How audience testing works

Show two versions to two fresh groups of people, one version per viewer, and get back a plain-language read on which one landed.

Audience testing shows two versions to two fresh groups of people and tells you, in plain language, which one landed. It is built for the call you cannot make from the edit bay alone: which opening holds attention, which version of the line reads clearer, which one people actually finish.

One viewer, one version

The test is monadic, which means every participant sees exactly one version, never both. Someone who watches version A is never shown version B, so nobody is weighing the two against each other in their head. Each group meets its version cold, the way a real viewer meets your work for the first time. That is what keeps the comparison honest: you are measuring two first impressions, not one person's preference between two options.

The six-step wizard

Setting up a test walks through six steps:

1. Pick your versions. Upload or choose the two you want to weigh against each other.
2. Write your questions. A short set: one primary question the test is really about, plus any follow-ups.
3. Choose your audience. A general-population group, or narrow by age and gender. See Choosing an audience.
4. Set the size. How many people watch each version. More people means a firmer answer.
5. Review the quote. You see the credit cost before anything runs. See Credits and packs.
6. Launch. The test fills, and you watch progress as responses land.

Screenshot: testing-wizard-1
The six-step setup: pick the versions, write the questions, choose who watches, set the size, review the cost, launch.

What a participant sees

Participants open a plain, confidential page: no studio branding, no logo, no color from your kit. Branding is a hint, and a hint would bias the answer, so the page is deliberately neutral grey. They watch their version once (the player does not let them skip ahead or rewatch), answer your questions, and that is the whole session. A single exposure keeps every reaction a genuine first look.

You get a report, not a dashboard

When the test finishes you get a written report that reads like a colleague talking, not a wall of charts. It opens with the plain verdict, carries the audience's own words, and puts the numbers last for the people who want them. It tells you what happened and what it means, including when the honest answer is that there is no clear winner. See Understanding your verdict.

When to reach for it

• Two openings and no agreement in the room about which one holds.
• A claim or a line you suspect plays differently than you think.
• A finished video you want to pressure-test before it goes out wide.