Free headline generator for a stronger first draft.

Take Cobb Lane Physio, a two therapist clinic that has built a specific programme for runners with recurring knee pain who are training for a first marathon. The page needing a headline is the programme page, the reader arrives from a search or a local recommendation, and the action is booking a first assessment.

The shortlist, after writing far more than three: "Knee pain at mile eight? Get to the start line without stopping training." Then "A six week plan for runners whose knees give out past mile eight." Then "Marathon training with knee pain: what to change before you rest it completely." The one that goes on the page is the second, with a subheading reading "Assessment, a written plan and two follow ups, built around the training block you are already in. First appointment forty minutes." The title tag for search is written separately as "Knee pain when running: six week rehab for marathon training", which lands at fifty eight characters.

The chosen headline works because every part of it is checkable. Six weeks is a duration the page can honour. Runners is the audience, stated so plainly that a swimmer knows to leave. Knees give out past mile eight is the reader's own phrasing rather than clinical language, and it describes a specific moment they have lived through rather than a category of injury. Nothing in it promises a cure, which matters because the page cannot promise one. The subheading then does the job the headline is too short for: it says what is included and how long the first appointment takes, which is the information that actually decides a booking.

The version the clinic wrote first: "Cobb Lane Physio: expert physiotherapy for all your needs. Our experienced team is dedicated to helping you move better and live pain free."

That headline is a label. It names the business and the category, which is only useful to someone who already knew both and was looking for the address. All your needs describes every physiotherapist in the country and therefore distinguishes none of them. Expert and experienced are claims a reader has no way to evaluate in the second they spend on a headline, so they read as filler. Live pain free is the opposite failure, an outcome the page cannot guarantee and the clinician would never say in the room, so a reader who books on the strength of it arrives with the wrong expectation. And a runner scanning results for their exact problem sees nothing about themselves in it, so they never reach the good clinical writing further down the page.

If the page underneath cannot deliver on any honest headline you could write, the headline is not the problem. A strong headline over a thin page produces a fast bounce and a reader who now knows not to trust you, which is worse than the weaker headline that would have set the right expectation. Write the page first, or at least its first section, then write the headline from what is actually there.

If the offer itself is vague, no arrangement of words will fix it. When you cannot say who the page is for, what they get, and roughly what it costs or how long it takes, the headline options will all be abstractions because the underlying thing is one. That is a decision to make in a conversation with whoever runs the business, not a drafting exercise. The same applies when three different audiences are being served by one page: the honest fix is usually three pages, and trying to cover all of them in one line produces the bland compromise that serves none.

If you write for a publication, follow its style guide rather than a generator. Editorial headlines are governed by house conventions on capitalisation, tense, length, whether questions are permitted and how numbers are rendered, and a sub editor will rewrite anything that ignores them. Legal and regulatory constraints also outrank craft advice: headline claims in health, in finance, and in anything comparing you to a named competitor sit inside rules that a general drafting tool knows nothing about.

And if the plan is to test two headlines against each other, check you have the traffic for it. Comparing conversion rates on a page with a handful of visitors a week will tell you nothing for months, and acting on an early difference is worse than not testing at all. On low traffic pages, judge headlines by reading them aloud beside the competitors that appear in a search results page, and keep the one a stranger could repeat back to you correctly.

Different placements truncate at different points and none of them count words. A page headline is limited by the line break rather than a character count: at a typical hero size, anything past about twelve words wraps to a third line on a phone and loses its shape, so six to twelve words is the usable range. A title tag in search results is cut on pixel width rather than characters, which is why the same character count survives in one headline and truncates in another, and why roughly fifty to sixty characters is a practical target rather than a rule.

The consequence is the same in both cases: the important part goes first. Front loading is not a stylistic preference, it is protection against a cut you cannot predict. Test a headline by reading only its first five words and asking whether they still say who this is for. Then read it at the size it will actually appear, in the real typeface, on a phone. A headline set in a light weight at a large size behaves differently from the same words in a search result, and a version that looks balanced in a document often breaks awkwardly in the layout.

Three readers judge a headline and they want different things. The scanner on a search results page is comparing yours with nine alternatives in a couple of seconds, so it is judged on relative specificity: whichever one most obviously matches the words they typed wins, and cleverness reads as noise there. The reader who has already clicked wants confirmation they landed correctly, so their headline is judged on continuity with whatever they clicked. The colleague or client approving it is judging something else entirely, usually whether it sounds impressive enough, and that is the pressure that turns specific headlines into vague ones.

The defence against that third reader is evidence rather than argument. Put the proposed headline beside the actual search results it will sit among, and beside the first paragraph of the page. If it is the least specific option on the screen it loses, however good it sounds read aloud in a meeting. The question to settle is not whether the headline is good writing, but whether a stranger who read only that line could correctly describe what the page offers.

A page headline, a title tag, an email subject line and a URL slug are separate artefacts, and people routinely try to satisfy all four with one string. The page headline is written for someone who has already arrived and needs their click confirmed, so it can be a full sentence and can assume context. The title tag is written for someone comparing you with competitors before clicking, so it front loads the searched phrasing and lives with a pixel width limit. The subject line is written for someone deciding whether to open, in an inbox where the sender name is already doing half the persuasion.

They can share words, and often should, since a violent mismatch between a title tag and a page headline is disorienting. But optimising one string for all of them produces a headline stuffed with search phrases and no rhythm. Write the page headline first, then write the title tag as a deliberate variant with the searched phrasing in front, then keep the slug short, lower case and hyphenated, carrying the two or three words that matter and nothing else. Once published, leave the slug alone: changing it later means redirects and lost links for very little gain.

Volume is the technique, but the sorting afterwards is where the work happens. Read all of them aloud and delete anything you stumble over, since a headline that trips the writer will trip the reader. Then delete every one that would be equally true of a competitor, which usually removes half the list at a stroke. Then check each survivor against the first section of the page and delete any the page cannot immediately keep.

What remains is normally two or three, and they tend to differ by which fact they lead with rather than by phrasing. Choose by what the reader is most likely to have been searching for, not by which sounds best to you. Keep the runners up in a note beside the page, because they are useful later as email subject lines, section headings, ad variants and social openers, and rewriting them from scratch in three months is wasted effort. Finally, set a reminder to reread the headline once the page has been live for a while. Offers drift, prices change and audiences narrow, and a headline that was accurate at launch quietly becomes a promise the page no longer keeps.