The advice most creators hear about face thumbnails goes something like this: faces increase CTR, so use your face. The evidence behind the first part is real. The conclusion people draw from it is not.
The same data that shows faces tend to improve click-through rates also shows considerable variation — across niches, across channel types, and most importantly, across what the face is actually doing in the thumbnail. A strong, specific expression in the right context consistently outperforms a generic one. In the wrong context, a face adds nothing, and the space it occupies could have communicated more.
The useful question isn't "should I use a face?" — it's "does this face, with this expression, in this context, communicate something the viewer will respond to?" The answer to that is nuanced enough to be worth working through properly.
Faces do tend to improve CTR, but the gain is almost entirely determined by the expression and context, not just the presence of a face. A specific, legible expression that relates to the video's content is a strong hook. A generic or posed expression adds little. In some niches, a well-chosen subject visual outperforms any face thumbnail. The expression matters more than the face itself.
What the data actually shows
The consistent finding across YouTube CTR studies and creator A/B tests is that thumbnails with visible human faces — particularly faces displaying a clear emotional expression — tend to outperform those without. YouTube's own creator resources have referenced this pattern, and enough independent channel experiments confirm it to treat it as reliable signal rather than myth.
The nuance that usually gets dropped: this is a population average across many channel types and niches. Channels that see the strongest gains from face thumbnails are almost always personality-led — channels where the creator is the product, and the viewer's ongoing relationship with that person is a significant part of why they click. The face in the thumbnail isn't just a visual element — it's a signal about whose perspective this video takes, and that signal has value beyond the expression itself.
For channels where the content is the primary draw rather than the creator, the face advantage shrinks. A tutorial channel, a product review channel, or a gameplay highlights channel often sees no meaningful CTR difference between face and no-face thumbnails — because the viewer's decision is driven by what the video covers, not who presents it.
Faces trigger immediate attention responses in viewers — this is well-established in visual psychology. But a YouTube browse feed is not a blank page. In niches where face thumbnails are the norm, the attention advantage of a face over no face is considerably smaller. A different compositional choice can stand out more than yet another face thumbnail. The competitive context of your specific feed matters as much as the general data.
When faces work
The conditions that make a face thumbnail perform are more specific than most creators realise. They're not about whether a face appears — they're about what the face is doing and what context it's doing it in.
A specific, readable emotional expression
The expression matters more than anything else. An expression that communicates "something unexpected happened here" or "I know something you don't" gives the viewer a piece of information about the video's content. That's a hook — an implicit promise of what watching will reveal. A default open-mouth surprise face, a neutral smirk, or a generic "engaged" look communicates nothing specific. The viewer has seen it on every channel. It creates no curiosity because it contains no information.
The practical test: can a viewer infer something meaningful about this video from the expression alone, before reading the title? If yes, the expression is working. If the answer is "a person is present and moderately interested in something," it isn't.
Eye contact or clear directional gaze
Faces that make direct eye contact with the viewer create an engagement pull that's well-documented in visual psychology. Alternatively, a face that's clearly looking at something specific within the frame creates curiosity: the viewer wants to know what the subject is looking at. Both work. A face looking vaguely sideways, slightly downward, or at nothing in particular creates no meaningful pull — the gaze has no target and communicates no intention.
The creator is part of the value proposition
For commentary, coaching, personal finance advice, fitness content, and vlog-adjacent channels, the creator-viewer relationship is a real component of the value. In these contexts a face in the thumbnail reinforces the ongoing connection — viewers who know the creator will click on a face thumbnail based partly on that relationship, independent of the specific expression. This is why established personality-led channels can get away with weaker expressions than newer channels — the face itself carries accumulated meaning.
When faces hurt
Faces work against you when they're in the wrong context, when the expression is generic, or when the face competes with the visual information the thumbnail actually needs to convey.
Niches where the content is the hook
Tutorial channels, technical how-to content, product reviews, recipe videos, and a significant portion of gaming content — particularly gameplay highlights, speedruns, and builds — often perform as well or better with a subject-focused thumbnail. The viewer is making a decision based on the video's content, not the creator's reaction to it. A thumbnail that shows the outcome, the subject, or the key visual of the video communicates the value proposition more directly than a reaction shot. The face here is not a hook — it's filler.
A posed expression rather than a genuine one
Viewers process authentic and performed expressions differently, faster than they can consciously articulate it. A "thumbnail face" captured specifically for the thumbnail — in front of a camera, without a natural trigger for the expression — often reads as hollow. It occupies the frame without communicating anything. This is one reason some creators perform significantly better with face thumbnails pulled from actual footage than from dedicated photoshoots. The moment is real, and the expression carries the weight of what actually happened.
When the face isn't large enough to read
A face that isn't legible at 168 pixels wide — the size at which thumbnails appear in a mobile browse feed — can't do the work that faces are supposed to do. A face taking up 20% of the canvas, surrounded by text elements and game UI and scene detail, is a postage stamp. The expression isn't readable, there's no engagement pull, and the space could have been used for something that communicates more clearly at that size. If you're using a face, it needs to be large enough to matter.
Before finalising any face thumbnail, resize it to 168×94 pixels and look at it among other thumbnails. If the expression isn't immediately readable at that size, it isn't contributing what you're hoping it contributes. Either scale the face up, simplify the composition around it, or reconsider whether a different focal point does more work.
The expression matters more than the face
This is the most practically useful takeaway from all of the above: the variable that determines whether a face thumbnail performs is the expression, not the face.
A face with a strong, specific, legible expression is a narrative element. It tells the viewer that something happened, that something is at stake, or that the creator's reaction to a specific event is worth seeing. Each of those is an implicit promise — a reason to click. A face with a generic expression is just a face. It tells the viewer a human being is associated with this video, which is true of every video on the platform and provides no differential information.
The practical implication: if you're going to use your face in a thumbnail, commit to an expression that means something for this specific video. Not wide-eyed surprise as a default mode — something that relates to what happens, what's at stake, or what the viewer is about to learn. The expression that's specific to this video will outperform the general "excited creator" face almost every time.
Some creators find their best thumbnail faces by screenshotting genuine reaction moments from the video footage rather than staging a shoot. The expression is real — triggered by something that actually happened — and it reads that way. It also tends to be more specific to the video's content than a posed shot taken before filming. Worth trying alongside your usual process.
How this plays out by niche
Niche conventions shape audience expectations, and audience expectations shape what works. "Use faces" means very different things across different YouTube communities.
Faces work strongly here. The creator-viewer relationship is central to the value — viewers are trusting someone with something important. Vulnerability, calm authority, and genuine "I've been there" expressions connect. Polished, overly professional expressions often underperform more candid ones. Authenticity registers particularly strongly in this niche.
Works well for coaching, transformation, and motivation content where the creator-viewer dynamic matters. For workout tutorials and nutrition breakdown videos, the subject — an exercise, a meal, a physical result — often communicates the video's value more directly than a creator reaction shot. Worth testing both directions.
Faces work, but expression matters particularly in this niche — finance audiences are sceptical by default. Genuine concern, calculated confidence, or "I know something you don't" performs better than performed shock or enthusiasm. The niche has strong conventions around credibility signals; an obviously performed expression reads as hype. See the finance channel thumbnails post for more on this.
Highly dependent on sub-niche. Personality-led channels — Let's Play, commentary, reaction content — benefit from creator face thumbnails when the expression is specific. Gameplay highlights, speedruns, builds, and tutorial content often perform as well without one. The gaming niche is also one where a strong, differentiated subject visual can stand out more than another face in a feed full of faces. See the gaming thumbnails post for the full picture.
Often performs better without a creator face. The viewer's decision is based on what the video teaches, not who presents it. Thumbnails that show the outcome, the interface, or the specific problem being solved communicate the value proposition more directly. Faces add something only when the creator has a significant established audience whose relationship with the presenter is a draw in itself.
Use the expression, not just the face.
The question isn't whether to use a face — it's whether this face, with this expression, communicates something specific about what this video offers. A genuine, legible expression that's relevant to the content is a strong hook. A posed or generic expression takes up space that a clearer visual could use better. Know your niche conventions, size the face so it reads at 168 pixels, and ask whether the expression tells a story — or just fills the frame.
On average, yes — thumbnails with visible human faces tend to outperform those without, particularly for personality-led channels where the creator-viewer relationship is part of the value. But the gain depends heavily on what the face is doing. A specific, legible emotional expression contributes meaningfully to CTR. A generic or posed expression adds little, and in some niches a strong visual without a face outperforms a weak face thumbnail.
No. Faces work best when the creator is the primary draw — commentary, coaching, personal development, vlog-style content — and when the expression communicates something specific about the video. Tutorial channels, product reviews, gameplay highlight content, and technical how-to videos often perform as well or better with a subject-focused thumbnail. The question is whether a face adds more information than the alternative.
Specific expressions outperform generic ones. An expression that communicates "something surprising happened" or "I know something you don't" tells the viewer something about the video. A default wide-eyed shock or a slightly-open-mouth expression that appears on every thumbnail conveys nothing distinctive. The most effective expressions are ones where a viewer can infer something about the video's content from the expression alone.
Tutorial channels, speedrun and gameplay highlight content, technical reviews, cooking recipe videos, and some finance channels perform competitively without creator faces — particularly when the video's subject is more visually compelling than a reaction shot. In these contexts the thumbnail communicates more by showing the outcome, the object, or the moment than by showing a creator expression. Worth A/B testing against a face version.
Large enough that the expression reads clearly at 168 pixels wide — the size at which thumbnails appear in a mobile browse feed. A face occupying less than about a third of the canvas, surrounded by other competing elements, is usually too small to contribute what faces are supposed to contribute. If the expression isn't legible at thumbnail size, the face isn't doing its job. Scale it up or simplify the composition around it.
Yes, significantly. The expression is the variable that determines whether a face thumbnail performs. A specific, genuine expression that relates to the video's content creates a narrative hook — the viewer wants to know what caused it. A posed or neutral expression simply occupies space. This is why some creators' thumbnails perform much better when the face is pulled from actual footage rather than a dedicated thumbnail shoot.
Get a thumbnail built around what actually drives clicks in your niche.
Brief, composition, expression framing, revisions — all included. One thumbnail, designed for your channel.
Test Drive — €4524h delivery · Unlimited revisions · No contracts