NextMars / Blog / Outsourcing

How to choose a game art outsourcing studio

Six tests to run before you sign with a game art studio, the brief that exposes a bad one in 15 minutes, and the red flags that predict a stalled pipeline.

Start a project →

17 November 2026 · 7 min read

Most bad outsourcing stories do not start with bad art. They start with a pipeline that quietly stalls in week three: an artist who has stopped asking questions, a backlog of unapproved assets, a style that has drifted two shades lighter than the approved key art, and nobody senior to notice any of it.

That failure is almost always predictable from the first two weeks of contact. Choosing a game art studio is not a portfolio contest — it is a test of whether that studio can hold a scope under deadline pressure, and the evidence sits in how they answer a small brief, not in the showreel.

Here is what to look at, in the order that actually predicts the outcome.

The five ways an art pipeline fails

Worth naming up front, because every test below exists to rule one of these out.

  1. The design was approved once, badly. One drawing was signed off, modelling started, and the thing that comes back cannot be built because nobody knows what the back looks like.
  2. Nobody owns the style. No art director on the vendor side, so each artist interprets the reference differently and the drift is invisible until you line the assets up.
  3. Feedback has no shape. Notes arrive scattered across calls, chats and forwarded screenshots. The studio bills the hours and delivers something that satisfies none of it.
  4. The engine was never in the room. Everything looks right in the DCC tool and breaks on import — scale, pivots, colour space, material names.
  5. The senior people are on the sales call, not the project. The portfolio you liked was made by someone who will not touch your files.

None of these are visible in a portfolio. All of them are visible in how a studio responds to a real brief.

Six tests before you sign

1. Portfolio depth, not portfolio breadth

Every studio shows twenty styles. The question is whether they have done your style more than once, and whether they can show you the process rather than only the result.

Ask for a project in your genre and, ideally, one with the same technical constraints: same engine, same platform, same art direction era. A studio that is excellent at painterly fantasy is not automatically excellent at flat vector mobile UI, and the flat vector UI quote will be a learning project at your expense.

2. Is there a named art director?

Ask directly: who is the art director on this project, and can I see their work. If the answer is a role rather than a person, that is your answer.

The art director is the person who keeps forty assets looking like one game. Without one, consistency becomes your job, and you will be doing it at midnight after your own day job.

3. Who answers you, and what do they ask?

Speed of reply matters less than the content of it. A good first reply contains questions — engine, resolution, style reference, deadline, whether files must be animation-ready. A poor one contains a rate range and an invitation to a call.

Calls are not bad. Calls that replace written scope are.

4. Revision policy, in writing

Two rounds per asset, one consolidated note per round, X days to respond. If this is not written down, it does not exist, and the disagreement will happen at the worst moment: when the asset is 90% done and everyone is tired.

5. The technical handover

Ask what exactly you receive: file formats, layer structure, naming convention, colour space, pivot and scale conventions, whether the source files come with it. For 2D animation, whether the Spine project is included or only exported atlases.

A studio that has integrated art into an engine before will answer this without needing to think. One that has not will answer with "PNG".

6. IP, NDA, and work-for-hire

Before any file moves: who owns the art, what happens on termination, whether the studio may show the work publicly, and whether they use subcontractors. If subcontractors are used, that is fine — but you should know, because it affects both confidentiality and the answer to question 2.

The 15-minute test

This is the single most predictive thing you can do, and it costs nothing.

Take a real asset from your backlog — one character, one environment, one UI screen. Write a four-line brief: what it is, the style reference, the resolution and engine target, the deadline. Send it to three studios and ask for a written scope, a fixed price and a date.

Then read the replies, not the prices.

What came backWhat it tells you
Scope, price, date, plus two or three clarifying questionsThey have done this before. Proceed.
Price only, no questionsEither the brief was complete, or they are guessing
"Let's jump on a call to discuss" with nothing in writingScope will be negotiated verbally forever
Quote far below the others, no assumptions listedThe excluded work will be billed later
Quote with an assumptions list you disagree withThe best possible outcome: it is visible before you sign

That last row is worth dwelling on. A studio that states its assumptions — "price assumes one approved direction, two revision rounds, no engine integration" — has just shown you its entire risk model. You can argue with it now, cheaply, instead of in week six.

A portfolio read as a set, not as individual pieces: one sheet pulled out and examined closely
A portfolio read as a set, not as individual pieces: one sheet pulled out and examined closely

Red flags that predict a stall

  • An instant quote with no questions about your engine, resolution or references.
  • "Unlimited revisions." This is not generosity; it means revisions are priced into the base rate, and the studio will eventually protect its margin by slowing down.
  • A portfolio with no process shots. Finished pieces only, no line work, no rejected directions, no WIP. It usually means the studio did not do the exploration, or is showing work it did not make.
  • No named contact who is not the salesperson.
  • A price 40% below every other quote with no explanatory assumption.
  • Reluctance to do a small paid trial asset.

Questions worth asking, verbatim

Print these; the answers are more informative than any case study.

  1. Who is the art director, and can I see their personal work?
  2. What happens if an asset comes back and the style has drifted?
  3. How many revision rounds, and how is a round defined?
  4. Who is on my project day to day, and for how many hours a week?
  5. What exactly is in the delivery package, file by file?
  6. Do you use subcontractors, and are they under the same NDA?
  7. Can we start with one paid asset before committing to the set?
  8. What has gone wrong on a project like mine before, and what did you change?

Question eight is the one that separates a studio from a vendor. Anyone can list strengths. An account of a real failure, and the process change that followed, is a much better predictor of how they will behave when your project hits trouble.

Start with one paid asset

Never begin with the full set. Commission one asset — the hardest one in the batch if you can, a hero character or the screen that defines the UI language. Pay the normal rate; a free trial is a different product made by different people with different incentives.

One paid asset tells you: how they interpret a brief, how they take notes, what the files look like, whether the style holds, and how the whole thing feels to manage. It costs a fraction of the set and de-risks all of it.

If the trial asset is wrong in a way that a second round fixes, that is a good sign. If it is wrong and the second round does not fix it, stop, and you have lost one asset rather than forty.

The contract, in eight lines

You do not need a forty-page agreement for a small commission, but these should be in writing:

  1. Deliverables, file by file, with resolution and format.
  2. Number of revision rounds and the definition of a round.
  3. Response times on both sides — theirs for delivery, yours for feedback.
  4. Payment schedule tied to milestones, not to a calendar.
  5. IP assignment on final payment, with source files included.
  6. NDA covering unreleased material, extending to subcontractors.
  7. Publicity: whether they may show the work, and when.
  8. What happens if the project stops mid-production.

The eighth line is the one people skip, and the one that causes the ugliest disputes: who owns the half-finished work, and what is owed for it.

How we work

We are a small studio, so this is easy to state plainly: one named art director on every project, a written scope with assumptions before any work starts, two revision rounds per stage, and a paid single-asset pilot when a client wants one.

If you want to run the test above on us, send a four-line brief. You will get a scope, a price and a date within one working day. If you would rather understand how the number gets built before you ask for one, read how much game art outsourcing costs — it explains why no honest public price list exists, and what actually moves a quote.

Have a brief in mind?

Send it and you get a written scope, a fixed price and a delivery date within 48 hours — no discovery call required.

Start a project →