Choosing a test type

The first decision in the wizard, organized by the question you are actually asking rather than by the name of a template.

Picking a test type is the first real decision in the wizard, and the honest way to make it is to name the argument you are having. The question sets are grouped by the stage of the funnel they read, so start with the stage.

Attention: does it stop the scroll and earn the watch?

Reach for these when the opening is what changed: which first seconds keep people watching, how the pacing feels, whether the piece would earn a stop in a fast feed. Skip them when both versions share the same opening, because there is nothing there to separate.

Comprehension: did the message land?

For when clarity is the variable: which version gets the point across, which explanation makes a new or complicated product make sense, which one makes the next step unmistakable. Skip these when the goal is a feeling rather than a message, and reach for an emotion set instead.

Emotion: how does it make people feel?

Brand films and response ads want different things. One set reads the breadth of warm feeling a long-game film creates; another reads raw intensity and how fast the brand registers. There is also plain enjoyment, and whether people would pass it on. Honest limit: self-report under-captures feeling, and a well-liked piece is not automatically the persuasive one.

Brand: does it build the brand?

Whether people remember your brand afterwards, whether they can tell whose piece this is and how quickly, and whether it reads as credible. Skip these when the brand appears identically in both versions, because then it is not the variable. To measure one finished piece against people who saw nothing at all, see Brand lift.

Intent: does it move people toward action?

Buying inclination, likelihood to recommend, how an offer framing lands. Read the result as which version wins, never as the share of people who will actually buy: stated intent runs ahead of behavior, and asking the question lifts it further.

Fit: right audience, right feel, better than the rest?

Whether it feels made for someone like the viewer, whether it reads as native content or as an ad, whether the talent lands, and how it ranks against the category people already carry in their heads.

Not sure? Start with Creative Comparison

It is the recommended default and the fastest broad read: an overall rating, an engagement question, one open question, nothing else. Run it when the argument in the room is "which one is better" with no sharper diagnosis attached, then move to a specific stage once you know which part of the work you disagree about.

What exists for each format

Video has a set for all six stages. Audio has all six too, minus the questions that assume a picture. Image tests cover five: there is no comprehension set for a still frame, because "did you understand it" is a weaker question of one image than of something that unfolds. If understanding is the question about an image, test it as video or audio instead.

Two labels on every card, and why to read them

Each card says whether the question set is research-backed or built from measurement principles. That is not a quality grade (several of the most useful sets are the second kind), but you should know which one you are holding. Each card also carries a "skip it if" line, and most carry a caveat naming the limit of what the answer can mean. Read both before you spend anyone's time.