FormatsLatestHow we measure
The library / Formats

The cover frame is a decision, not a leftover

It is the still that represents the video everywhere the video is not playing. Most people accept whatever the editor picked.

GrowthCamel research desk · 2026-10-04
Image — to addChoosing which frame represents the whole thing/images/library/cover-frame.jpg · 1200 × 675 (16:9)Photorealistic editorial documentary photograph, shot on a 35mm full-frame camera with a fast prime lens, natural available light, shallow depth of field with a soft falloff, fine natural film grain. Warm muted palette of creams, oatmeal, warm greys, wood and skin tones. Real lived-in spaces with visible imperfection. Candid, unposed, caught mid-moment; subjects absorbed in what they are doing and not addressing the camera. Natural casting of ordinary-looking people of varied ages and backgrounds, ordinary clothes, no styling. Close over-the-shoulder view of hands holding a phone, scrubbing slowly back and forth through a video timeline to land on one frame, the rest of the room dark and out of focus. Deliberate, fiddly, unhurried. AVOID: no text, no lettering, no captions, no numbers, no charts, no graphs, no infographics, no UI mockups, no logos, no brand marks, no watermarks, no recognisable public figures, no stock-photo posing, no direct-to-camera smiling, no thumbs up, no handshakes, no glossy corporate office, no oversaturated colour, no HDR, no lens flare, no 3D render, no illustration.
In this story

The cover is the frame that stands in for the video on your profile grid, in search, and anywhere the video is not currently playing. Most creators accept whatever frame the editor offered, which means the most reproduced image of their work was chosen by a default.

This mattered moderately when the grid was the only place it appeared. It matters more now that Instagram content can surface in search, where the cover is the image attached to a result.

Two jobs, often in conflict

In the grid, the cover is seen alongside your other covers by someone deciding whether to follow you. Here consistency wins: a grid that looks like one body of work reads as a practice, and a grid of unrelated stills reads as a feed.

In search and recommendation, the cover is seen alone by someone with no context. Here legibility wins: one subject, readable at thumbnail size, and obviously about something.

These pull in different directions. A consistent grid tends toward similar covers; a legible result tends toward whatever is clearest for this specific video. Decide which surface you are optimising for rather than failing at both.

A default cover is a decision made by an editor that has never seen your grid.

What makes a cover work

  • One subject, large. Thumbnails are small. A wide shot containing the whole scene contains nothing at 120 pixels.
  • A moment, not a pose. The frame mid-action outperforms the frame where someone is standing still waiting to begin.
  • Not the first frame. The first frame is usually someone settling. The good frame is typically a second or two in.

The text question

Adding a caption to the cover is common and it carries a specific risk. Instagram has said it reduces the visibility of reels that are majority text, and a cover dominated by a headline is exactly that. A few words over a legible image is fine; a text card is not a cover.

There is also a quieter cost: a cover built around text is unreadable at the size it is most often seen, so it fails at the job it was added for.

The cheapest version of this

Spend fifteen seconds choosing. That is the whole intervention. The gap between a default cover and a chosen one is larger than almost any other fifteen seconds available in the edit, and unlike most format advice it costs nothing and risks nothing.

Sources

growthcamel.
What actually works in short-form, measured.
ProductPricingLibrary TermsPrivacy © 2026 Growth Camel