Outsourcing decisions in operations and IT are made with spreadsheets. Outsourcing decisions about creative work are usually made with instinct, and the instinct is generally the same one: this is too close to the brand to hand out. It is a reasonable instinct that leads to expensive outcomes, because it treats a divisible job as indivisible.
Creative work splits cleanly along a line most organisations never draw. On one side is direction: the judgment about what the work should communicate and why, which encodes knowledge of the business, the customer and the market. On the other is production: the execution of that judgment at volume, which encodes craft. The first is genuinely hard to delegate. The second is delegated successfully every day, at scale, by industries whose entire economics depend on it.
The clearest example is one most executives never look at. Game studios are the most art-intensive software businesses that exist, and virtually all of them keep art direction internal while buying production capacity externally. Not because it is cheaper per hour, though it usually is, but because demand is spiky in a way that headcount cannot follow. A seasonal release might need several hundred finished assets on a fixed date, followed by weeks of nothing.
Three Habits Worth Copying
They write the brief before they choose the vendor. An underspecified brief is the single largest predictor of a disappointing engagement, and it is almost always the buyer failure rather than the vendor. A workable brief states the deliverables and formats, the technical constraints, reference material that demonstrates the target rather than describing it, the approval sequence, and a budget range. Withholding the budget range does not improve your negotiating position; it produces proposals aimed at the wrong problem.
They buy a paid test before they buy the programme. A small batch, paid at normal rates, reveals in two weeks what references and portfolios cannot: whether the vendor reads a brief accurately, hits dates, and handles a revision round without friction.
They keep a style document and treat it as infrastructure. When multiple contributors work on one visual system, drift is the default outcome. A short document that fixes palette, proportion rules and the specific things that must never vary costs a day to write and prevents the class of problem that customers register as sloppiness without being able to articulate.
The vendor-selection question deserves a note of its own. Portfolio quality is a weak signal, because a portfolio shows the best work a vendor has produced under unknown conditions. Stronger signals are how they describe their process, whether their pricing structure is legible, and what they say happens when something goes wrong. A studio that publishes its methodology openly is easier to evaluate than one that publishes only results; a reasonable illustration is this guide to outsourcing 3D art, which sets out the stages and the failure points rather than the highlight reel.
For organisations comparing providers, it is also worth understanding how the market segments before talking to anyone. Price bands correspond to real differences in what is included, and an overview of how outsourcing studios compare makes the trade-offs visible in advance rather than during negotiation.
The underlying point is not that creative work should be outsourced. It is that treating it as a single indivisible thing guarantees you either overpay for capacity you use intermittently, or underinvest in the direction that actually differentiates you. Draw the line, and both problems become tractable
