Short answer: hire a senior freelancer when the scope is clear and one person can hold it, hire an agency the moment the work needs more than two people, and hire in-house only when design decisions are entangled with product decisions every single week. Most early-stage founders get this wrong in the same direction, and it costs them months.

I have worked all three sides. I was in-house at Novo, a US fintech serving small businesses. I was the founding product designer at Sully.ai, an AI healthcare startup, where I was the only designer. And I have spent the last several years independent, working directly with founders across fintech, healthcare, biotech, and AI SaaS. This is the comparison I wish someone had given me before I picked any of them.

In-house: slow cycles, compounding context

At Novo, a feature took roughly two months from a product manager bringing the problem to something shipped. Discovery, design, build, release. And then it kept going, because the version that launches is never the version that stays. You iterate on it after it is live, with real usage data, for as long as the feature matters.

That rhythm is the actual argument for in-house. When I joined there were three or four designers and a head of design. By the time I left the team was past seven. What that scale buys is not throughput, it is continuity: a designer who was in the room for the last four decisions makes a better fifth one, and no external partner can replicate that no matter how good they are.

The other thing in-house gives you is proximity to everyone else. At Novo I sat in on customer interviews, worked with researchers, and went back and forth with developers as things got built rather than handing over a file and hoping. That is the part I genuinely miss. As an external partner I hand off work, and while I document it properly, it is not the same as being in the meeting when an engineer says the interaction I designed will take three weeks.

In my first year at Novo I redesigned the website and the app and shipped features on top of that, and the team gave me their MVP award, most valued person in product. I bring it up for one reason: that kind of end-to-end ownership across a whole product surface is available in-house in a way it rarely is on a contract.

The failure mode at seed stage is not the salary. It is utilisation. A senior designer with three days of real work a week gets bored and leaves. One designer owning brand, product, and marketing at once burns out. Both cost more than the hire did.

Founding designer: in-house on paper, something else in practice

I was the only designer at Sully.ai. I owned the website, the web app, and the platform, and I worked directly with the founders with nobody in between.

We had a rule: improve one percent every day. In practice that meant shipping something every single day, even if it was small. I would research competitors, go through user requests, and come back not with mockups but with proposals about what we should build and why. The job was not executing a brief. It was writing the brief, then executing it, then defending it, often in the same week.

Things broke. We were moving fast, so of course they did. But the way to read that is not recklessness. We were not pushing enormous features on instinct; we were shipping small, A/B testing what we shipped, watching what failed, and refining. Breaking something small on Tuesday and fixing it Thursday is a much cheaper way to learn than spending six weeks in Figma being certain.

If you are hiring a founding designer, screen for that specifically. Ask a candidate about a time they changed the direction of a product, not a time they executed a design beautifully. Somebody waiting for a brief will stall in this seat.

Agency: coordination is what you are buying

You are not paying an agency for better craft. Plenty of independent designers out-execute plenty of agencies. You are paying for parallel capacity and for someone other than you holding the timeline together.

That matters when brand, website, motion, and copy all have to land on one fundraise announcement date. It matters much less when you need one clearly-scoped thing built well, because you are also buying an account layer between you and whoever is actually designing. Every question routes through a translation step, and at early stage, where the brief moves weekly, that translation step is the expensive part.

Independent: fastest start, hard ceiling

The honest pitch for hiring someone like me is speed and directness, not price. You talk to the person doing the work. There is no discovery phase that exists to justify a retainer.

From a signed contract I usually start inside a week. First real work ships in two to four weeks depending on how fast feedback comes back, and longer than a month if the scope is genuinely large. Compare that to a two-month product cycle in a staffed in-house team, or to the weeks of recruiting before an in-house designer exists at all.

The real tradeoff is capacity, and anyone who tells you otherwise is selling. One person has a ceiling. If your scope triples mid-project, there is no bench behind me. An agency absorbs that. I cannot.

When I tell founders to hire someone else

My rule is simple: if the scope genuinely needs more than two people, you want an agency, not me.

I am a sole designer. I collaborate with one or two people when a project calls for it. But if you need a 3D artist, an illustrator, a motion designer, and a developer all pulling in the same direction on the same deadline, an agency has those people under one roof and I would be assembling a temporary team for you. That is a worse version of what a studio already does well, and you would be paying me to project-manage rather than to design.

The same applies if you need a committee to sign off, or if you need steady output every week indefinitely. A subscription service or an in-house hire is a better instrument for continuous production than a project with me.

The decision, compressed

One clearly-defined thing, built well, with direct access to you for a few weeks: hire an independent senior. More than two disciplines on one deadline: hire an agency and pay for the coordination. Design decisions entangled with product decisions every week, permanently: hire in-house and accept the lead time.

Scale your design capacity to what actually needs shipping this quarter, not to what looks legitimate on an org chart.