For most of the last decade, media buying was where paid performance was won: audiences, bids, structure. That has inverted. The major platforms now infer who should see an ad primarily from the ad itself, which means the creative is doing the targeting and the media buyer's remaining levers are comparatively small.
The consequence is an arithmetic one that most teams have never run. Winning assets fatigue on a predictable clock, and only a fraction of new concepts ever become winners. If your production throughput multiplied by your hit rate is lower than your fatigue rate, the account declines — regardless of how good any individual asset is, and regardless of what the media team does. Four excellent assets a month cannot hold a position that needs twenty.
That reframes creative as a supply problem with a testing discipline attached, not as a taste problem. It changes what you resource, how you brief, and what you measure — and it means the most valuable thing a creative team can produce is not a great ad but a reliable understanding of what this audience responds to, because that is what makes every subsequent round cheaper.
What decides outcomes here
Measure the opening separately from the body
Hook rate and hold rate answer different questions and need different fixes. Blended into one completion number, a strong argument nobody watches looks identical to a weak one — and those need opposite responses.
Brief by variable, not for one great ad
Assets that differ along several dimensions at once produce results you cannot attribute to anything. Hold everything constant except the thing you are actually testing, or the round teaches you nothing transferable.
Separate concept testing from production polish
Applying full finishing to every asset, including the majority that will be killed within a fortnight, is the most common reason throughput stays low. Test rough, finish only what proves out.
Build modular, shoot once
A component library — hooks, product beats, demonstrations, testimonial fragments — turns a new concept into an edit rather than a production. This is what makes volume sustainable after an agency leaves.