Locale: en
When promoting an article on social media, an overly helpful summary can remove the reason to click. But a vague “shocking truth inside” sounds suspicious. The useful zone sits between those extremes: tell people what the story is about, why it matters, and what kind of evidence supports it—then leave one meaningful piece unresolved. Social copy should work more like a 15-second trailer than a plot summary.
A common social post looks like this:
“A study found X. The answer is Y. The reason is Z. Read more here.”
Very helpful.
Also very effective at eliminating the need to read more.
It is like a movie trailer that reveals the killer, the motive, the twist, and the ending. The audience can reasonably respond, “Thanks. I think I’m done.”
The opposite is not better:
“You won’t believe the shocking secret they don’t want you to know…”
Now there is a gap, but no trust.
The goal is not to hide everything. It is to show enough value to justify attention while keeping one consequential piece for the article itself.
1. Strong social copy does not hide the topic
A reliable structure is:
- Show a topic the reader can immediately understand or relate to
- Add a reason to believe the article is worth reading—research, data, a comparison, or a concrete observation
- Leave one unresolved element: “one trait,” “one factor,” “an unexpected difference”
- Rule out one obvious guess
- Hand the reader to the URL
For example:
“People who connect quickly with strangers seem to share a common trait. Research found that one ability mattered in particular—and it was not simply being a great talker.”
The reader already knows the subject and why it may matter.
But one useful question remains: “Then what is it?”
That is the point.
The best copy often does not ask the reader a question explicitly. It causes the reader to ask the question internally.
2. Information gaps work—but total vagueness does not
Information-gap theories of curiosity propose that curiosity can arise when people become aware of a missing piece of knowledge and want to close that gap.
A 2024 study by Qiu and Golman analyzed more than 100,000 WeChat news articles. Reading was associated with headlines that made a question salient, appeared important, and involved more surprising topics. The same study also reported a small negative correlation between clicks and the likes-to-clicks ratio, suggesting that attracting curiosity and producing post-click satisfaction are not identical goals.
In practical terms:
“Now I want to know” can be a powerful entrance.
But:
“I clicked and got nothing” is still a bad experience.
A curiosity gap is not free attention. It is attention borrowed against a promise that the article must later repay.
3. Too vague and too complete can both lose
A registered Scientific Reports study analyzing 8,977 headline experiments found that concreteness did not have a simple linear relationship with click-through behavior.
When a baseline headline was too vague, making it more concrete could help.
When a headline was already highly concrete, adding even more information could reduce the benefit.
The useful lesson is an optimal middle level of concreteness.
A practical editorial rule is:
- Do not hide the topic
- Do not hide the general direction of the finding
- Show what kind of evidence exists
- Keep only one meaningful element unresolved
People sometimes summarize this as “show 80%, hold back 20%.” That can be a useful metaphor, but it is not an empirically established universal ratio.
4. One small expectation violation is enough
A strong hook often follows this pattern:
“You probably expect A. The result was not A.”
Research by Lewry and colleagues in 2023 found that minimally counterintuitive stimuli—items that violated one intuitive expectation—generated more curiosity than merely improbable stimuli in their experiments. Making the stimuli radically stranger did not produce endlessly increasing curiosity.
This was not a study of social headlines, so it should not be treated as a plug-in formula for CTR. But the editorial lesson is useful:
You do not need:
“EVERYTHING YOU KNOW IS A LIE.”
Often you only need:
“Most people assume speaking skill is the key. The research pointed somewhere else.”
One turn is enough. There is no prize for performing three midair twists and landing in a credibility crater.
5. Simple language beats impressive language
A 2024 Science Advances paper analyzed more than 30,000 field experiments run by The Washington Post and Upworthy. General readers tended to choose simpler, more readable headlines over more complex alternatives.
Compare:
“Predictors of social-bond formation in interpersonal interactions were identified.”
with:
“People who make friends quickly tend to share one trait.”
The first sounds like the reader has accidentally volunteered to peer-review a manuscript.
The second gives the brain an easy entry point.
The rule is not “remove expertise.”
It is make the doorway simple, then provide the technical detail inside.
6. A question mark is not a growth hack
The evidence here is nuanced.
The 2024 WeChat analysis found higher reading associated with question marks in headlines.
But a 2026 paper by Fang and Wheeler drew on more than 53,000 Reddit posts, roughly 3.08 million academic papers, 22,743 online-news A/B headline experiments, and a preregistered study. Across those settings, question-framed titles tended to reduce engagement, in part because they were perceived as less informative.
The practical synthesis is not “never use a question.”
It is: do not mechanically turn every title into one.
Instead of:
“Why do some people connect so quickly?”
try:
“People who connect quickly share a measurable pattern.”
Now the sentence gives information while still causing the reader to wonder what the pattern is.
Put the question in the reader’s head, not necessarily in the punctuation.
7. Seven hook structures worth rotating
If every post says “Scientists discovered a surprising factor,” the account starts to sound like an automated vending machine.
Rotate structures.
1. Missing-answer hook
“People who do X more successfully tend to share one behavior.”
Hide only the behavior.
2. Expectation-reversal hook
“A is usually treated as the key to X. The study found the important difference somewhere else.”
Disconfirm one obvious guess.
3. Result-first hook
“Doing A increased B. But the reason was not what researchers expected.”
Give the result; hold back the mechanism.
4. Comparison hook
“When A and B were compared, a clear difference appeared. The deciding factor was not the usual one.”
Useful for product, behavior, or condition comparisons.
5. Failure-pattern hook
“Cases that go badly often share the same overlooked pattern.”
Useful for troubleshooting content.
6. Scale hook
“More than 30,000 real-world tests revealed a surprisingly simple pattern in which headlines people choose.”
Sometimes the evidence size is itself the hook.
7. Small-mystery hook
“Two sentences can say nearly the same thing, yet one gets chosen more often. The difference is not an advanced copywriting trick.”
Start with a concrete puzzle.
8. The worst mistake is finishing the article inside the post
Automation often produces:
“New article.” “Title.” “Summary.” “Conclusion.” “URL.”
The URL is present, but the post has already performed the article’s job.
The reader has the useful answer and keeps scrolling.
Social copy is not a miniature article.
Its job is to create a justified reason to enter the article.
There is one important exception: high-stakes information.
For medical, legal, financial, safety, disaster, or similar topics, do not hide warnings or decision-critical facts as bait. When withholding information could change a consequential decision, safety outranks suspense.
9. The line between a trailer and clickbait is payoff
The difference is not how dramatic the wording sounds.
It is whether the article delivers what the post promised.
Bad copy:
- Implies a conclusion the article does not support
- Turns correlation into causation
- Inflates numbers
- Adds fake authority such as “experts are stunned”
- Promises an answer that never appears
Good copy:
- Opens only questions the article can actually resolve
- Identifies whether the evidence is research, observation, comparison, or experience
- Preserves uncertainty where the source is uncertain
- Never changes a fact merely to make the hook more surprising
- Produces a satisfying “Ah, that’s what it was” after the click
Clicks can be borrowed.
Trust has a credit limit.
10. If you automate it, do not optimize only for CTR
An automated distribution system should not optimize solely for clicks.
Track a bundle of outcomes:
- Click-through rate
- Reading depth or dwell time
- Movement to another article
- Saves and shares
- Return visits
- Negative feedback
- Repetition of the same hook pattern
The 2024 curiosity study is a useful warning: more curiosity-driven clicks do not automatically mean stronger post-click approval.
The real sequence you want is:
“Clicked. Read. The promise matched. Came back.”
not merely:
“Clicked. Victory.”
11. A practical generation policy for AI-written social copy
A reliable generator can follow this sequence:
- Extract the article’s actual reader value in one sentence
- Identify the evidence type
- Estimate the reader’s most obvious first guess
- Select one unresolved element that the article truly answers
- Withhold only that element
- Rewrite in plain language
- Rotate the hook structure so posts do not all sound the same
- Add the URL
- Compare the copy with the article and reject exaggeration, fabrication, or unresolved promises
Useful prohibitions:
- Do not begin every post with “New article”
- Do not paste the full summary
- Do not spam “shocking,” “revealed,” or “actually”
- Do not invent numbers
- Do not invent conflicts that are absent from the source
- Do not hide critical warnings for dramatic effect
- Do not translate all languages word-for-word from one source syntax
For multilingual publishing, translate the function, not the sentence shape.
Preserve:
“This line establishes relevance.” “This line establishes evidence.” “This line leaves one gap.” “This line gently violates the obvious expectation.”
Then rebuild those functions naturally in each language.
12. The reusable formula
The most useful default is:
Value
→ Evidence
→ One information gap
→ One small surprise
→ URL
Do not tell everything.
Do not tell nothing.
Social copy is not a compressed version of the article.
It is the trailer that gives the reader a reason to enter.
Do not reveal the killer in the trailer.
But at least tell people what movie they are watching.
References
- Qiu, J. & Golman, R. (2024). Curiosity in news consumption. Applied Cognitive Psychology. https://doi.org/10.1002/acp.4195
- When curiosity gaps backfire: effects of headline concreteness on information selection decisions. Scientific Reports. https://doi.org/10.1038/s41598-024-81575-9
- Shulman, H. C., Markowitz, D. M., & Rogers, T. (2024). Reading dies in complexity: Online news consumers prefer simple writing. Science Advances. https://doi.org/10.1126/sciadv.adn2555
- Fang, D. & Wheeler, S. C. (2026). Titles framed as questions reduce reader engagement. Journal of Consumer Psychology. https://doi.org/10.1002/jcpy.70031
- Lewry, C., Gorucu, S., Liquin, E. G., & Lombrozo, T. (2023). Minimally counterintuitive stimuli trigger greater curiosity than merely improbable stimuli. Cognition, 230, 105286. https://doi.org/10.1016/j.cognition.2022.105286
- Blom, J. N. & Hansen, K. R. (2015). Click bait: Forward-reference as lure in online news headlines. Discourse, Context & Media. https://www.sciencedirect.com/science/article/pii/S0378216614002410
