How to Write Social Copy That Stops the Scroll: Hooks, Captions, and CTAs

The advice on writing scroll-stopping copy has itself become the thing people scroll past. Every guide offers the same swipe file of openers, which is precisely why those openers no longer work.

You know the list. Start with a bold statement. Ask a question. Use "nobody talks about this." Create a curiosity gap. Add line breaks and emojis. End with a clear call to action.

None of it is wrong. All of it is now so widely followed that the pattern itself has become a signal — readers recognise the shape of a templated post before they read a word of it, and scroll on. Which means the useful version of this article is about the mechanics underneath the templates, and about two or three things that decide the outcome before your writing gets a chance.

You're not competing with your competitors

Start here, because it reframes everything else.

A reader deciding whether to stop on your post is not comparing it to your competitor's post. They are comparing it to whatever is directly beneath it in the feed — which might be a friend's holiday photo, a news story, or something funny. That's the actual competitive set, and it's a much harder one.

The practical consequence: "better than what other brands in our category post" is a low bar that feels like a high one. Your copy has to earn attention against content that had no commercial motive at all, from people the reader actually knows.

It also explains why category-standard copy underperforms even when it's well written. Being the best-executed version of a familiar format still reads as familiar, and familiarity is what the thumb is filtering for. It is the same dynamic that has made everyday creators outperform polished endorsements — the less it looks like marketing, the further it gets before being filtered.

Your hook isn't the first line — it's the first visible line

This is the mechanical point that almost nothing written on this topic covers, and it decides more posts than any amount of craft.

Feeds truncate. Your caption is cut off with a "more" link long before the platform's hard character cap, at a point that varies by device, orientation and app version. Everything past that cut is invisible until someone taps — and nobody taps because of text they can't see.

So the real question isn't "is my hook good." It's "is my hook above the fold."

Two rules follow from this, and they're worth more than any hook template:

Featured Recommendation AD · AFFILIATE

4.4 / 5.0

Wati

WhatsApp Business API platform for team inbox, broadcasts, chatbots and automated customer conversations.

Best for: SMBs automating sales & support on WhatsApp

Visit Webs

View details Compare

• Put the specific thing first. Not the setup, not the context, not "As marketers, we all know that…" — the actual claim. Context can come after the cut, because by then they've decided to keep reading.
• Engineer the line before the cut. If the visible portion ends on a complete, satisfying thought, you've given the reader permission to move on. If it ends mid-tension, the tap is nearly automatic.

Why swipe-file hooks stopped working

Every one of those templates worked when it was rare. Each became less effective in proportion to how widely it spread — and the popular ones are now saturated enough to be actively counterproductive.

The mechanism worth understanding A hook works because it's specific, not because it's a hook. Copy the shape of a good opener and you get the shape. The specificity was doing the work all along.

What actually makes an opening line stop someone, underneath the templates:

• A specific claim. "Most captions bury the point past line three" beats "your captions might not be working." Numbers, named things, concrete situations.
• A named reader. Making it obvious in the first five words who this is and isn't for. The people it excludes were never going to convert; the people it includes now feel addressed.
• A contradiction. Not manufactured contrarianism, but a genuine correction of something the reader probably believes. It works because it creates a real question in their head, and the only way to resolve it is to keep reading.
• Something only you could say. A specific from your own work, a number from your own data, a thing that happened. Unswipeable by definition, which is the whole point.

The hook has to be a cheque the post can cash

Curiosity-gap openers get taught as a technique and used as a trick, and the difference matters more than it used to.

If your first line promises a revelation and the post delivers a platitude, you've won one tap and taught that reader something about your account. Do it consistently and you're training your own audience to skip you — while the platform, watching people expand and immediately exit, draws its own conclusions about the value of showing you to more people.

The rule is unglamorous: only open with a promise you're going to keep in the next forty words. If the payoff isn't there, the hook is the problem, not the reader.

One idea per post

Most weak captions are two decent posts fighting each other.

It happens because writing a post feels like an opportunity, so everything relevant gets included. The result is a caption where the reader has to work out what they're supposed to take away, and the usual resolution to that work is scrolling.

The discipline is to write the one sentence you want someone to repeat, then delete anything that isn't in service of it. Whatever you cut is a future post — that's not a loss, it's a content calendar, and it's a good deal cheaper than the process described in building a social media content calendar from scratch.

If you find yourself unable to name the one sentence, the problem is usually upstream of the caption. A post with no single point is often a symptom of not having decided what the account is for, which is territory covered in building a social media strategy from scratch.

The CTA that quietly costs you reach

This is where the standard advice does the most damage, because it's given confidently and it's often wrong now.

Feeds generally deprioritise posts that send people off-platform. So the "link in bio" or in-post external link — the CTA every guide recommends — can suppress the distribution that would have made the link worth having in the first place. You get the click-through rate you wanted on a much smaller audience.

Meanwhile "comment below with your thoughts" is engagement bait, which platforms have spent years learning to discount, and readers have learned to ignore.

A more reliable hierarchy:

• Ask for a save. The strongest signal available on most platforms, because it means the content had future value. It's also honest — you can only ask for a save if you've written something worth keeping.
• Ask for a share. Nearly as strong, and it's distribution rather than just signal.
• Ask a real question. Not "thoughts?" — a specific question that a specific person would actually have an answer to. The difference between engagement bait and a conversation is whether you'd be interested in the reply.
• Put the link in the first comment or the profile, and mention it plainly rather than making it the post's purpose. Detail on why in why captions and on-screen text now matter more than hashtags.
• Sometimes: no CTA. A post that's genuinely useful and asks for nothing performs better than most people expect, and it earns you the right to ask next time — which is the slow mechanic behind growing an engaged community from zero.

The caption does a different job on each platform

Cross-posting one caption everywhere means it's optimised for none of the jobs it has to do.

• Under short-form video, the caption supports something the viewer is already watching. It should add context the video can't carry, not repeat what they can see — an extension of the thinking in short-form video for brands and in the split between clips for discovery and longer video for trust.
• On text-first platforms, the caption is the content, and the hard limits above force real editing. Constraint is genuinely useful here.
• On LinkedIn, the caption usually carries the whole argument, and the truncation point matters most because the reader is deciding whether to commit to a longer read.
• On Pinterest, the caption functions largely as search text — which is a different craft entirely, closer to writing for discovery than for a feed.

That last point generalises further than most people realise. Social copy is increasingly indexed and surfaced through search, both on-platform and off, which is the argument in making your posts discoverable in search. Writing captions purely for the feed leaves the longer tail of discovery on the table.

If maintaining genuinely distinct copy for six platforms sounds unsustainable, that's a real signal — and usually an argument for fewer channels done properly, as in why brands are cutting their channel list.

Write for the metric you actually want

Likes measure agreement. Saves and shares measure usefulness. These pull in different directions and you have to choose before you write.

A post confirming something your audience already believes collects likes cheaply and produces nothing. A post teaching something specific collects fewer likes and more saves, and saves correlate far better with people who act later.

The copy implication is direct. Writing for agreement means broad statements and shared frustrations. Writing for usefulness means specifics, numbers, steps, and the willingness to be less immediately likeable. Most brands say they want the second and write the first, then conclude social doesn't work — a diagnosis worth testing properly with a proper social audit before you accept it. The shift in what engagement data shows points the same way.

The self-edit pass

Run every caption through this before it goes out. It takes two minutes and catches most of what's above.

• Read only what's visible before truncation. Is the specific idea in there?
• Does the visible portion end mid-tension, or does it let the reader off?
• Delete the first sentence. Is it better? (It usually is.)
• Could a competitor publish this caption unchanged? If yes, rewrite.
• Does the post pay off what the opening promised, within the next forty words?
• Is there exactly one idea, or two fighting?
• Does the CTA send them away from the platform — and is that worth the reach it costs?
• Which metric is this written for, and is that the one you actually want?

Treat the results as data rather than opinion. Openings are testable in exactly the way described in creative testing — same content, different first lines, and the platform will tell you which one it prefers faster than any style guide will. If copy is the persistent bottleneck rather than distribution, that's usually a social media resourcing question rather than a talent one.

The short version

You're competing with the next post in the feed, not with your category. Your hook is whatever is visible before truncation, so put the specific claim in the first line and engineer the cut to land mid-tension. Skip the swipe file — the specificity was always what worked, not the shape. Only promise what you'll pay off in the next forty words. One idea per post; the rest is next week's content. Be careful with the CTA, because sending people off-platform can cost more reach than the click is worth, and asking for a save beats asking for a comment. And decide which metric you're writing for before you write, because likeable and useful are rarely the same sentence.

The reason most social copy fails isn't that it's badly written. It's that the point was on line four.