How captions carry the muted watch

Why captions decide outcomes#

A large share of Feed and Reels viewing starts with sound off, and many sessions stay muted throughout. Without captions the viewer gets motion without meaning, which flatters view counts while starving comprehension and clicks. Captions are therefore structure, not accessory. Start at the Meta Ads hub for the family view, and see Meta own video specs for caption support.

Captions also widen access for viewers with hearing needs and for bright commutes where audio stays down. Hooks depend on them, and video craft assumes them throughout.

Legibility rules for phones#

Set captions large enough to read at arm's length on a mid range phone: short lines of four to six words, heavy weight, high contrast with a solid or soft edged backing plate. Pure white on busy footage without backing fails; yellow on white fails faster. Keep lines to two at once so the visual still breathes.

Time captions to speech with clean line breaks at phrases, never mid phrase, and hold the payoff line slightly longer than speech pace. Safe areas decide vertical placement, and our overlay guide keeps captions distinct from in image text.

Burned in versus SRT uploads#

Burned in captions guarantee placement, style and timing on every surface, survive reuploads and reframes, and suit Reels where interface crowds the frame. SRT uploads keep text selectable and translatable, suit longer explainers, and need en_GB spelling checked line by line because auto sync mangles product names. Many teams ship both: burned in for control, SRT as backup where supported. Meta documents uploads in its captions help article.

Whichever route, archive a textless master so future edits never fight burned in words. Diversification compounds when winners stay editable.

Placement that survives interface#

Reels interface occupies the base and top of frame, so centre captions in the middle third with clearance top and bottom, and never park disclaimers in the bottom 40 per cent. Feed is kinder but still overlays calls to action, so lift captions above the action row. Test on a real phone in Feed, Reels and Stories before spending.

Caption style should match brand voice without novelty fonts: one family, two weights, consistent colour. Novelty per video reads as noise in a mixed ad set. Advantage+ creative can adjust text placement, so review its previews with captions on.

Caption QA before spend#

Run every cut through a muted two second test: mute, play, pause at second two and ask what the viewer knows. If brand, viewer and payoff are unclear, rewrite before re-editing. Then watch fully muted for sync drift, spelling, product names and line breaks. Fix at source in the edit project, not in export patches.

Log caption variants with thumb stop and completion rates; readable captions lift both. Ask our team for a caption audit when completion trails views, and revisit A/B discipline before blaming the concept.

Fix at source in the edit project, never in export patches, so the next reframe inherits clean timing. A captioned master with tidy project files pays back across translations, ratio reframes and seasonal relaunches.

Multilingual captions and tone#

Translate meaning, not words: short lines survive translation better than dense sentences, so simplify source copy before sending it out. Keep product names, offer terms and calls to action identical across languages, and recheck line breaks after translation because word lengths shift timing and layout. Use en_GB spelling throughout source files so variants stay consistent.

Tone should match the market: plain and direct where clarity decides, warmer where community decides, always in one caption family. Review translated cuts muted on a phone with a native reader before spending, because auto translation mangles idiom. Diversification multiplies when each language carries its own captioned cuts, and the 3-2-2 method shows how to slot them into tests.

Frequently asked questions#