Ghost founder John and I have been talking content strategy lately. We've been really focused on blue-sky-synergy-sessions (obviously), but we've also been getting into the nitty-gritty on a few issues.
One that took us some time to hash out was how we approach headlines for the blog. John is a fairly pragmatic guy, while I tend to be too idealistic for a career in marketing. We debated how sensationalist our blog post titles should be, and eventually settled on "we don't know, so we'll need to experiment".
But the discussion got me thinking. I realised there are probably some other people who've explored the importance of headlines, and whether it really is necessary to write over-the-top "clicky" titles for your blog posts if you want them to be read widely.
I decided to look into it while we continue to experiment with our headlines here on the Ghost blog.
The state of headlines on the internet
Though there are many fewer newspapers, we all speak headlines now. — Tom Leonard, journalism history professor at the University of California, Berkeley
I don't know about you, but as a reader I'm increasingly aware of headlines. I notice when they're sensationalist, I notice when they're lists, and I notice when the curiosity gap is so big there's nothing but gap.
I've also noticed that the success of sites like BuzzFeed and Upworthy have kicked off bigger trends across the web. The sensationalist, emotional, curiosity-inducing headlines we know from those sites are proliferating throughout the content we read all over the web.
As BuzzFeed CEO Jonah Peretti points out in the quote below, we can go too far with headline optimisation.
Headlines optimization is a dangerous game. Realtime click data causes many publishers to over-optimize and manipulate readers into clicking stories they don’t actually want to read… In most cases it would be better for readers if the information was included in the headline so you only click if you actually are interested in reading the whole story. — Jonah Peretti
John suggested an interested way of thinking about headlines in our discussion. He thinks of headlines more like an ad for the content itself, rather than part of it. This idea made me rethink where to draw the line between the integrity of my work and the way I craft headlines to encourage readers to click through.
The key is to keep finding new ways to engage people with your content by being playful with the creative and ruthless with the testing. — Democratic digital consultant Joe Rospars
As Joshua Benton points out in a post for NiemanLab, the more of our traffic that comes from social media, the more our headlines have to fight for attention:
Every day, social’s share of online news traffic grows as more and more people get headlines from Twitter on their phones rather than a news website at their desks.
That new environment is, for instance, the main reason headlines have become so much more emotional and evocative. In a print newspaper, a headline is surrounded by lots of other contextual clues that tell you about the story ... Online, headlines often pop up alone and disembodied amid an endless stream of other content. They have a bigger job to do.
Over at Vox, Matthew Yglesias has struggled with the same issue John and I have been discussing. He points out something I noticed during our discussion: the term "clickbait" has developed from a term that referred to content that failed to deliver on a sensationalist, traffic-driving headline to a blanket term for content we don't like.
In a similar way, I've found myself using "spammy" to refer to content and headlines that are superficial or sensationalist. I've recently realised this is unfair, as they're not actually spam—I just think they're a waste of space.
The dream of every tech-savvy publisher in America is to develop algorithmic distribution tools that will know who you are and what you will like, ensuring that you are served all and only headlines and articles that you find both clicky and pleasing to click on. Thus the scourge of "clickbait" will someday be eliminated from the earth. — Matthew Yglesias
Do headlines even matter?
When it comes to drawing readers in and helping them understand your content, do headlines even make any difference?
Studies have shown that headings help to anchor our understanding of text and help us remember what we've read. It seems that a heading helps us to use knowledge we already have to anchor new knowledge we take in when reading, which makes it easier to remember the new facts later.
This is especially relevant when you consider a blog post that could have different headings depending on the angle you choose. This post, for instance, could have been called "6 ways to improve your headlines to drive more traffic" or "Blogs are the new magazines (or, why you should care about headlines)". Either of those alternative titles give the content a very different focus, and you would likely have felt different about it as you were reading, since I set you up to think a particular way with my headline.
However, this power, as all power does, comes with responsibilities. If you misrepresent something with your headline, readers not only form the wrong impression, but they're generally unable to change their mind as they read your post. First impressions have such a strong influence over us that we have trouble updating the information in our memory as we read. So even if you correct it in your post, using misrepresentation in your headline to increase clicks is a bad idea.
I can share something on Twitter with the wrong headline and it goes nowhere; with the right one it takes off. — Jay Rosen, NYU journalism professor
And, as Frank Moraes points out on his blog, Frankly Curious, it's common for a blog these days to operate like a magazine—that is, to only show you a post title to draw you in. Newspapers, on the other hand, tend to show you all—or most—of an article right under the headline. If your readers are seeing a list of headlines when they hit your blog, you don't have that extra context to rely on.
Those headlines have a big job to do.
Upworthy famously makes their writers create 25 different headlines for each piece of content before choosing the best one. In fact, they often choose more than one from the list of 25 and test them with posts on social networks to see which performs best.
The difference between a good headline and a bad headline can be just massive. It’s not a rounding error. When we test headlines we see 20% difference, 50% difference, 500% difference. A really excellent headline can make something go viral. — Upworthy co-founder Peter Koechley
Then again, some blogs have managed to grow incredibly popular without giving in to the headline hype. Stratechery, for instance, is featuring a post today called "Aggregation and the new regulation". Wait But Why's latest post as I write this is called "How (and Why) SpaceX Will Colonize Mars" and the one before is called "Why I'm Always Late".
The most recent post on A List Apart is titled "Thinking Responsively: A Framework for Future Learning". And the homepage for Signal v. Noise is currently rocking titles like "Less than perfect" and "The difference between time and attention".
As Joel Canfield says, if your blog is designed for your fans, headlines matter a lot less:
My blog is not the first place people will encounter me. My blog is for my fans, not strangers. Anyone who’s already here doesn’t care what the headline of the blog post is (do you?) They’ll read whatever I write because they’re fans.
This also points to an interesting difference between online communities where different headline styles are more effective. If you've ever posted anything on Hacker News, you might have noticed than sensationalist headlines don't trend too well with that audience. Unlike Twitter, where the audience is more broad and content is coming from all over, Hacker News is a community of people with similar interests who don't like clickbait.
Knowing your audience can make a big difference to how important your headlines are. If you're focusing on search traffic, SEO-friendly headlines will obviously be important. If you want to build a community around your blog of returning readers who become fans, you probably don't need to focus on clicky headlines so much as consistent, high-quality content.
What works best in a headline?
So you want to beef up your headlines, now that you agree they have the tough job of drawing readers in? I found a couple of people who'd analysed headlines online to find out what people read, click on, and share most often. Here's what I learned about what works and what doesn't.
Positive emotions get shared more, but negative superlatives get more clicks
A study from the University of Pennsylvania used online news to test what makes for a "most shared" article. They found two main aspects that predicted how successful an article would be:
how positive the message was
how much the article excited the reader.
While "articles that evoked some emotion did better than those that evoked none", happy emotions outweighed sad ones. And the more a story aroused the reader's emotions (i.e. how "excited" they were), the more likely they were to share it. Being highly anxious or extremely angry encouraged sharing just as much as a cuddly panda story.
The researchers also found that high-quality content and practical value made a difference to the success of a story. As did social currency—when sharing a story makes the reader feel like they're "in the know" as a meme does, for instance—and stories that triggered personal memories.
The irony, of course, is that the more data we mine, and the closer we come to determining a precise calculus of sharing, the less likely it will be for what we know to remain true. — Maria Konnikova, writer for The New Yorker
A study by Outbrains analysed roughly 65,000 titles from paid links that ran in the company's content advertising network over a three-month period.
The analysis found that negative superlatives like "never" or "worst" generated 63% more clicks than positive superlative like "always" or "best".
Compared to headlines with none, negative superlatives performed 30% better.
In a study by marketing management company Conductor, readers were tested to see how many superlatives they prefer in a headline.
From options ranging between zero and 4 superlatives, 51% chose 0-1. 25% chose 4—the most superlative-laden option.
It seems we like either few superlatives or a lot.
Curiosity works, though it won't forever
A content analysis of over 15,000 items from a popular Israeli website found that arousing curiosity encouraged more clicks, but not necessarily lots of comments. The items with the most comments tended to be "political/social topics and [items with] controversial elements".
In the past Upworthy has shown how arousing curiosity can increase clicks with headlines like these:
- The Things This 4-Year-Old Is Doing Are Cute. The Reason He's Doing Them Is Heartbreaking.
- She Has A Horrifying Story To Tell. Except It Isn't Actually True. Except It Actually Is True.
- This Amazing Kid Got To Enjoy 19 Awesome Years On This Planet. What He Left Behind Is Wondtacular.
Although this has helped Upworthy stories garner millions of views, the team is moving away from this tactic. Adam Mordecai, one of Upworthy's "star curators", says so many other sites are copying this tactic that it's no longer working as well as it did. It worked for Upworthy "because people weren’t used to it", he says, but descriptive headlines that explain what's in the content rather than teasing the reader are becoming Upworthy's best performers.
Even Facebook is in on the descriptive headline approach. A Facebook newsroom post from August 2014 explains that headlines which tease the reader to click are considered clickbait and will be pushed out of the News Feed by Facebook's algorithms.
The algorithm change was prompted by a Facebook user survey in which 80% of people said they preferred descriptive headlines in their News Feed.
Negative words draw us in
When the marketing team at Takipi analysed data from 100 blogs, they found some fascinating insights about which headlines perform best. One was that negative words like "kill", "fear", "dark", "bleeding", and "war" draw more shares. The analysis used tech blogs, so these words weren't used literally, but seemed to add enough colour to a topic to encourage more sharing.
Other negative words like "no", "without" and "stop" also increased shares, even for posts about thing you should stop doing.
We love lists
You probably caught onto this one already, but the Takipi team also found that lists draw clicks. In particular, it seems the bigger the number, the more we care.
Matthew Yglesias made a great point about lists in his piece on clickbait for Vox:
Magazines, which were traditionally bought off of newsstands packed with other magazines, are really good at choosing appealing headlines and striking images, and promising content the reader will enjoy reading. Lists weren't invented by Buzzfeed — they're all over the cover of every magazine.
We prefer title case
The Conductor study mentioned above also asked readers whether they preferred headlines to be written in sentence case, title case, or all capitals.
Sentence case is when only the first letter and proper nouns are capitalised.
Title Case Capitalises Pretty Much Every Word Longer Than Three Letters.
This is another area John and I debated, with neither of us knowing what would work best. I suggested we use sentence case on the Ghost blog because I like it, but in the Conductor study, 64% of people chose title case as their favourite.
Surprisingly, the second most popular choice was all capitals, with 21% of the votes.
This data made me rethink my choice, and you might have noticed I've been using title case around here lately.
Although, I was fascinated to find during my research into Upworthy's headlines that the older titles I used earlier in this post were written in title case, but I wasn't able to find a single title case headline on Upworthy's front page today. Even paging through their content I found all of their recent posts were headlined in sentence case. Does this mean they've used their stacks and stacks of data to realise we're all getting sick of title case and sentence case is what stands out now? Hmm...
So what have we learned after all that? It was a lot to take in, so I'll try to sum it up:
- Headlines can make a big difference to how many clicks a post gets.
- We share stories that focus on positive emotions.
- We click on stories using negative superlatives or words like "kill".
- The curiosity gap works, but it's getting worn out.
- We love lists.
- We prefer title case headlines.
So the takeaway lesson is to spend time on your headlines, pay attention to what works, but don't overuse the same tactic too much.
And take heart in this study, which found that above all, creativity in headlines is the most important factor in grabbing a reader's attention:
... headline readers tend to disregard standard norms such as length, clarity, and information as long as headlines rivet their attention in terms of creative style regardless of underdetermined semantic meaning.