Run a Thumb-Stop test

The share of people who say they would stop for this in a moving feed, plus what actually registered in the glance.

Thumb-Stop asks the feed question directly: would this earn a stop, or would a thumb keep moving? It is scored as a rate rather than an average.

What it measures

The primary question offers seven answers from "Definitely not" to "Definitely", and the verdict is decided on the share of people who picked one of the top two. A rate, not a mean. The card labels this set as built from measurement principles rather than drawn from a specific body of research.

What people are asked

• "If this appeared in your feed, would you stop to watch it?" across those seven answers. This is the question the verdict rests on.
• "Once you were watching, how interesting was it?" on a 0 to 10 scale.
• A pick-any memory check on what registered: title or caption text, a face, the product, a logo or brand name, the colors, or none of these.
• An open question on what it seemed to be about, and another on what would make them stop.
• A 0 to 10 rating of how much it would stand out from everything else in their feed.

The memory check is asked from recall, after the piece is gone, so nobody sees the work twice.

When to pick it

Pick Thumb-Stop when the fight is about packaging: the cover frame, the first image, the thing that has to survive a scroll. Skip it when the piece is not feed-native, because pre-roll and page embeds do not reach people that way and the feed framing will not match. Once someone is already watching, Hook Strength and Attention and Hook are the sharper reads.

What the verdict looks like

A winner on the stop rate is named only when both groups reached the size committed at launch with balanced dropout and a gap clear of chance. Rates need more people than averages to separate, so a small test here often lands at "too close to call". The memory check and the open answers still tell you what got through. See Understanding your verdict.

What it cannot tell you

Hypothetical feed behavior overstates real stopping. Everyone in a study is paying attention in a way nobody scrolling is, so read the number as a relative comparison between versions and never as the share of a real feed that would stop.

Formats

Thumb-Stop is available on video tests and on image tests.