Do Faces Boost Thumbnails?

A large-scale analysis of around 300,000 viral YouTube videos suggests that using faces in thumbnails does not automatically lead to better performance. While faces are widely used, the data shows they perform broadly the same as thumbnails without faces, with results varying significantly depending on niche, format, and audience.

The idea that faces are either essential or harmful has recently gained traction online. One claim circulating on X suggested that showing your face in thumbnails could be reducing views, especially for lesser-known creators. This sparked debate, as such advice often gets simplified into hard rules that do not reflect how YouTube actually evaluates content.

Nate Curtiss, Head of Content at 1of10 Media, challenged the claim and pointed to a detailed dataset analysing high-performing videos across tens of thousands of channels. His argument was that thumbnail performance is far more situational than blanket advice suggests.

The dataset examined “outlier” videos from 2025, measuring performance based on how much a video exceeded a channel’s typical view count. When comparing thumbnails with faces to those without, the overall difference was minimal, despite faces appearing in a large proportion of videos.

More meaningful insights emerged when the data was broken down further. Larger channels saw a small benefit from using faces, while smaller creators did not experience the same lift. Performance also varied by category. For example, finance-related content tended to perform better with faces, while business-focused videos often performed worse. Thumbnails featuring multiple faces were shown to outperform those with a single face.

YouTube’s own guidance helps explain why click-through rate alone does not tell the full story. The platform’s thumbnail testing tools prioritise watch time rather than CTR, meaning thumbnails are judged on both their ability to attract clicks and to keep viewers watching once they arrive.

According to YouTube’s Creator Liaison, thumbnails are tested until one version generates more total watch time. This approach is designed to discourage clickbait, as a thumbnail that attracts clicks but leads to poor retention is unlikely to perform well overall.

YouTube has increasingly encouraged creators to use A/B testing within Studio, allowing them to test multiple titles and thumbnails to see which combinations deliver the best results in real viewing conditions.

Audience familiarity also plays a role. YouTube advises creators to think differently about subscribers, casual viewers, and new audiences. Familiar faces may work better for established channels, while broader visual cues may be more effective when targeting viewers who do not yet recognise the creator.

The key takeaway is that there is no universal rule. Faces are not a guaranteed advantage, nor are they a disadvantage by default. Their effectiveness depends on the topic, the audience, and how well the thumbnail aligns with the video content.

YouTube increasingly treats the title, thumbnail, and opening moments of a video as one combined package. A thumbnail should set an expectation that the video immediately delivers on, particularly on mobile, where viewers often see the transition from thumbnail to video almost instantly.

Ultimately, the ongoing debate reflects a desire for simple answers in a system that rewards nuance. As top creators have pointed out, it is not about whether a face appears, but how it is used. With better testing tools now available, creators are better served by experimenting and focusing on watch time rather than relying on one-size-fits-all thumbnail advice.