Creative direction is a decision function, not a taste function
The job is not having the best ideas. It is holding a standard across work made by many people under pressure.
In short
Creative direction is the function that holds a consistent standard across work produced by many people over time. It is often mistaken for a taste function, as though the job were having better ideas than everyone else. In practice it is a decision function: deciding what the work is trying to do, what good looks like for this specific thing, what is finished, and what does not ship.
You can identify an organization with no real creative direction without seeing an org chart. The work is uneven in a particular way: individual pieces are fine, the body of work does not cohere, quality tracks whoever happened to make each thing, and nobody can explain why one piece was approved and a similar one was not.
Quality decays by default
This is the part that surprises people. Standards do not hold on their own, and they do not decay through anyone's carelessness. They decay through a series of individually reasonable decisions.
A deadline moves, so something ships slightly below the bar. That piece becomes the reference for the next piece, because the most recent work is always the working definition of the standard. The next compromise is measured from there. Nobody made a bad decision at any point, and eighteen months later the work is visibly worse than it was, with no single moment anyone can point to.
Creative direction is the counterweight to that drift. Not through veto power, though that has to exist, but by keeping an explicit reference for what good means that is not simply the last thing produced.
The most recent work always becomes the working definition of the standard. That is why standards decay without anyone deciding to lower them.
The four decisions
- What is this for
- The most commonly skipped one. Not the deliverable, the job. Work briefed as a deliverable produces something that looks correct and achieves nothing, and it is unreviewable because there is no stated intent to review it against.
- What does good look like here
- Good is contextual. The bar for a paid social variant is not the bar for a brand film, and pretending otherwise wastes effort in one direction and embarrasses you in the other. Setting this before work starts is what makes later feedback feel fair rather than arbitrary.
- Is it finished
- Distinct from whether it is good. Plenty of good work ships unfinished because someone ran out of time and nobody was watching the difference. Finish is where craft lives and it is the first thing sacrificed under pressure.
- Does it ship
- Somebody has to be able to say no on quality alone and have that hold. If no is always negotiable against a date, the standard is decorative.
Feedback that improves work
Most creative feedback makes work worse, and it does so through a specific mechanism: it substitutes a solution for a diagnosis. When a director says make the logo bigger, they have given an instruction that can be followed exactly and still not fix the underlying problem, which might have been that the hierarchy is unclear or the eye enters the composition in the wrong place.
- Say what is not working before saying what to change. The maker usually has a better solution once they understand the problem.
- Separate reactions from decisions. Not liking something is information, not an instruction, and conflating the two is how work gets flattened toward whatever offends nobody.
- Give feedback against the stated intent, not against what you would have made. These are different, and the second one is not your job.
- Be specific about what is working. Not encouragement, information. If the maker does not know what to protect, the next round will lose it.
- Deliver it once, in one pass, from one person. Serial feedback from multiple directions produces work that is an average, and averages are the thing creative direction exists to prevent.
Direction at volume
The shape of the job has changed. When production was expensive, direction happened per asset. Now that a team can generate hundreds of variants in an afternoon, per-asset review is not possible, and the function moves upstream: direction is exercised through the system, the references, the templates and the constraints, rather than through individual approvals.
That makes the reference material more important than it has ever been, because it is now the thing doing the directing when nobody is in the room. It also makes the review sample the critical control. You cannot look at everything, so what you look at, how often, and what happens when it falls short is the entire mechanism.
The uncomfortable part
Creative direction only works if someone is accountable for the result and has the authority to refuse work. Where that authority is nominal, where any no can be escalated around, the function is theater. The organization has a job title and no standard, which is more expensive than having neither, because it produces the reassuring feeling of quality control without the substance.