# When does a startup need a designer? | Other Land

URL: https://otherland.studio/blog/when-a-startup-needs-a-designer

> In-house, embedded, fractional or freelancer: how to tell which design model fits your startup, with real pricing thresholds and where each one breaks.

Know a team that needs this? Send it their way

## When does a startup actually need a designer, and which model fits

In-house, embedded, fractional, freelancer: what each one actually costs, and when each one breaks.
Veronika Bogush, Operational Director October 22, 2026 · 12 min read

### Key takeaways

- Funding stage predicts little: three teams with almost the same story needed three different design models.
- Three questions decide it: a stream of decisions or a task with an end, whether anyone inside can evaluate design, and whether the direction is locked and funded for a year.
- Embedded fits a continuous stream of decisions, a project fits work with a finish line, and fractional fits a product in decent shape with a team that can carry it.
- With an embedded designer from Other Land, we pick the designer for your problem, a design director and a product manager stay on the account, a designer who does not fit is replaced without a hiring round, and the contract ends or extends when you decide.

Most startups need design help earlier than they hire for it, and most of them need a different model than the one they’re considering. The short version: if design decisions come up every week and shape what gets built, you need a designer inside the team. If the product is already working and you mostly need a second opinion plus occasional execution, you still need a designer, just not a permanent one. Those two situations look almost identical from the inside, and they cost about four times different.

I run operations and partnerships at Other Land, a product design studio. Five years of my job has been the conversation that happens before any design work starts: a founder describes where they are, and we figure out together what form of collaboration actually fits. Sometimes the honest answer is that we’re the wrong fit entirely. This article is that conversation, written out.

### Funding stage tells you almost nothing

Every article on this topic sorts the answer by round. Seed means freelancer, Series A means your first hire, after PMF means a team. It’s a clean structure and it’s mostly wrong.

Here are three teams that came to us within a few months of each other.

#### A US product

Two founders and a small team, an investor on board, and real customers. The engineers built the product with AI tooling and can assemble interfaces themselves.

#### A team in the UAE

Selling six AI modules to government departments through private contracts. Product works, partnerships signed, everything built by their developers.

#### A French platform

Pivoted from selling its own courses to letting creators upload theirs. Backend solid, new audience incoming.

Same description almost word for word. Product assembled by engineers using AI, first customers or partners already there, and a growing sense that the thing looks like something a couple of people put together last month rather than a company you’d wire money to.

They got three different answers. One took a full-time embedded designer for ten months. One is weighing embedded against a first in-house hire. One needed a bounded two to three month project and then to be left alone. If stage predicted the model, that couldn’t happen.

### The three questions that actually decide

#### Is this a stream of design decisions, or a task with an end?

This is the first fork and the cleanest one. A stream means design questions come up in planning, not after it: what this feature is for, how it should behave, what makes it successful. A task with an end means you know what needs to exist, roughly, and you need someone to build it well.

#### Can anyone inside your company evaluate design?

Not “do you have taste”. Can someone look at a portfolio and tell whether that person is good, and then look at the work six weeks in and tell whether it’s going well. If nobody on the executive team can do this, hiring in-house means choosing blind and then managing blind. That’s the single most expensive mistake available to you at this stage, and it isn’t fixed by budget.

#### Is your direction locked, and is there money for a year?

A full-time hire is a bet that the next twelve months look roughly like the last three. That bet is fine when the product is scaling on a settled direction and revenue is predictable. It’s a bad bet while you’re still finding the shape of the thing.

Everything else, the round you raised, the headcount, the industry, mostly falls out of these three.

### The five options, in plain terms

Before the table, worth defining these, because the words get used loosely and two of them are routinely confused with each other.

#### In-house hire

A designer on your payroll. You recruit them, you evaluate them, you manage them, and you carry the cost whether or not there’s a full week of design work. Roughly three months from opening the role to someone being productive, if the search goes well.

#### Freelancer

An individual you contract directly, usually per project or per month. Cheapest real option and often the right one. What you don’t get is anyone checking their work, so the quality of the outcome is the quality of that one person, and you’re back to needing someone inside who can tell.

#### Embedded

A designer from a studio who works inside your team as if in-house: your tools, your rituals, your roadmap, your Slack. The difference from outstaffing is what stands behind that person. At our end there’s a design director and a product manager attached to the account for its whole duration, validating decisions and stepping in when something needs escalating. You’re buying one designer’s capacity and a studio’s judgment.

#### Fractional

A fixed monthly allocation of senior design time rather than a person. Ours starts at 40 hours a month. The team stays in your context (we join a weekly meeting), executes a defined amount of work, and reviews what you built without us. It is not a discounted embedded designer, which is the confusion I correct most often on calls. It’s for a different situation entirely, and I’ll come back to that below.

#### Project engagement

A scoped piece of work with a start, an end, and a handover. A rebuild, a new flow set, a design system. When it’s done, we leave.

There’s a sixth thing on the market, the design subscription, where you buy a queue and submit requests. It’s built for volume across many small tasks and it’s genuinely good at that. It’s poor at depth in one product, because nobody in that queue is carrying your context between tickets. If you’re comparing it to fractional, that’s the difference.

### The models, and where each one fits

Situation
Model
What decides it

Design decisions come up in planning and shape what gets built. Volume of iteration is high
Embedded full-time
You need judgment in the room, not delivery after the fact

You need to look mature within six months, nobody inside can evaluate design, and what happens after is genuinely unknown
Embedded instead of a first hire
Hiring risk plus an undefined horizon

Design is needed across the product and everything around it: investor decks, conference materials, ad creative, site
Embedded full-time
One senior generalist covers what would otherwise be three part-time hires

Product is scaling on a settled direction, budget covers a year, and someone with design expertise is available to help you hire
In-house hire
Permanence matters more than flexibility

A defined rebuild with a clear end, after which the team ships on its own
Project engagement, design system handed over
The work has a finish line

The stream of decisions has stopped, but consistency and outside perspective are still needed
Fractional
Design is needed regularly, not daily

Small volume, the team builds most of it, they need feedback on their own work and targeted fixes
Fractional
No case for full-time, but a permanent low-level need

Long horizon: work always appears, with a large project every couple of years
Fractional plus project bursts
Continuity between the bursts

Budget below the threshold
Freelancer
Our cost structure doesn’t compress that far

#### Embedded: when design decisions are continuous

We worked with Persona, a US product, on an embedded model for about ten months. The designer was full-time on the team: in every planning ritual, in direct conversation with everyone, aware of the strategic roadmap, part of the discussion about what features meant and how success would be measured. Roughly 80% of her time went to the product, both improving what existed and defining what was new. The other 20% went to everything that comes with scaling: the investor deck, materials for a conference, creative for a paid campaign.

She worked in Figma and also used Cursor to assemble new features. That combination is worth naming, because a lot of founders assume that a designer who uses AI is doing less. The volume of iteration was the reason embedded made sense, not the tooling.

Then Persona changed direction. The focus moved from selling the product to selling the data the product generates, which made them their own primary user. End-user experience stopped being the thing that determined the company’s success. Within a month they told us the contract was ending, and they were right to.

That’s the honest failure mode of full-time embedded: it’s sized for a stream of decisions, and streams can stop.

The second embedded case is a live conversation rather than a finished story. That UAE team with six modules is choosing right now between a first in-house hire and working with us. Their situation has three properties that push toward embedded. They know they need design for the next six months and genuinely don’t know about after: if the six modules get rebuilt well, that might be enough for a while. Nobody in the company can evaluate a designer. And a bad hire, made blind, costs them the six months they’re trying to save.

#### The 10 to 30% conversation

Embedded is often the more expensive option on paper. Depending on the market, it can land anywhere from below the cost of a local mid-to-senior in-house designer to about 10 to 30% above it.

I’d rather write that number down than have it come up on a call, because the interesting part is what the difference buys.

#### You don’t manage the designer

There’s a design director and a product manager on the account for its whole duration, validating decisions, weighing in on the harder product questions, and stepping in if something goes sideways between the teams. For a team with no design expertise inside, that’s the entire point: you’re not qualified to give design feedback, and with this model you don’t have to be.

#### You get a department rather than a person

If the designer turns out to be wrong for the product, they’re replaced, and you don’t run a hiring process to do it.

#### You can stop

Contracts end or extend on a decision you make later rather than one you make now.

If your direction is locked, you have a year of runway, and you have someone with design expertise to help you hire well, none of that is worth paying for and you should hire.

The premium is insurance against uncertainty. When there’s no uncertainty, don’t buy insurance.

#### Project: when the work has an end

The French platform pivoting to creators needed the product rebuilt on the UI side, new flows for a new audience, and the whole thing made coherent. They didn’t need a permanent designer because the volume wasn’t there and their team assembles most things with AI already. We scoped it at two to three months with a design system handed over at the end, so they could keep building without us.

Scoping this as embedded would have been more expensive and worse. A stream of decisions requires someone in the room. A rebuild with a known endpoint doesn’t.

#### Fractional: not a cheaper embedded

This is the most misread model on the list, so I’ll be blunt about it. Fractional is not a discount version of having a designer. It’s a different situation.

#### The Key Academy, a German education product

Runs on 40 hours a month with us. Their team builds a lot themselves and has no design expertise inside. Every week we join the team meeting to stay current on the product and the business, we execute the design work most likely to move their numbers (conversion is their main challenge), and we give feedback on what they’ve built without us.

#### Giftmall, a gift card marketplace

Has been with us for seven years. Every couple of years there’s a large project: the mobile app, a full site redesign. Between those, the same 40 hours a month. Their task volume doesn’t justify a hire or an embedded designer, but work always appears, and having someone permanently in context means it gets picked up and everything stays consistent.

#### Persona, after the pivot

Moved from embedded to fractional rather than leaving. The stream of product decisions was gone. The need to look grown-up and consistent wasn’t.

What these three have in common isn’t budget. It’s that the product is in decent shape and the team can genuinely carry it forward.

Fractional works when someone competent needs a designer nearby. It doesn’t work when a startup needs to seriously improve a product; 40 hours a month won’t hold that.

### The numbers, and where we stop being the answer

Our pricing, plainly:

- Embedded full-time starts at $2,000 per sprint week, invoiced every four weeks, four week minimum. Over a year that’s around $104,000 and up.

- Fractional starts at $2,000 a month for 40 hours.

- Between those, the realistic shape for a team that needs actual output but can’t fund full-time is 80 hours a month, or full-time for a focused four to six months to fix everything at once. Forty hours won’t produce enough in a month to feel like progress.

Which gives a threshold worth stating out loud.

If you have less than roughly $50,000 a year for design and you need continuous product improvement, we’re the wrong studio for you.

That situation is common right now. Early-stage team, product assembled with AI, and they can see it looks generated. If it’s a complex SaaS product it also looks confusing. Their decks aren’t consistent either. All of that is real and all of it is fixable, but our cost structure doesn’t compress: senior design engineers are expensive, and the design director and PM are on every account by construction.

For those teams, a freelancer is the right answer. You lose the validation layer, which is a real loss. You still end up with a substantially better product than you have now, for much less money. Better than nothing is an underrated position, and I’d rather say this than take a project that leaves a team with no runway and a half-finished redesign.

### Why this choice matters more than you’d think

You will probably live with whatever you pick for years.

Almost nobody moves from us to an in-house hire in the way the internet implies they should. It happens, but rarely, and usually after a large change in the company rather than as a natural graduation. Most engagements just continue. Our longest unbroken one ran four years, with Cricut, and Giftmall is now in its seventh year with us. The reason is unglamorous: a designer who is deep in the context of your product is expensive to replace, and onboarding a new one costs months, so as long as it’s working there’s no reason to change it.

Which means the decision you’re making isn’t “what do we do this quarter”. It’s closer to choosing a structure you’ll still be inside in two years, adjusted at the edges. Worth the extra week of thinking.

### When the right model and the affordable model disagree

I want to be honest about something we don’t usually put in writing.

Sometimes the model that fits isn’t the model a team can pay for. Embedded is clearly right, the budget covers fractional. We don’t refuse those, we adapt the roadmap: fewer things in parallel, tighter prioritisation, more of the execution left to the client’s team.

But you should know what you’re trading:

- Slower cycles.

- Less depth, because someone who is in your product two days a week knows it less well than someone in it five.

- No same-day response when something urgent lands.

Fractional-instead-of-embedded is a real option and often the right call, as long as nobody pretends it’s the same thing at a discount.

If you’re sitting with this decision now, the fastest way through it is to answer the three questions at the top honestly. Stream or task. Can anyone here evaluate design. Is the direction locked and funded. In my experience the answer falls out of those three faster than out of any comparison of agencies.

If you want to talk it through, that conversation is what I do. Book a call.

### Other Land's design services

Product DesignProduct design agency for SaaS and AI productsDesign EngineeringDesign engineering agency for teams that ship in codeGrowth DesignGrowth design agency for increasing conversion, activation and retentionDesign SystemsDesign systems for SaaS, AI products and mobile appsUX AuditUX audit services for SaaS and AI productsMVP DesignMVP design and build for early-stage SaaS and mobile startupsWebsite DesignWebsite design and development for product companiesBrand DesignBrand design and brand refresh for product companies

Blog

### Other Land articles on AI product design

AI product design
October 12, 2026

#### Why AI-generated products look generic, and what designers do differently

by Artem Ivanov, CEO
AI product design
May 28, 2026

#### How AI Quietly Changed Modern UX Patterns

by Other Land team
AI product design
December 8, 2023

#### AI Workflows and Modern Application Design Patterns

by Vadym Oliinyk, Product Designer