These used to be two different businesses. A creative agency made the assets; a media agency bought the placements. On modern app platforms that split has stopped making sense, and choosing between them as if they were alternatives is usually how teams end up with creative nobody can evaluate.
The more useful question is not which type to hire. It is who owns the read on whether the creative worked.
Page Contents
The three models you will actually be offered
| Model | What you get | What you still own | Fits when |
|---|---|---|---|
| Advisory | Strategy, direction, review. No production, no buying. | All execution | You have a capable team and want a second opinion |
| Creative only | Concepts and production. Your team runs the ads. | Media buying and test structure | You have a strong media buyer and a creative bottleneck |
| Creative plus media | Concepts, production, and running the account | The product, and the final call | You want one team accountable for the number |
The prices look similar. The engagements are not.
Why the split stopped working
Targeting used to be the job. You built audiences by hand, stacked interests, layered lookalikes, and a good media buyer beat a bad one on that skill alone.
That work is largely automated now. You give the platform a budget, a conversion event and a creative, and it finds the people. The creative has become the thing being optimised.
Which means media buying decisions are now downstream of creative decisions. Budget should move because a creative earned it. Ad sets should be structured so individual creatives can be read. If those two jobs sit with two parties who do not talk, the account is being run by nobody in particular.
The thing that actually matters: who reads the results
This is the part most buyers do not think to ask about, and it decides whether the engagement works.
Whoever makes the creative has to see how it performed. Not a monthly summary. The actual per-creative read: which concepts survived, which died, what the survivors had in common.
Without that, a creative team is guessing. They ship a batch, hear “it didn’t work,” and make another batch on the same assumptions. That is not iteration, it is repetition with a delay.
So even on a creative-only engagement, where your team runs the ad account, the creative side should still be reviewing performance and feeding recommendations back: how campaigns are structured, how tests are set up, what to test next, and where the creative strategy should go. A creative partner who never looks at the numbers is flying blind, and so are you.
When evaluating anyone, ask: “who looks at the results, and how often?” If the answer is “you do, and you tell them,” that is production, not a creative programme.
When creative-only is genuinely right
It is the right call when you already have a media buyer who structures creative tests well, and your actual bottleneck is the supply of concepts. Adding an agency to the buying then pays for a seat that is already filled.
But even then, make the creative partner accountable for understanding what is being tested. They should know the test structure, see the results, and be able to tell you whether a concept failed or was never given a fair read. A starved test and a bad idea look identical in a report, and only someone looking at the setup can tell them apart.
When one team for both is right
When nobody currently owns the link between what gets made and what gets spent. That is the more common situation, and it is where splitting the work costs the most: the creative team blames the buying, the buying team blames the creative, and the account drifts.
Questions worth asking either way
- What is one test, and what is the unit? Concepts or ads. Agencies mean very different things by “test”, and the difference shows up in your invoice.
- Who reviews performance, and how often? If the creative side never sees results, they cannot improve.
- Who structures the tests? In creative-only engagements this is usually you, and it decides whether the creative can be judged at all.
- What happens when a concept fails? The answer tells you whether they iterate on the idea or just produce more variants.
This post is part of our guide to how to choose a mobile UA agency.
Frequently asked questions
What is the difference between a creative agency and a media buying agency? Historically creative agencies produced the assets and media agencies bought the placements. Platform targeting is now largely automated, so the creative is what actually gets optimised and media decisions follow creative results. The distinction matters less than who owns the read on performance.
Should I hire a creative agency or a media buying agency? Hire creative-only if you already have a media buyer who structures creative tests well and your bottleneck is concepts. Hire one team for both if nobody currently connects what gets made to what gets spent, which is the more common situation.
Can a creative agency work without running my ads? Yes, and it is a common setup. The condition is that they still see performance data and review how tests are structured. A creative partner who never sees results cannot tell a bad concept from a badly-run test.
Who should structure the creative tests? Whoever can guarantee each creative gets a readable result: one creative per ad set, enough budget per test, consistent naming. If that is not agreed explicitly, it usually happens by accident and badly.
