If you own a blog, sooner or later you end up staring at a title that feels a little off and wondering how to fix it. This guide walks you through how to write a headline in eight steps, so the searcher who needs your article can tell at a glance that it's the right one. By the end you'll have a repeatable way to draft, check, publish, and measure a title, and you'll know which small mistakes tend to cost you clicks without anyone noticing.
One thing to say up front is that no wording guarantees clicks or rankings. A good headline helps the right person recognize your page, and that's a job worth doing well. You'll need a target query for the article, a rough idea of who it's for, and access to Search Console if the page is already live.
Step 1: Choose one search task for the headline
Before you write any words, write down one line that covers the query, the reader, and the job they're trying to get done. For this article it might read: "A marketing lead searching for a repeatable method to write blog titles that earn relevant clicks." That single sentence is your headline brief.
If you're improving a page that already exists, open the Search results report in Search Console and look at the Queries and Pages tabs. They show which searches really bring up that page, and that's often different from the ones you had in mind when you wrote it. Pick the task the article actually fulfills, and don't try to fit every related phrase into one title.
You're done with this step when a colleague can read your line and name one reader, one task, and one promised answer without opening the article.
Where people go wrong here is treating a keyword list like a headline brief. Something like "headline formulas, SEO titles, CTR, AI, content marketing tips" is a topic inventory, and it doesn't tell a reader anything. A page that gets impressions for several queries usually needs one carefully chosen main promise, not all of them squeezed into the title.
Step 2: Check the results and what your article actually delivers
Search your target query and look at the results the way a reader would. Notice what the competing titles promise and what kind of answer seems to be wanted: a step-by-step guide, a list of examples, a definition, or a comparison. That tells you a lot about search intent, which is simply the job the searcher is trying to finish. If you want a deeper look at how intent changes what a page should do, our piece on SEO content for different search intent types covers it.
Then take an honest inventory of what your article can really deliver. It might be a process, a set of examples, a few checks, or a list of mistakes and fixes. Look for a specific, useful angle that's true of your page, and don't copy a rival's phrasing.
You're done when your planned headline names a format and a benefit that the article demonstrably supplies.
Where people go wrong is promising a checklist when the article only has broad advice, adding a year that has no real meaning for the content, or calling something "ultimate" when it isn't. Google's guidance on helpful content asks whether a title or main heading summarizes the page accurately and avoids exaggeration, so this is a good habit as well as a safe one.
If your team keeps its products, audiences, voice, and brand rules in one place, that helps at this step. Deep IQ stores that kind of brand context as structured data, which keeps the audience and product claims in a headline grounded in the brand. It doesn't replace checking what a real searcher wants, and it doesn't validate a headline against the finished article. That part stays with you. Our guide to grounding content in brand voice and product facts goes deeper on the stored-context side.
Step 3: Write a plain-language working headline
Start with a subject people will recognize and add the specific action or benefit. "How to Write Blog Headlines That Match Search Intent" tells a reader far more than "Better Titles for Better Results." Use the words a reader would type to find the subject, and put them early. Google recommends putting familiar words prominently in the page title and the main heading, and the opening words matter most because a narrow display can cut off the rest.
Nielsen Norman Group has long advised that headlines should make sense out of context, specify useful information, use familiar words, and put the informative terms first. Their advice was published in 2015, and it still reads as good plain sense, because a headline often shows up in a list, a feed, or a search result with nothing around it.
You're done when someone who sees only the title can tell what the page covers, who it helps if that matters, and why it's worth opening.
Where people go wrong is leading with a long brand prefix, an unexplained slogan, or an abstract benefit before the topic is named. If the first three or four words don't say what the page is about, move things around until they do.
Step 4: Draft variants with headline formulas, then keep only the truthful ones
Now write a few alternatives, and let each one take a different clear angle. If you're wondering how to write clickable headlines, this is the step where most of that work happens, and it goes better when you keep the options honest instead of being a keyword permutation of the last. The headline formulas below are drafting patterns you can adapt. They aren't proven to raise CTR, and each has a check you should run before you use it.
| Pattern | Illustrative headline | Check before using |
|---|---|---|
| How to + task + relevant outcome | How to Write Blog Headlines That Match Search Intent | Does the page teach the full task? |
| Task + constraint or audience | Blog Title Writing Tips for Small Marketing Teams | Is that audience really addressed? |
| Number + specific category | 7 Headline Mistakes That Make Blog Posts Harder to Click | Are there exactly seven substantive mistakes? |
| Specific question + direct answer | What Makes a Blog Headline Worth Clicking? | Does the page answer the question directly? |
| Process + differentiating detail | Write Blog Headlines in 8 Steps, From Query to CTR Review | Are all eight steps present and in order? |
| Outcome + method | Match Search Intent With a Clearer Blog Title | Is the method explained, not just promised? |
These sample headlines are illustrations for drafting. They aren't measured winners, and they don't stand for real articles.
You're done when you have a few workable headlines with clearly different promises, and you can point to a section or example in the article that backs up each one.
Where people go wrong is adding numbers, superlatives, questions, or emotional words just because a formula says they work. If curiosity-gap wording tempts you, our piece on whether clickbait drives sustainable traffic covers that separately. A Backlinko analysis of Google results found no significant difference in organic CTR between question titles and non-question titles in its dataset. So let the reader's task pick the formula, and don't lean on folklore. If a numbered format fits, our guide on how to write a listicle that earns AI citations has useful notes on making the list honest.
Step 5: Make the search title and the on-page heading agree
There are two things to line up here. The on-page headline is the title people read on your article, usually the H1. The HTML title element is the title stored in the page's metadata, and Google reads it as an input when it builds the clickable line in search results. That clickable line is called the title link, and Google generates it automatically.
Pick an H1 that reads naturally on the page and an HTML title that works in a search result. They don't have to match character for character, but they should describe the same topic and make the same promise. Use your main search phrase once where it fits naturally, and take out repeated variants. If your brand name appears, keep it short. Also check that the page has one clear, visually prominent main title, ideally the first H1 a visitor sees.
You're done when the H1, the HTML title, and the article all describe the same answer, and each page on your site has its own title element.
Where people go wrong is repeating "how to write a headline" and "how to write clickable headlines" in the same title, adding the same long tagline to every post, or showing several headings that all look like the main one. Google may build a different title link if the title is inaccurate, repetitive, or a poor description of the page.
DeepSmith's content pipeline builds brand context, SEO structure, and publishing metadata into each article as it's produced. That gives you a consistent starting point, but the editorial check is still yours to make: read the title and the finished article side by side and confirm they make the same promise. DeepSmith doesn't have a headline scoring tool or a headline A/B test, and nothing in it promises a CTR gain.
Step 6: Run a five-question editorial check
Before a headline goes live, put it through these five questions, which double as blog title writing tips you can reuse on every post. This is an editorial checklist, and it isn't a search engine score or a tested threshold.
- Can a reader name the subject from the first few words?
- Is the outcome or useful information specific?
- Does the wording match what the page actually delivers?
- Is there one natural primary phrase, rather than a pile of variants?
- Would the title still make sense outside its original context and on a smaller screen?
When something fails, edit for familiar language and cut filler before you shorten anything essential. For a full line-level pass on the article itself, use our AI draft editing checklist. A title like "Content That Works: Everything You Need to Know About Writing the Very Best Blog Headlines" buries its subject. "How to Write Blog Headlines That Match Search Intent" is clearer, and it's an illustration of clarity, not a tested performance claim.
You're done when every answer is yes and nothing in the title relies on an invented result, feature, statistic, or date.
Where people go wrong is trimming a title to hit a character count and deleting the words that made it useful. Google sets no fixed length limit for the HTML title element, and the title link it shows can be shortened to fit the device. Preview how your title reads, especially on mobile, but don't treat 60 characters as a Google rule.
Common mistake: A page with lots of impressions and a low CTR doesn't automatically have a bad headline. Check the query, the page, the title Google displayed, and the position before you change anything. The page may be showing up for a task it doesn't fully answer.
That's the main reason to pay attention to the other headline mistakes to avoid, which are gathered in a table further down.
Step 7: Publish, then check what searchers actually see
Once the article is live, confirm the rendered H1 and the saved HTML title, and don't rely on the draft field alone. After Google has processed the page, search for it or for a relevant query and note the title link it displays. If that differs from what you wrote, look at your page title, your main heading, any repeated branding, the accuracy of the wording, and the language you used. Google can draw on more than the title element. A change to your title can take a few days to a few weeks to show up, because Google has to recrawl and reprocess the page.
You're done when the published page has the title metadata and on-page heading you meant to publish, and any difference in the displayed title has been written down and not assumed away.
Where people go wrong is assuming the CMS title is the same as the Google title link, or deciding a revision failed the day after it went live. Google also doesn't offer a switch to force a particular title link for one page, so the best you can do is make the page's signals clear and consistent.
DeepSmith has a review and publish step in Produced Content where you can edit the article body, title, and slug before it goes to your CMS. Use that moment to read the final promise one more time. It doesn't control what Google displays, and it shouldn't be treated as if it does.
Step 8: Measure relevant clicks and revise one promise at a time
In Search Console, open the Search results report and turn on CTR, which is clicks divided by impressions. Look at the Pages tab and focus on pages that have enough impressions to be worth your time. Filter to one page, look at its Queries, and compare similar date ranges. Read clicks, impressions, CTR, and average position together, and split by device when it matters. To go the other way, pick a query in the Queries tab and check which URLs sit behind it in Pages.
Keep a simple record for each change: the old title, the new title, the day you published it, and the day the changed version showed up in search, if you can see that. As an illustration of the arithmetic, 120 clicks from 4,000 impressions is a 3% CTR. That example is only arithmetic, and it isn't a benchmark for a good CTR.
You're done when your team can say which query and page you changed, what the new promise was, and how clicks and impressions behaved alongside position.
Where people go wrong is treating a sitewide CTR change as proof that one title worked. CTR moves with query mix, position, device, how the result is presented, and the time period. Average position in Search Console is an aggregate across many searches, so it isn't a fixed ranking. A before and after comparison can guide an editorial decision, but it isn't a controlled A/B test, so avoid setting a universal CTR target or a fixed number of days to prove success. If you're looking into a steady ranking with falling traffic, our walkthrough on diagnosing a CTR-driven traffic drop shows how to separate title problems from everything else. The SERP itself also matters, and how the modern SERP crowds out organic clicks explains why a good title can still see fewer clicks. If AI Overviews are part of your worry, we also look at the CTR data on AI Overviews.
Headline mistakes to avoid, and the fix for each
Most of the mistakes below are quiet, which is why they last. Nobody complains about a vague title. It just gets fewer clicks than it might have, and you rarely find out why. Use this table as a last pass on any headline.
| Quiet click-killer | What a reader sees | Fix |
|---|---|---|
| Vague subject | "Ideas to Improve Your Content" | Name the actual task: writing blog headlines. |
| Format mismatch | "Checklist" for a page with no checklist | Match the label to the deliverable. |
| Keyword repetition | Several near-identical versions of one phrase | Use the clearest phrase once and write naturally. |
| Unsubstantiated number or claim | "Proven 3x More Clicks" without evidence | Remove the performance claim and use a benefit you can verify. |
| Generic or repeated title | Every page ends in the same long descriptor | Give each page a distinctive, descriptive title and shorten the branding. |
| Important words buried | The topic follows a long preamble | Move the subject and task to the front. |
| Search and page promise diverge | Search title promises examples, page provides definitions | Revise the title or supply the promised material. |
| Conflicting prominent headings | Several headings look like the main title | Make the main page title visually clear. |
| False freshness cue | A year appears without a meaningful update | Include a date only when the content earns it. |
| Mechanical length editing | An informative title is cut down to a vague one | Keep the meaning and check how the title renders. |
| Declaring a winner from aggregate CTR | A changed query mix goes unexamined | Look at the same page and relevant queries, with position and device in view. |
What the research says, and what it doesn't
A few studies come up whenever people talk about headlines, and they're worth using with care.
Google says title links are often the main information people use to decide which result to click. Its guidance asks for a title element on every page, with concise, descriptive wording, distinct titles across pages, no keyword stuffing, and a clear main title on the page. It doesn't say that editing a title alone will raise rank or CTR.
BuzzSumo looked at 100 million headlines published from January 1, 2019 through December 31, 2020, and measured social sharing on Facebook and Twitter. It reported 11 words and 65 characters as an effective length in that social dataset. That's about shares on social platforms, so don't carry those figures over to Google Search as an ideal title length. The study itself notes that past sharing doesn't establish what will work next.
Backlinko's analysis, updated April 16, 2025, covers 4 million Google search results, including CTR data across 1,312,881 pages and 12,166,560 queries. It reports higher observed CTR for titles in a 40 to 60 character range and no significant difference between question and non-question titles. Those are observed associations. They aren't a Google character rule, and they don't prove that resizing a particular title will lift its CTR.
Google's own guidance for generative AI search says you don't need to rewrite in a special style for AI systems. There's no verified headline formula that guarantees an AI citation, so a clear, useful title on a page that delivers on it is the sensible approach. If AI answers are on your mind, our article on how to write a blog post AI answer engines cite picks up where headlines leave off.
Keep clicks, rankings, social shares, and AI citations separate in your head. They're different measures, and a headline that does well on one says little about the others.
What to do next
Pick one live article that already gets impressions, and walk it through steps 1, 5, and 8 this week. Write the one-line brief, compare the H1 to the HTML title, and note the current CTR and average position for its main query. If the headline needs work, change one promise at a time and record the date. Small, well-documented changes will teach you more about your own readers than any formula list. It also helps to have your voice and product facts written down once, and our guide on documenting brand voice and tone shows how.
If you'd like a platform that produces publish-ready articles with brand context and metadata already built in, you can start a DeepSmith free trial and see how the pipeline handles a real piece of yours.
