Captions get less attention than they deserve, partly because their effect is indirect. A caption does not obviously drive reach the way a thumbnail does. But it determines whether someone stops, whether they expand, whether they comment — and comments and dwell time are among the strongest distribution signals on every major platform.
The truncation problem
Almost every platform cuts captions off. Instagram shows roughly the first line or two before "more." LinkedIn truncates behind "see more." TikTok shows very little. The practical consequence is that your caption has two distinct parts with different jobs, and most people write them as one continuous block.
| Part | Job | Length |
|---|---|---|
| Visible fragment | Earn the expansion, or deliver value on its own | One sentence, roughly 8–15 words |
| Expanded body | Deliver the substance and prompt a response | Whatever the content needs |
The expansion click is itself a tracked engagement signal on several platforms, which means a caption that earns expansions is doing measurable good beyond just being read.
Openings that earn the expansion
Patterns that reliably work, with the mechanism behind each:
- A specific number. "I cut my hashtags from 30 to 6 and reach went up 40%." Numbers signal there is actual content behind the line, not vibes.
- An admitted mistake. "I did this wrong for two years." People expand for the correction — and it builds credibility rather than spending it.
- A contradiction of accepted advice. "Posting more often made things worse." Tension is the cheapest reason to keep reading.
- A direct question with a real answer coming. "Why do some tags stop working after a month?" Questions work when the caption actually answers them; they fail as pure bait.
- A concrete situation. "A client sent me 40 hashtags and asked which to keep." Specificity implies a story.
What goes in the body
Once expanded, the reader has committed. The body should reward that rather than padding it.
- Deliver the specific thing the opener promised. Do not make them wait through three lines of preamble.
- Keep paragraphs to one or two sentences. Mobile screens make dense blocks genuinely hard to read, and readers bounce from walls of text.
- Put one clear idea per caption. Captions that cover three topics get remembered as covering none.
- Close with something answerable. Not "thoughts?" — which is too vague to answer — but a specific question: "Which platform do you find hardest to get reach on?"
The comment prompt, done properly
Comments matter more than likes on nearly every platform, because they indicate genuine engagement rather than a reflex tap. But most comment prompts fail for a specific reason: they ask a question that requires effort to answer.
"What’s your content strategy?" is a hard question. Answering it well takes three paragraphs, so nobody answers. "Do you use more than 10 hashtags — yes or no?" is trivially answerable, which is why it gets replies. Lower the cost of responding and the response rate rises.
Two other things that work: asking people to pick between two named options, and asking about a common frustration. Both are easy to answer from experience without composing anything.
Platform differences worth respecting
- Instagram. Longer captions have been performing well when the first line earns the expansion. Storytelling works here.
- LinkedIn. The first two lines are critical. Line breaks matter enormously — dense paragraphs perform badly. Professional register, but personal specifics.
- TikTok. Short. The spoken hook carries the work; the caption is mostly context and searchable phrasing.
- YouTube. First two lines of the description are visible and should restate what the video delivers in searchable language.
- Pinterest. Not really a caption — a description, written for search rather than for engagement.
- X. The whole post is the caption. Every word competes for a hard character budget.
A quick test before posting
Read only the first line. Would you expand it? If the honest answer is no, rewrite that line and leave the rest alone. It is the highest-return edit available on any caption, and it takes thirty seconds.