Making Headlines for Google News

An article headline may seem a singular thing, but in the context of websites and SEO there are actually at least four different headlines for every article.

What is a headline? Seems like a silly question. It’s the sentence at the top of an article, right? It’s what announces what the article will be about.

Correct, of course. Yet there’s more to it than that. Headlines serve multiple purposes, and in the context of SEO there are actually four distinct types of headline that each have a different purpose for the article’s overall visibility in organic search.

So let’s talk about these headlines.

1. Visible Headline

The first and most obvious headline that every news article on the web has is, of course, the visible headline at the top:

The headline serves an important purpose: it announces the topic of the article.

If you’re writing purely for human readers only, you can have a lot of fun with headlines. Puns, word jokes, cryptic phrases, you name it - as long as it grabs the reader’s attention and makes sense in the context of the news website, you can do whatever you want with headlines.

Search engines, however, need a bit more. It’s easy to forget that Google is still a machine. The technology that powers search engines continues to improve, and at times it seems that Google can understand human language perfectly.

But that’s not entirely the case. Context is something that is almost intuitively grasped by most humans, but it’s something machine systems still struggle with.

So, to ensure both humans and machines can understand our article headlines, we need to make them clear and unambiguous. This means we have to forego those puns and clever jokes and instead write our article headlines to be descriptive and meaningful.

The headline I’ve shown above is a good, well-written example. It contains all the right phrases (‘Autumn Nations Cup’, ‘James Ryan’, ‘Ireland’) so that it’s clear what the article will be about.

(As a sidenote, I’m an Irish rugby fan so it pains me somewhat to use this example, but it’s a good one so I’ll suffer through this and stick with it.)

Visible headlines on an article can be as long and intricate as you want. The only real limits are what your editor (or, sometimes, the website’s article template design) lets you get away with and what makes sense for your readers.

Other types of headlines, however, have some stricter limitations.

2. HTML <title> Tag

The second type of headline is all too familiar to SEOs: the <title> tag.

A webpage’s <title> tag is a HTML tag in the page’s source code which contains the webpage’s headline. Google uses this title in the same way as humans use article headlines: to determine what the content of the page is about.

<title>Autumn Nations Cup: James Ryan urges patience among Ireland fans after defeats - BBC Sport</title>

Title tags are one of the most important on-page SEO elements, and SEOs spend a lot of effort ensuring every page has a proper well-optimised title tag.

For news articles, the <title> often mimics the visible article headline. This is important, as they both serve the same purpose: indicating what the page content is about. While the visible headline and the <title> don’t have to be identical, they should more or less align to convey a similar meaning and contain the same keywords and phrases.

If your visible headline is very different fron your <title> tag, it can confuse search engines about the contents of the page.

For example, the <title> of the article I’ve been using as an example could be something like:

<title>Andy Farrell praises Ireland squad after gritty defeat to England - BBC Sport</title>

This title would contrast with the article headline to such an extent that it would confuse Google about the article’s actual contents. Is it about James Ryan or about Andy Farrell? If someone searches for ‘james ryan’ in Google, should we show this article? Or should we show it when someone searches for ‘andy farrell’?

Confusion is bad for search engines. In fact, I believe most good SEO is about removing confusion and contradicting signals for Google. It’s about simplification, clarity, and unambiguity.

So make sure your <title> tag is clear and unambiguous and aligns with your article headline.

While visible headlines don’t really have a length limitation (within reason, of course), <title> tags are somewhat more restricted. Generally, it’s recommended to limit <title> tags to around 60 to 65 characters. This is primarily because Google uses <title> tags as the webpage’s heading when the page is shown in its search results, and that heading has a visual limit:

Here we see that this article’s <title>, at 90 characters, surpasses that limit. As a result, it gets shortened by Google with ‘. . .’ and some of the title contents is omitted.

In this instance it doesn’t have too big of an impact, as the general gist of the headline is still clear for people scrolling through Google’s results. Sometimes, however, a title that is shortened by Google loses a lot of its meaning, which is why it’s generally recommended to try to stick to that 65-character visual limit.

(Note: the 65 character visual limit is actually a 512 pixels width limit. The actual number of characters depends on which characters you use. More ‘w’ means fewer characters, more ‘i’ and ‘l’ means more characters to play with. It’s also why using vertical pipes ‘|’ as separator characters is usually better than horizontal bars ‘-’.)

Now, that 65 character visual limit is mainly valuable if you care about how the webpage shows up in regular Google search results. And, as we discussed last time, this is not necessarily where most news publishers will get much organic traffic from.

Instead, the main source of search traffic to publishers is the Top Stories box. There we play by different rules and, you guessed it, Top Stories uses a different headline entirely.

3. NewsArticle Structured Data Headline

Most news websites have extra code in the HTML source of every article. This code is called structured data, and is intended to give search engines like Google more information about the contents of the webpage.

There are many different types of structured data. For news articles, the recommended type to implement is Article or NewsArticle, and it looks a little bit like this:

The purpose of structured data is to make the page easier to understand for Google. Often search engines have to make a ‘best guess’ about the contents of a webpage. Is it a product page on an ecommerce site? A personal blog? A business listing? With structured data, there is no guesswork - the structured data snippet tells Google exactly what the page is. This makes it easier for Google to index the page and, more importantly, to use that page in its search results in the right way.

So for news articles, the NewsArticle structured data tells Google all it needs to know about the article. And that means Google can use the article as a news source, like so:

All the relevant elements that Google needs to show articles in Top Stories like this - the headline, the main featured image, the publisher logo, the article’s timestamp - are all included in the NewsArticle structured data snippet.

And yes, the headline you see there in Top Stories almost always matches the headline that’s included in the NewsArticle structured data snippet.

Again, this headline can be different from the visible headline and the <title> tag, but it’s recommended to keep it as close to those two as possible. All three of these headlines serve the same purpose and so should have the same meaning and contain the same keywords.

(Interestingly enough, it looks like the BBC Sport website doesn’t have any NewsArticle structured data implemented in their articles. Without that structured data, most publishers won’t have a hope of appearing in Top Stories. I reckon the BBC is a rare exception.)

The NewsArticle structured data headline has a different character limit, and this is not a visual limit or a ‘soft’ limit. It’s a hard limit of 110 characters. If the headline attribute in a NewsArticle structured data snippet is more than 110 characters, it breaks the validation of the structured data.

So you have a hard limit of 110 characters for the headline attribute in your NewsArticle structured data. Because the structured data snippet tends to be automatically generated by the technology that powers your publishing site, this is something usually handled by the developers that implement this.

Sometimes when an article fails to show up in Top Stories, it can be traced back to a problem with the NewsArticle structured data. Excessive headline length is an issue I see pop up from time to time where the publisher’s CMS doesn’t enforce that hard 110 character limit in its structured data snippet.

Now this doesn’t mean you should go wild and use all 110 characters all the time. As with regular search results, an article’s presence in the Top Stories box on Google also has a visual headline limit before it gets shortened with ‘. . .’, as shown here:

These three headlines all get shortened in the Top Stories box. That limits the headlines’ impact and is likely to result in fewer clicks to these articles.

Again, as this is a visual pixel limit, it depends on the width of the characters in your headline. Generally speaking, the break-off point seems to be around 70-75 characters. More than that and it’s very likely the headline will get shortened when the article is shown in Top Stories or a similar news box on Google.

The folks at NewsDashboard did research about the average headline length in Top Stories in the course of 2018, which showed that most headlines average between 58 and 72 characters.

Most of the publishers I’ve worked with simply take the visible article headline and chuck that into the NewsArticle structured data. That’s something to keep in mind when crafting your headlines - it may get used beyond just the visible article heading, and that means you may need to keep these restrictions in mind.

(Note: I’ve seen some instances where Google uses the Open Graph HTML meta tag title instead of the NewsArticle headline:

<meta property="og:title" content="Autumn Nations Cup: James Ryan urges patience among Ireland fans after defeats">

This is exceedingly rare and usually points at a problem with the article’s NewsArticle structured data snippet.)

Now there’s one more type of headline that an article may have.

4. Section Article Headline

Articles don’t exist purely on their own - they are part of a news site, and that means they are shown on homepages and section pages.

Often these pages are laid out in specific ways to make them look appealing and easy to navigate. And that means an article’s full headline can be too long to fully show, so a shortened version is needed.

The BBC Sport website is a great example of this. Each of the articles listed on their section pages are shown with shortened headlines that are abbreviated versions of their actual visible headlines.

This is a necessity because the layout of the site doesn’t cater for full 75-character long headlines. Using full headlines doesn’t fit in the site’s design, so shortened headings are needed for articles when they’re listed on a section page.

The purpose of these shortened headings is primarily to entice readers to click on them. For SEO their value is limited, so you don’t need to worry too much about them from that perspective.

Sticking with the James Ryan article example, the abbreviated section page headline reads “Ryan urgens patience among Ireland fans”. It omits his first name and the ‘Autumn Nations Cup’ key phrase, which are both present in the article’s full headline.

On its own, that shortened headline would not be good for the article’s performance in Google.

Fortunately, this section heading plays only a very small role in the article’s overall optimisation, with the other headlines sending much stronger relevancy signals to Google. So we don’t have to try and artificially optimise this headline to help the article’s Google rankings.

As long as it fits in the section page design and is sufficiently interesting for readers to click on, that’s all that matters.

Summarised

Something as simple as an article headline is actually a multi-faceted attribute that is subjected to different limits and best practices depending on when and how it’s used.

For optimising articles in Top Stories, the NewsArticle structured data headline is probably the most important one, just ahead of the <title> tag and the visible headline.

Make sure all three of these are optimised for SEO; that means they should include the keywords you’d like the article to rank for, and they adhere to the recommended character limitations.

If you’re depending on auto-generated titles and headline attributes based on the visual headline that you create, keep the restrictions for each in mind when you craft your headline.

If you’re lucky enough to work for a publisher that has separate fields in the CMS for each of these headlines, use that to your advantage to deliver the best possible headline for each type.

---

Miscellanea

I’ve come across a few great articles in the last while which you may find interesting:

• Google Discover: How to Rank and Drive Traffic

The folks at Ahrefs have compiled a very solid guide on optimising for Google Discover, with contributions from excellent SEO specialists Dan Taylor, John Shehata, Suganthan Mohanadasan, and Kevin Indig. Definitely worth a read for anyone involved in publishing SEO.

• How the BBC World Service migrated 31 million weekly readers to an isomorphic react app and improved page performance by up to 83%

This is a great piece about how BBC World migrated from a PHP technology stack to a React-based single page app. While SEO isn’t specifically mentioned, if you read the piece you do get the impression that technical SEO best practices were definitely part of the new platform’s spec.

• New and improved crawl stats for your site

Google has launched an improved Crawl Stats report in Google Search Console, which is superbly useful. Especially large websites (like news publishers) can find some seriously interesting information there about how Google is crawling your site, which may lead to some improvement opportunities for your dev teams.

---

Thanks for reading this second issue of my SEO for Google News email newsletter. We’re up to 640 subscribers now, a week after launch! I really appreciate all your interest in publishing SEO, and I hope I can do it justice.

If you have any feedback at all, send me an email (barry@polemicdigital.com) or leave a comment on the SEOforGoogleNews.com website. I want to make this newsletter as good as it possible can be, and your input is most welcome.

For the tweeters among you, I share more SEO stuff there - both general and news-specific - so if you’re not adverse to the odd bit of profanity you should follow me:

And if you feel anyone else would find this newsletter useful, do send it on to them!

Cheers,
Barry Adams