Most readers only see your title, so it has to carry the claim

For nearly everyone who meets a post, the title is the post. It turns up in a search result, a feed, a newsletter digest or a line inside an AI answer, and most people read it, form an impression and move on. A headline written to make them curious hands that majority nothing, and leaves them with whatever they guessed. The title has to make the argument on its own, and the body exists for the minority who want the reasons.

The click is the exception, not the rule

The numbers here are hard to measure honestly, so we will lean on one study that measured them properly. In 2016 researchers from Columbia University and Inria published Social Clicks: What and Who Gets Read on Twitter?, which tracked a month of links to five major news sites. Their dataset held 2.8 million shares, responsible for an estimated 75 billion potential views, and 9.6 million actual clicks. Potential views are not real impressions, and Twitter a decade ago is not your search result today. But the shape of the finding is the one every publisher recognises: links get seen far more often than they get opened.

AI answers push the same way. We argued in arriving from an AI summary that the reader who clicks through has usually been given the answer already. The reader who does not click has been given something too, and quite often it is your title, cited as a source next to a sentence the model wrote. If that title is a riddle, you have contributed a riddle to the one place your brand appeared.

A teaser asks the reader to do the work

The curiosity-gap headline works by withholding. "The one mistake that kills most company blogs" promises an answer and keeps it behind the click. For the reader who clicks, that is a mild annoyance. For the reader who does not, it is worse than nothing, because they leave with the impression that something is wrong with company blogs and no idea what.

A claim title does the opposite. "Why in-house blogging quietly stalls at month four" tells the non-clicking reader something they can keep and repeat. It can be wrong, and that is the point: a claim can be checked, disputed, and remembered. A teaser cannot be any of those things, because it has not said anything yet.

Google asks the same question in its own words. Its guidance on helpful, people-first content includes two self-assessment questions about exactly this: whether the main heading or page title gives a descriptive, helpful summary of the content, and whether it avoids exaggerating or being shocking. A teaser fails the first question by design.

Google will rewrite a title it does not trust

There is a mechanical reason as well. Google's title link documentation says the title shown in results is generated automatically, and that the HTML title element is only one of several sources. It may use the main visual heading, other prominent text on the page, or even the anchor text of links pointing at the page. The documentation lists the cases where it adjusts, among them titles that are obsolete, titles that do not accurately reflect the page, and pages with several headings of equal prominence.

So a headline that says less than the page does is not guaranteed to survive. The most reliable way to keep your own words in front of the reader is to write a title the system has no reason to improve on: descriptive, specific, and consistent with what the page actually argues.

What a claim title costs

We should concede the trade. A title that states the conclusion gives some readers everything they wanted, and they will not click. If your only measure is click-through rate, a teaser will sometimes win, and a few subjects genuinely resist compression into one sentence.

But the reader who reads the claim and does not click has still learned what you think, with your name next to it. Repeated across a year of posts, that is how a company becomes associated with a position, which is the commercial point of the blog in the first place. The reader who does click arrives knowing what they are about to read, which is a better start than arriving to find out whether the bait was worth it.

This matters more for an automated blog than for a hand-written one. A model asked for a catchy headline will reach for the patterns that dominate the material it learned from, and the curiosity gap is everywhere. Our own rule is the one in our cornerstone post, whose title is its entire argument: if a reader saw only the headline, they should still know what we believe.

Write the title for the people who stop there

The practical test takes ten seconds. Read the title alone and ask what someone who never clicks would now believe. If the answer is "nothing yet", it is a teaser. If the answer is something you would be happy to have repeated with your name attached, it is a title.

Write the post for the reader who clicks. Write the title for everyone else, because there are far more of them.