How To Write Headlines That Get Clicks

A headline has one job before it has any other job: it has to make someone stop scrolling. Everything else — SEO value, shareability, brand voice — only matters if the headline clears that first bar. Most headlines fail not because they're badly written but because they try to summarize the article instead of creating a reason to open it.

The gap between curiosity and clickbait

The mechanism behind almost every headline that works is what advertising writers have long called the curiosity gap — the reader knows just enough to be interested and not enough to be satisfied. "5 Mistakes New Runners Make" works because it implies you might be one of the five without confirming it.

The difference between a curiosity gap and clickbait is whether the article actually closes the gap. A headline that promises an answer and delivers one is doing its job. A headline that promises an answer and delivers vague filler is clickbait, and readers remember which sites do that to them. It costs you the second click far more than it earns you the first.

What actually makes a headline specific

Vague headlines ask readers to trust that the content is good. Specific ones show it. "How to Write Better" asks for trust. "How to Cut Your First Draft's Word Count by a Third Without Losing the Point" shows a reader exactly what they're getting and why it's worth their time.

Specificity usually comes from one of a few places:

• A number — but only when the number is real and the list underneath actually has that many distinct, substantial points. Padding five ideas out to ten hollow ones is more damaging than just writing "5."
• A named outcome — what changes for the reader, stated in concrete terms rather than in something like "success" or "results."
• A named audience — "for beginners," "for freelancers," "if you're switching from X" — narrowing the headline often increases relevance more than it narrows the audience.
• A contrast or tension — "why the advice you've heard about X is backwards" works because it sets up a disagreement the reader wants resolved.

Length is a constraint, not a rule

There's no fixed ideal character count that applies everywhere — it depends on where the headline is being read. A headline in a Google search result gets truncated at a certain pixel width, not a fixed character count, so previewing how it actually renders matters more than hitting an arbitrary number. On social platforms or in an email subject line, the constraints are different again. The real rule is simpler: cut every word that doesn't earn its place, and stop once the headline says exactly one clear thing.

Testing without fooling yourself

Formal A/B testing needs meaningful traffic volume to produce a result you can trust — a small blog running two headline variants for a day and picking whichever got three more clicks isn't learning anything reliable; that's just noise. If you don't have the traffic for a real test, the more useful habit is writing several genuinely different headline angles before publishing (not just synonym swaps) and picking the one that would make you personally click, read cold, with no context.

How headlines change across the places people see them

The same article often needs more than one headline, because the context changes what the reader already knows and what's competing for their attention. A search-result title is competing against nine other blue links answering the same query, so specificity and a clear match to the search intent tend to matter more than novelty. A social media headline is competing against posts from friends and unrelated content in a feed, so it has more room to lean on curiosity or a stronger emotional hook. An email subject line is read alongside dozens of other subject lines in an inbox, often on a small screen, where the first few words carry disproportionate weight because the rest may get cut off. Treating all three as the same problem — writing one headline and reusing it everywhere — usually means it's optimized for none of them particularly well.

Power words work only when they're specific to the claim

"Power words" — proven, ultimate, essential, incredible — get recommended constantly in headline-writing advice, but they only work when they're attached to something specific enough to back them up. "The Ultimate Guide to Running" says nothing a reader can evaluate. "The Only 3 Runs You Need in Marathon Training" makes a specific, checkable claim, and the word "only" is doing real work because it's tied to a number and a claim the article can actually defend. Used as decoration rather than as part of a real claim, these words tend to blur together across headlines and stop registering as meaningful at all — readers develop a kind of blindness to generic superlatives the same way they tune out banner ads.

Frequently asked questions

Does a headline need to include the exact keyword someone would search for? For search-driven content, including the core term a reader would actually type in is still one of the more reliable ways to signal relevance, both to the reader scanning results and to how the page gets matched to queries. It doesn't need to be the literal first words of the headline, and forcing an awkward exact-match phrase in usually reads worse than working the same idea in naturally.

Is it ever fine to write a headline before the article is finished? Drafting a working headline early can help focus the piece, but the version that actually gets published is worth revisiting after the article is done — it's common for the article to end up making a slightly different point than planned, and a headline written before that shift will describe a piece that no longer quite exists.

Do question headlines perform differently than statement headlines? Question headlines can work well when the question is one the reader has actually asked themselves — they create instant relevance. They tend to fall flat when the question is generic enough that almost any headline could be rephrased as one ("Want to Improve Your Writing?"), because the specificity that makes a headline work is missing regardless of the sentence structure around it.

How many headline variations should I actually write before picking one? There's no fixed number, but writing only one and moving on skips the part where most of the improvement happens. Writing five to ten genuinely different angles — not minor word swaps — before picking usually surfaces at least one clearer or more specific option than whatever came to mind first.

Does the same headline advice apply to podcast episode titles and video titles? Largely yes, with one added wrinkle: those formats are also competing inside a platform's own recommendation and search system, not just in front of a human reader deciding whether to click. Specificity and an honest, checkable promise still matter most, but it's worth also checking how the title behaves as plain text without any visual thumbnail support, since it often has to work harder on its own than a headline sitting next to article body text.

The practical takeaway

A good headline is a specific, honest promise about what's inside, pitched at a reader who knows nothing about the piece yet. Chase specificity and a real curiosity gap before you chase length rules or power words, adapt the headline to where it's actually being read, and never let it promise something the article doesn't deliver — that trade might win you one click, but it loses you the reader.