How much creative does an account actually need?
The honest answer is not a number someone else published. It is whatever keeps a validated replacement ready before the current winner decays.
Work backwards from decay
Start with the observed half-life of a concept in your account — the cumulative impressions at which conversion rate drops meaningfully. Divide monthly impressions by that figure and you know how many concepts the scale block consumes per month. That is your replacement rate, and it is specific to your market, budget and vertical.
Then add the failure rate
Most tested concepts do not graduate. If roughly one in four passes, and you consume three concepts a month, the test block needs twelve entering per month to keep the bench stocked. Production volume is a derived number, not an ambition — and once derived, it is defensible in a budget conversation.
Volume without a graduation rule is expenditure. Volume with one is supply.
Variations are not concepts
Ten aspect ratios of one video is one concept. Counting variations as volume is how teams convince themselves they are testing while running the same argument in different frames. Track concepts and variations separately: concepts open new audience pools, variations extend the life of a pool you already have.
Where volume comes from cheaply
- Modular production. Shoot proof, demonstration and endings separately, then assemble. One shoot yields a batch, not an asset.
- Hook variation. The first three seconds carry most of the decay. Re-cutting openings on a proven body is the cheapest volume available.
- UGC pipelines. A repeatable creator process produces native volume at a fraction of studio cost, provided the brief is specific.
- Localisation. A validated concept re-voiced for four markets is four assets from one idea — as long as it is rewritten, not subtitled.
The ceiling
There is a point where more creative stops helping: when the test block cannot give each new concept enough budget to reach a decision. Past that, additional production makes results less readable, not more. If you are there, the constraint is budget structure, not the creative team — and producing more will simply cost more to learn less.
Takeaways
- Derive volume from observed half-life and graduation rate, not from published benchmarks.
- Count concepts and variations separately — only concepts open new pools.
- Stop increasing volume when the test block can no longer fund a decision per concept.