Curiosity is the most defensible idea in thumbnail advice. People do want to close a gap in what they know, and that has been studied for decades. What the advice usually gets wrong is how much to hold back. Real click data shows that the vaguest teases lose, and so do packages that give everything away. This article explains what the research supports, how to apply it to a thumbnail, and how to test it on your own channel.
Key takeaways
- Curiosity is the pull to close a gap in what you know. Lab studies found that it makes the answer more engaging and easier to remember.
- On 8,977 randomized headline tests, clicks peaked at a middle level of concreteness: too vague lost, and too specific lost too.
- Controlled studies found that clickbait-style teases did not reliably win clicks. “You won’t believe what happened next” is the vague form that loses.
- A question mark is not curiosity on its own. On real headline tests, the headline with a question mark won only 42.3% of its pairs.
- For a thumbnail, show the subject clearly and hold back one thing: the result, the method or what happens next. Then test it on the same video with A/B Testing.
Why curiosity can raise clicks, and what the research measured
Loewenstein’s information-gap account, from 1994, describes curiosity as the feeling that arises when people notice a gap between what they know and what they want to know. Lab studies have since found that curiosity about an answer engages reward-related brain regions and improves memory for the answer, as Kang and colleagues showed with trivia questions in 2009.
Those studies used trivia, not thumbnails. The step from “curiosity makes an answer more engaging” to “a curious thumbnail gets more clicks” is a reasonable inference, but the evidence on real clicks is narrower than most advice suggests, and it comes with an important limit.
How much to hold back: the middle level of concreteness
The clearest evidence comes from a registered report by Aubin Le Quéré and Matias, published in 2025, which analysed 8,977 headline experiments from a news site that showed different headlines for the same story to randomly assigned readers. The effect of making a headline more concrete depended on the others in the test. When the other headlines were vague, the more concrete one won. When they were already very concrete, making one even more concrete lost clicks. Clicks peaked in the middle.
For a thumbnail, the middle is easy to describe. The viewer should be able to tell at a glance what the video is about: the subject, the situation, the stakes. What you hold back is one thing the video exists to reveal.
| Too vague | Middle ground | Gives it away |
|---|---|---|
| A shocked face and “WAIT FOR IT” | A sourdough loaf cut open, the crumb hidden behind the knife, “Day 30” | The finished loaf with “PERFECT CRUMB: 75% WATER, 4 FOLDS” |
| A blurred box and “???” | Three microphones side by side with one price tag hidden | The cheap microphone with a “WINNER” badge and its price |
The examples are illustrations, not tested thumbnails. In the middle column the viewer knows what the video is about and has one specific reason to click.
Why vague teases and clickbait do not win
Two controlled studies tested clickbait directly. Molina and colleagues (2021) ran three studies and found that clickbait headlines did not reliably attract more clicks or engagement than plain ones. Scacco and Muddiman (2020) found that curiosity-style news headlines did not outperform traditional ones in a lab experiment or in a field test on a news site. Other studies of curiosity headlines have found mixed results, which is why the middle-ground finding above, drawn from thousands of real tests, is more useful than a simple yes or no.
That fits the middle-ground finding. “You won’t believe what happened next” withholds everything, including what the video is about, so it sits at the vague end. It also creates an expectation of a surprise that the video then has to meet.
Are question marks a good curiosity trigger?
Not by themselves. We checked the question-mark rule against real clicks on 17,236 pairs of headlines that ran against each other on the same image. The headline with a question mark won only 42.3% of its pairs, and 33.8% of the pairs where the difference in clicks was clear. That is worse than a coin flip.
A question can still be the right wording when the video genuinely answers it, such as “What’s inside a $5 smartwatch?” on a teardown. What does not work is adding a question mark to make an ordinary statement look mysterious.
Visual curiosity elements that stay honest
- A partial reveal: the object, the result or the screen is shown but one decisive detail is hidden or cropped.
- A before without the after: the starting point is clear and the outcome is what the video shows.
- A result without the method: “Cut my editing time in half” with the editing timeline visible, and the how left for the video.
- A real reaction: a face reacting to something in the video. Faces draw attention, and on Instagram photos with faces received about a third more likes and comments, though that study did not measure clicks on videos. The reaction must match what happens in the video.
Each of these holds back one thing and makes the rest clear. If the viewer cannot tell what the video is about, the thumbnail has moved to the vague end.
Curiosity versus clarity in your own packaging
Curiosity and clarity pull against each other, and the second one usually matters more. In our test of 321 pairs of videos from the same channel, the curiosity rating of the video that beat its channel’s average was only slightly higher than that of the one that fell short (an effect size of 0.24). Promise clarity separated them more (0.31), the largest gap of all the factors we score.
A practical check: when you add mystery to a thumbnail, ask whether a viewer could still say in one sentence what the video is about. If not, you have bought curiosity with clarity, and the evidence suggests that is a poor trade. You can score your thumbnail and title and read the curiosity and promise clarity ratings side by side: when clarity falls as curiosity rises, the gap has become a blind spot.
How to test a curiosity thumbnail on your channel
Test on the same video. YouTube Studio’s A/B Testing lets eligible long-form videos try up to three thumbnails, shows them to different viewers, and picks the winner on watch time rather than click-through rate. That is useful here, because a curiosity thumbnail that wins clicks but loses viewers in the first seconds will not win the test.
- Change one element: hide the result in one version and show it in the other, keeping the subject, colours and text otherwise the same.
- Let the test finish. YouTube says a test should be completed within about two weeks.
- After publishing, check the Intro figure in the audience retention report, the share of viewers still watching after 30 seconds. If it drops, make sure the opening delivers what the thumbnail set up.
Do not compare two separate uploads or two different videos and treat the result as a test. Their audiences and timing differ in ways you cannot see.
Update, 19 October 2026. This article replaces our two earlier articles on curiosity in thumbnails, including an earlier version of this page; the other one now redirects here. It was rewritten from the sources below on 16 October 2026: the earlier versions gave advice with no source, such as presenting question marks as an effective trigger, which real click data contradicts.
Sources
- Loewenstein, G. (1994). The psychology of curiosity: A review and reinterpretation. Psychological Bulletin, 116(1), 75–98. doi:10.1037/0033-2909.116.1.75
- Kang, M. J., Hsu, M., Krajbich, I. M., Loewenstein, G., McClure, S. M., Wang, J. T. and Camerer, C. F. (2009). The wick in the candle of learning: Epistemic curiosity activates reward circuitry and enhances memory. Psychological Science, 20(8), 963–973. doi:10.1111/j.1467-9280.2009.02402.x
- Aubin Le Quéré, M. and Matias, J. N. (2025). When curiosity gaps backfire: effects of headline concreteness on information selection decisions. Scientific Reports, 15, 994. doi:10.1038/s41598-024-81575-9
- Molina, M. D., Sundar, S. S., Rony, M., Hassan, N., Le, T. and Lee, D. (2021). Does clickbait actually attract more clicks? Three clickbait studies you must read. Proceedings of CHI 2021. doi:10.1145/3411764.3445753
- Scacco, J. and Muddiman, A. (2020). The curiosity effect: Information seeking in the contemporary news environment. New Media & Society, 22, 429–448. doi:10.1177/1461444819863408
- Bakhshi, S., Shamma, D. A. and Gilbert, E. (2014). Faces engage us: Photos with faces attract more likes and comments on Instagram. Proceedings of CHI 2014, 965–974. doi:10.1145/2556288.2557403
- YouTube Help. A/B test titles & thumbnails. support.google.com/youtube/answer/13861714; Measure key moments for audience retention. support.google.com/youtube/answer/9314415
- ThumbSignal. We tested our title rules against real clicks, for the question-mark result, and the method and evidence page for the 321-pair test.
Frequently Asked Questions
Does curiosity increase CTR on YouTube thumbnails?
It can, within limits. On real headline tests, clicks peaked at a middle level of concreteness, and vague teases lost. No study we found measured curiosity in YouTube thumbnails directly, so test it on your own videos.
Should I use question marks in thumbnails or titles?
Only when the video genuinely answers the question. On 17,236 pairs of real headline tests, the headline with a question mark won only 42.3% of its pairs. A question mark does not create curiosity by itself.
How much should a curiosity thumbnail reveal?
Enough for the viewer to know what the video is about at a glance, and no more than that about the result. Hold back one thing the video exists to show, such as the outcome, the method or the winner.
Can curiosity thumbnails hurt my channel?
They can when the curiosity is vague or the video does not pay it off. A viewer who cannot tell what the video is about may scroll past, and one who clicks and finds nothing surprising may leave early. Check the Intro figure in your retention report after publishing.
Can I use curiosity thumbnails without text?
Yes. A partial reveal, a before without the after, or a real reaction can carry curiosity without words, as long as the subject of the video is still clear from the image alone.

