Cost per tested creative: production is the smaller half

Cost per tested creative equals what it cost to make plus what it cost to reach a decision about it.

Most versions online stop at the first term. That is why the answers look encouraging and the months do not.

Production here is one batch of 100 credits for a source take of up to three minutes. Export adds 20 credits per output minute.

A forty second ad is about 113 credits. A sixty second one is 120.

The second term is the media spend needed to get enough impressions on that creative to decide whether to keep it.

Add them and you have the real cost of one attempt. That is the number that belongs in any plan about finding a winner.

Write both terms side by side rather than as a total. The moment they are separate, it becomes obvious which one you can move.

Decide what signal you need. Usually that is enough impressions to tell a clear winner from a clear loser at your baseline rate.

The sample size approximation for comparing two proportions is 16 times p times one minus p, divided by the difference squared.

Here p is your own baseline rate as a decimal. The difference is the gap you want to detect.

Detecting a large gap is cheap. Detecting a small one is not, and most people are trying to detect a small one without knowing it.

Multiply the impressions you need by your cost per thousand impressions. That is the spend floor per creative.

That number, not the production cost, caps how many creatives you can honestly test this month.

Do the calculation once and keep it. It changes when your baseline rate changes, which is quarterly at most.

Most people have never computed it, which is why most testing plans are decided by whoever sounds most confident.

If your spend floor per creative is large next to production, halving production cost changes almost nothing.

It becomes decisive at the other end. You have budget to read twenty creatives and can only afford to make six.

That situation is more common than it sounds. Production usually involves somebody's diary as well as their invoice.

One filmed take of up to three minutes covers several scripts if you record them in one session. Each batches separately at 100 credits.

Rendered presenters change the arithmetic sharply. At 320 credits per minute plus 70 for a generated read, a forty second rendered ad is about 374 credits.

That is over three times the filmed cost per attempt. Use the render for the scripts nobody will say out loud, and film the volume.

Cost per tested creative is an input. Cost per winner is the output somebody actually cares about.

Multiply cost per tested creative by thirteen to twenty. That range comes from a published winner share of roughly 5 to 8 percent.

Motion's analysis of 550,000+ Meta ads is the source. No tighter number is defensible.

On the production side alone, twenty filmed attempts is about 2,260 credits. That fits inside a month of Basic at $39.99 plus one top-up.

Add the media floor and the picture changes completely. That is the point of doing both terms.

If cost per winner comes out above what a winner is worth to you, the fix is upstream of any tool.

The formula assumes creatives are independent attempts. Six variations of one idea are closer to one attempt than to six.

It assumes you will act on the reading. A test you do not act on cost the same and returned nothing.

It ignores the value of a losing creative. A loser sometimes tells you more about the audience than a winner does.

And it assumes somebody will film. The cheap production term depends on one talking-head take.

Two practical facts sit alongside those. Every export is 9:16, and nothing of yours is stamped on the file.

With all of that stated, the number still does its job. It stops a plan built on production cost alone.

Anybody quoting a cost per tested creative without a media term has calculated the cheaper half and called it the answer.

Ask for both terms next time somebody presents one. Most creative plans get considerably more honest in the same meeting.

Then price the production half here, where the fifteenth attempt costs the same 113 credits as the first.