great teams have to actually work together. Trust it.
4. Tolerating the skin-crawl. Hiring someone whose competence looks right on paper while your gut says no, then doing it again with the next hire. Some people never learn that skin-crawl is disqualifying.
5. Never blocking the time. “I’ll get to it this week” is how the urgent eats the important. Twenty to thirty hours of definition work doesn’t happen unless it’s on the calendar and defended.
6. Treating the definition as permanent. Last year’s MOC for this year’s very different company.

In the end, the failures are a tragic comedy. I call it “being the warden of a prison of your own making.” I watch constantly as a coach: the sheer volume of bitching leaders do about the people on their teams, next to the excuses they make for skipping the work up front. It’s one thing to bitch about your team when you defined the role, communicated the expectations, and ran the process. It’s another thing entirely to skip the definition, communicate nothing, and then blame everyone around you. Most relationship problems on teams are exactly this: not knowing what you want, not saying what you need, and then resenting people for failing a standard that lived only in your head.