The hardest part of cutting from 68 to 24 employees was not the conversations themselves, though those were brutal. It was realizing, somewhere in the middle of that process, that almost every hire had felt right at the time.
Not reckless. Not obviously wrong. Right. Each one was someone talented, energized, who believed in what we were building. The signal that you're over-hiring does not announce itself. It just quietly narrows your options until the math becomes inescapable.
At Realm, after our Series B closed, we scaled aggressively. We went from a lean team to 68 people across San Francisco and Copenhagen. It felt like graduation. The product was working. The community was growing. The team was energized. Then the fundraising market tightened. The conversations that had been warm went cold almost overnight. We didn't have the revenue to support 68 people without the C-round, and the C-round wasn't coming. The math was simple and brutal.
What I've carried from that experience is not a rule about headcount — it's a discipline about the hiring decision itself.
The bar most startups set wrong
Most early-stage founders hire when they feel the pain of not having someone. The logic is intuitive: you're stretched, something important is slipping, so you hire to address it. But this bar is set wrong in two directions.
Hire too reactively and you bring in people before you understand the role. The job description was written under stress, the interview was compressed, the onboarding was improvised. Six months later, the person is doing a version of the job you needed but not quite the one that exists now.
Hire too proactively — "get ahead of the need" — and you add coordination overhead before you have the problems those people are meant to solve. The worst version of this is the stage-inappropriate hire: bringing in a VP of something because that's who you'll need at the next stage. But a great VP of Sales at Series B builds and manages a sales team. At seed, you need someone who can close deals personally. Those are different people, often with incompatible instincts about what success looks like.
The question is not "do we need this person?" It is: "what specific, measurable thing is not happening because this role doesn't exist yet, and does it matter to the business right now?"
If you cannot answer that question with a specific answer — not "we need more engineering capacity" but "we cannot ship X by Q2 without another backend engineer, and X is the primary thing blocking our next funding conversation" — the hire is probably premature.
What I learned at Nokia that I didn't appreciate until Realm
Before Realm, before any of the founder narrative, I spent over a decade at Nokia. Software configuration management. Build system tooling. Infrastructure for software deployed on hundreds of millions of devices. Unglamorous, foundational work.
One thing you see clearly from inside a large software organization is the coordination overhead of headcount. It does not scale linearly. Every new engineer added to a team changes the communication surface area for everyone already there. The hidden infrastructure — the build systems, the release pipelines, the standards nobody wrote down because they were implicit — starts to crack. I watched teams of 300 engineers slow down when teams of 30 had moved faster, not because the 300 were less capable, but because the organization had outgrown the structures that made the 30 effective.
We repeated a version of this at Realm. Not at Nokia scale, but the mechanism was the same. When we went from 24 to 68, we didn't just add people — we added coordination. Every decision that used to happen in a room between four people now required a process. Every piece of work that used to move quickly now had handoffs. The product didn't slow down dramatically, but the friction was real and it accumulated.
The lesson is not that small teams are always better. It's that team growth has costs that don't show up on the hiring plan. The cost of not having someone is visible. The cost of having them — in coordination, in management overhead, in the constraints their presence puts on the team's flexibility — is invisible until it isn't.
Three tests before the next hire
What I run now, and what I'd give to a seed or Series A founder thinking about their next hire:
The absence test: What specific thing is not getting done because this person doesn't exist? Not a category of work — a specific deliverable or outcome. If you cannot name it, you're hiring to reduce anxiety, not to address a constraint.
The stage test: Is this the person for where the company is, or where you want it to be? Hiring for the next stage feels responsible but often isn't. The person you'll need to manage a team of fifteen engineers is not the person you need to build the product alongside two other engineers. Optimizing for the future role costs you the present one.
The founder-time test: Is this hire freeing up founder time for high-leverage work, or is it delegating something the founder shouldn't be delegating yet? Some things — early customer relationships, core product decisions, the hiring bar itself — should stay with founders longer than feels comfortable. Hiring away from those responsibilities too early doesn't extend your capacity. It disconnects you from the feedback that keeps your product decisions grounded.
The principle that stuck
From 68 to 24, I came away with one principle I haven't compromised since: the company has to be able to survive without the next round. Fundraising is optional, not a necessity. If the business cannot support its own headcount, you've built a structure that depends on external validation to exist.
That principle shows up in hiring more than anywhere else. Headcount is the largest controllable expense in most startups and also the hardest to reduce. You can stop buying software. You can delay a marketing campaign. You cannot easily lay off twelve people when you need to. The decision has a human cost that is disproportionate to the financial calculation — and the human cost is real, not a rounding error.
The right discipline is not to be permanently understaffed or to refuse to grow. It's to make each hire as deliberate as the first one. At the beginning of a startup, founders agonize over every hire because the consequence of a bad one is immediately obvious. As the company grows and there's capital in the bank, that deliberateness erodes. The C-round was supposed to fund growth. It felt like permission. It wasn't.
The team you have at 24 people can still be the right team at 24 people. It's worth knowing the answer to that question before you scale to 68.
This is part of a series on startup fundamentals — the operating decisions that determine whether a company survives. Related: Default Alive for Real and The Six Board Slides Every Seed Company Needs.