Fullslate

Most teams commission the number of videos they can get produced. If that number is four a month and the channel could run forty, the limit sits in production rather than in media spend.

Generating a large batch of cuts from one brief is now inexpensive, but checking them still takes a person. A team going from four videos a month to forty has to look at ten times as much material before anything goes out. We build on the assumption that review is the binding constraint, so most of the engineering goes into keeping off-spec material away from the reviewer.

• A production line that runs from brief to finished cut, using your brand kit, your fonts, and cuts you have already approved.
• A conformance check that runs before any person sees the batch. It tests what a machine can test: aspect ratio, duration, captions present, on-screen legal text, audio loudness, safe zones. Anything that fails is held back.
• An approval step. Nothing is published until someone on your side approves it.
• Several hook variants for each concept, so you can test openings at a set budget.

The signal

First-pass acceptance rate: the share of delivered assets your reviewer approves without asking for changes. Where you run paid media, we also track three-second retention across hook variants at a set budget.

Where it comes from

Your reviewer’s approve, edit and reject decisions, recorded as they work. For retention, the ad platform’s own delivery report.

How fast

Acceptance rate gives a readable weekly figure once you are past about thirty assets. A hook test settles in around seventy-two hours.

The trap

We have no reliable machine test for whether a video is any good. Acceptance rate tells you how often your reviewer agreed, which depends on that reviewer and will drift if they start approving without looking. We record how many assets were opened alongside how many were approved, so you can see whether that is happening.

A quality-control report on a real batch of forty: how many were held back, which check each one failed, and the frames that failed it. Alongside that, first-pass acceptance rate week by week.