How much time do you have to read this? Switch to 5 minutes version.
Our research shows that only one in five people read web content word for word. Majority scan, skip and only read key items of interest.
We asked the group that doesn’t read full content for their reasons and this is what they said:
There’s your quick answer. Some of my readers will leave at this point and that’s fine because the rest of the content is for those willing to dive deeper.
The Iceberg Model
So here’s a weird thing. This is a 400 word article that can also be 4000 words long. Being presented with everything at once can discourage reading so I only show the most important parts and let my readers carve a personalised path through available content by clicking onhypotext [+]←click to expand for more information
hypo ˈhʌɪpəʊ/ from Greek hupo ‘under’.
Hypotext is a way of revealing content on-demand. It acts like a traditional link, but it doesn’t interrupt user experience by sending readers to another page. Once clicked, the extra information is injected into a desired spot in the page. Another click hides it away.
This aids user experience in several ways:
– Supports easy scanning and better content overview by removing visual clutter.
– Encourages content consumption through low word count in its skeletal form.
– On-demand information retrieval enabled interactivity and personalisation.
– In-line citations ad sources improve trust and credibility of content.
– Users stay on the page they’re reading which minimises interruption.
This form of visual information compression enables users to gain the overall content allowing them to carve a personalised path through various bits of information.
Journalists on average tend to write in a more web-friendly way than most other online publishers. There are many great examples [+] out there but their format still appears limited by static medium principles (press).
One of my favourites is “Early morning is actually the worst time to drink coffee” article by Kia Kokalitcheva from Fortune. Her article is only 240 words which means that even a busy “scanner” like myself might make it to the end of the article. Kia gives the answer right in the beginning of the article and supplies her readers with background information in form of hyperlinks.
I’ve also seen experiments which combine inverted pyramid model with narrative. One example is “Health experts have figured out how much time you should sit each day” by Brigid Schulte from The Washington Post. Her piece is 1355 words long, more than most people are willing to read. Brigid does give the answer in the first paragraph, but the rest of the article continues without much structure making it hard to ascertain what value the remainder of the text holds for the reader. My guess is that this hybrid content model serves both types of readers, those who are in just for the quick answer and those who may be bored on the train looking to kill five minutes.
In response I created a modified version of Inverted Pyramid [+] in hope to align user experience with user expectations by looking at the web for what it is – a dynamic medium.
The inverted pyramid is a metaphor used by journalists and other writers to illustrate how information should be prioritized and structured in a text (e.g., a news report). It is a common method for writing news stories (and has adaptability to other kinds of texts, e.g., blogs and editorial columns). This is the best way to understand the basics about a news report. It is widely taught to mass communication and journalism students, and is systematically used in Anglophone media. Source: http://en.wikipedia.org/wiki/Inverted_pyramid
On the left we see a classic inverted pyramid. On the right is a modified model with background information folded in hypotext. This meta data is retrievable on-demand in a non-linear mode. Supplementary information does not have to appear at the bottom of the content. Its placement can and should be contextual.
It’s not them it’s us
How we write isn’t compatible with how we read on the web.
What’s amazing about our findings is that they’re identical to those published by Nielsen [+] back in 1997. Basically, how we write isn’t compatible with how we read on the web and we haven’t done anything about it in nearly two decades. I don’t know about you but I see a great opportunity here.
The first stage of content consumption research took place in April 2015 and it included five hundred Australian Android users who were asked how they read online. Only 15.9% of our respondents tend to read content in full.
I read most but may skip some parts
I skim through looking for key items
I read everything in full
A headline is enough information
Our results were identical to those published by Jacob Nielsen in 1997:
In research on how people read websites we found that 79 percent of our test users always scanned any new page they came across; only 16 percent read word-by-word.
The follow-up research took place in May 2015 and included one hundred pre-qualified Australian Android users who never or almost never read content in full. Respondents were offered a list of six reasons for skipping content and could select multiple items if applicable.
I’m impatient to find quick answers
Find text to be too long to bother
Lose interest in what I’m reading
Poor layout and text formatting
A difficult or confusing read
Don’t trust the website
From additional replies we’ve learned about several other significant factors which include procrastination, dislike for aggressive advertisements and certain content formats such as lists and clickbaits:
Advertisements too annoying.
Generally learn all I need to know in the opening paragraph or subsequent paragraph.
I am a speed reader.
I’m looking for a specific piece of information no need to read irrelevant information.
Web content is being produced at an unprecedented rate. Article marketing, guest blogging, advertorials, opinion pieces, interviews, lists, clickbait! We’re saturating digital space with content pieces aggressively competing for user attention, causing content fatigue [+].
Question: What is the right dose of exercise for a longer life? Answer: 450 minutes per week
Would you believe it took Gretchen Reynolds of The New York Times 900 words [+] to say this?
The Right Dose of Exercise for a Longer Life
Exercise has had a Goldilocks problem, with experts debating just how much exercise is too little, too much or just the right amount to improve health and longevity. Two new, impressively large-scale studies provide some clarity, suggesting that the ideal dose of exercise for a long life is a bit more than many of us currently believe we should get, but less than many of us might expect. The studies also found that prolonged or intense exercise is unlikely to be harmful and could add years to people’s lives.
No one doubts, of course, that any amount of exercise is better than none. Like medicine, exercise is known to reduce risks for many diseases and premature death.
But unlike medicine, exercise does not come with dosing instructions. The current broad guidelines from governmental and health organizations call for 150 minutes of moderate exercise per week to build and maintain health and fitness.
But whether that amount of exercise represents the least amount that someone should do — the minimum recommended dose — or the ideal amount has not been certain.
Scientists also have not known whether there is a safe upper limit on exercise, beyond which its effects become potentially dangerous; and whether some intensities of exercise are more effective than others at prolonging lives.
So the new studies, both of which were published last week in JAMA Internal Medicine, helpfully tackle those questions.
In the broader of the two studies, researchers with the National Cancer Institute, Harvard University and other institutions gathered and pooled data about people’s exercise habits from six large, ongoing health surveys, winding up with information about more than 661,000 adults, most of them middle-aged.
Using this data, the researchers stratified the adults by their weekly exercise time, from those who did not exercise at all to those who worked out for 10 times the current recommendations or more (meaning that they exercised moderately for 25 hours per week or more).
Then they compared 14 years’ worth of death records for the group.
They found that, unsurprisingly, the people who did not exercise at all were at the highest risk of early death.
But those who exercised a little, not meeting the recommendations but doing something, lowered their risk of premature death by 20 percent.
Those who met the guidelines precisely, completing 150 minutes per week of moderate exercise, enjoyed greater longevity benefits and 31 percent less risk of dying during the 14-year period compared with those who never exercised.
The sweet spot for exercise benefits, however, came among those who tripled the recommended level of exercise, working out moderately, mostly by walking, for 450 minutes per week, or a little more than an hour per day. Those people were 39 percent less likely to die prematurely than people who never exercised.
At that point, the benefits plateaued, the researchers found, but they never significantly declined. Those few individuals engaging in 10 times or more the recommended exercise dose gained about the same reduction in mortality risk as people who simply met the guidelines. They did not gain significantly more health bang for all of those additional hours spent sweating. But they also did not increase their risk of dying young.
The other new study of exercise and mortality reached a somewhat similar conclusion about intensity. While a few recent studies have intimated that frequent, strenuous exercise might contribute to early mortality, the new study found the reverse.
For this study, Australian researchers closely examined health survey data for more than 200,000 Australian adults, determining how much time each person spent exercising and how much of that exercise qualified as vigorous, such as running instead of walking, or playing competitive singles tennis versus a sociable doubles game.
Then, as with the other study, they checked death statistics. And as in the other study, they found that meeting the exercise guidelines substantially reduced the risk of early death, even if someone’s exercise was moderate, such as walking.
But if someone engaged in even occasional vigorous exercise, he or she gained a small but not unimportant additional reduction in mortality. Those who spent up to 30 percent of their weekly exercise time in vigorous activities were 9 percent less likely to die prematurely than people who exercised for the same amount of time but always moderately, while those who spent more than 30 percent of their exercise time in strenuous activities gained an extra 13 percent reduction in early mortality, compared with people who never broke much of a sweat. The researchers did not note any increase in mortality, even among those few people completing the largest amounts of intense exercise.
Of course, these studies relied on people’s shaky recall of exercise habits and were not randomized experiments, so can’t prove that any exercise dose caused changes in mortality risk, only that exercise and death risks were associated.
Still, the associations were strong and consistent and the takeaway message seems straightforward, according to the researchers.
Anyone who is physically capable of activity should try to “reach at least 150 minutes of physical activity per week and have around 20 to 30 minutes of that be vigorous activity,” says Klaus Gebel, a senior research fellow at James Cook University in Cairns, Australia, who led the second study. And a larger dose, for those who are so inclined, does not seem to be unsafe, he said.
Software engineers now write bots, using rudimentary AI to summarise articles. When facing a long article many skip its content altogether and look at the comments in hope to get quick answers there. Here’s one I prepared earlier [+].
11 Foods You Don’t Need to Refrigerate
Reader Allyson writes: “You should do a video where John Green talks about foods people can stop putting in the fridge. Maybe he can leave out foods for a few days and try them, as a test.”
While I do like this idea, I’m not sure it’s worth poisoning our wonderful host. But lucky for us (especially John), someone already filmed something sort of similar. Here’s a quick rundown of foods you don’t need to refrigerate and what happens to them when you do.
Prior to publishing this article I reached out to a number of authoritative figures in our industry seeking peer review on use of hypotext and the web-friendly inverted pyramid model. The feedback has been outstanding but it will cost you nine minutes to read. It’s your call [+].
I’m a fan of cutting things to a minimum.
Webmaster Trends Analyst
That’s a neat idea. We’d probably devalue the hidden content, but maybe that’s fine… My initial thought was if it’s not critical to the article, why include it at all? I’m a fan of cutting things to a minimum, and sometimes spend more time removing things than I spent writing it in the first place.
My other thought was that you’ve reinvented footnotes. I imagine for some audiences this will work great, I don’t know if the effort required to do it would be worth it under the line though. You have to write & review the full content, then work out where to split things out, and find clean transitions that make it possible to hide blocks without impairing the readability of the rest (I also had a hard time figuring out where to click — dotted links can be lots of things, I’d test various graphic designs). You’re probably doing more work than if you created two versions of it separately.
At any rate, this is one of those things that’s probably easy to instrument for measuring (track views, clicks, time spent reading at which part of the page), and where real data can help to make a decision. I’d try to track as much as possible in those tests, for example, how far they’ve scrolled and how long they spend looking at a specific piece of the content.
If you could automatically recognize what kind of reader was viewing the page (do they spend time, or are they just scrolling?), blend in a subtle dialog on the side or even just display the content below the viewing area, that would probably increase the usability by almost completely removing the UI.
Most people don’t know how to write for the web. They still treat it like writing a magazine or something and those are not the same things. I usually tell people that web writing is as different as grant writing is from haiku. I draw on things such as Krug’s Don’t Make Me Think, Nielsen’s repeated research in this area and Schwartz’s Paradox of Choice.
The goal is to reduce cognitive strain when you approach content. Do that and you’re well ahead of the game. Unfortunately, very few think about such things. The hypotext is interesting but I don’t like the on-demand aspect of it. That’s a bit of friction – of effort – that I think probably reduces the chances of that text even getting read at all. It would be interesting if those things auto expanded as you scrolled to them – so that readers weren’t overwhelmed at the beginning and were given well structured content as they continued.
Also – related to this – I believe the ‘reading time’ metrics (started on Medium if I’m right) are another way to help improve readability. Finally, here’s what I wrote on the subject back in 2012.
Any piece of quality content should scream: READ ME!
In my opinion, the biggest problem with content nowadays is the misalignment between user searches and intentions with the content on a page. It’s common that site owners and marketers forget that content doesn’t necessarily correlate with text or keywords; useful content can adopt many forms, like graphics, video, audio, you name it. The problem is that most times we fall into the trap of writing for search engines and forget to make their content useful and relevant… FOR PEOPLE!
Content creators should focus on creating the best content, and then together with marketers and SEOs find ways to structure and amplify it in a way that makes it attractive and user friendly. Any piece of quality content should scream “READ ME! Look how great I am” instead of just being another dull wall of text that gets defeated by the scroll wheel.
Summarising. Content should:
– Be relevant and aligned with the nature of the website;
– Be useful for that intent and solve a user’s need;
– Be structured so it’s easy to consume and enjoy;
– Be creative in the way of conveying information;
If your content doesn’t follow these guidelines, very likely you’re wasting your time, money and making unhappy users.
It’s a more complicated situation than is being considered.
I think it’s a more complicated situation than is being considered and depends entirely on the content subject itself. For example – I just got a headline alert from the NYT saying “Top Soccer Officials Indicted for Corruption” – I didn’t click through – the headline told me all I needed to know. But if I got an alert saying “Supreme Court Issues Verdict on Gay Marriage” I’m going to click through and read most, if not all of that content. If someone tweets “Foo Fighters Announce Tour Dates” I’ll look at that and head straight to wherever the west coast dates are listed, ignoring the rest of the content. If it’s new research/data/science – I will read it all in full. This is also how I read the LA Times and the Wall St. Journal.
Users aren’t inclined to take action unless the value proposition is clear and present.
I like the idea of what the hypotext concept is trying to accomplish. I do wonder if users will understand what the UX will be when/if they click those dotted line links. They may either expect to leave the page (as in a traditional hyperlink), or like me, initially they may think those dotted line links lead to ads. So that behavior may have to be trained.
As always, users aren’t inclined to take action unless the value proposition is clear and present. I wonder if there may be a way to provide some visual cue as to what will be revealed upon a click, perhaps a translucent thumbnail or something.
Another thing I struggle with: where does it stop? In the example on your page there seems to be just two layers of depth. But in theory one could click on a hypotext link, reveal a block of additional content, which itself has hypotext links and you could keep on exposing layers. Pretty soon the whole universe of information is a giant expandable outline, like Workflowy. Or a fractal. Users sometimes lose comfort when they can navigate endlessly.
The Paradox of Choice.
Our robots would index the whole content.
Director of Technologies
I think all search engines use the same practice to value only visible content. We have no penalty for invisible content, but for some cases we use a headless browser to detect which part of text are visible for users. I checked your draft and for Yandex is OK. Our robots would index the whole content.
If a technology proves to be popular by users Google will embrace it.
There seems to be a lot of assumption implied in saying users generally merely screen text content. As you know, statistics like that are often a matter of looking at them. Why whether users briefly screen content or read it in full length or do something else completely can be debated, it seems more probably that user behaviour varies strongly depending on the type of content or vertical, e.g. academic paper vs. entertainment industry news. Having said that if the past years have taught us anything than that adaptation of new technology takes often longer than expected. Mobile is a good example there.
Google however seems to be very consistent in their strategies and largely immune to short lived trends. They seem to be focused on user experience, much more than anything else. It’s safe to assume that if a technology proves to be popular by users Google will embrace it. Otherwise it will likely receive little attention.
I don’t think it’s advantageous for the publisher or writer to deliver the answer in the first paragraph.
I don’t think it’s advantageous for the publisher or writer to deliver the answer in the first paragraph so the user can quickly bounce away. The format of your article seems to facilitate this, while our objective is to keep readers engaged on our platform. I think you have it right in your first paragraph – Majority scan, skip and only read key items of interest. Taking your article as an example, the hypotext seems to hide key interest points that scanning readers would stop to dwell on, like the summary table of research replies and diagrams.
I believe the first few lines of an article are better served to capture interest but still leave something the reader wants (namely, the promise in the headline).
I think the idea is excellent. In many ways the principle has been used by newspapers where the Headline, subheading, image and caption tell the story in broad strokes. There are two reasons newspapers did that: one, it allowed readers to prioritise their reading and go deep only in stories that interested them – two, it allowed readers to read-scan a paper and still think it was worth the money they paid for it, even though they did not really read it.
There are two issues with using this that i can see – the first one is quality (and you can’t do much to improve this). Some people will be really good at putting it together in a way that does the job (journalists are trained specifically and they have editors) – others not so much. Because in principle it is a very attractive way of actually putting a story together we may end up with even more rubbish on the web than we have at the moment where writing is a sort of barrier as well, not everyone can do it.
There is nothing you can really do to improve quality here of course, but it is worth keeping in mind the second one is implementation. Most CMS websites allow a very specific way of creating content that most users can follow the second obstacle is, to my mind the biggest of the two.
I have long maintained that we are not yet truly digital – all we have done is taken the tech of the past and digitized not but not really used digital to rethink it in a new radical format that takes advantage of what the medium has to offer. This is definitely a step in the right direction – Bravo!
The number 1 reason given for bouncing was that the headline they clicked on didn’t deliver on its promise when scanning the page.
Our internal UX team at News Corp mentioned that during a series of test where a user clicked an associated story (presented in the sidebar of the page or promoted with a catchy “what happens next will shock you…” bait headline) they found the number 1 reason given for bouncing was that the headline they clicked on didn’t deliver on its promise when scanning the page for the same relevant phrase contained in the heading link.
Editorial Golden Rule – “Headlines make a promise to the reader, body copy must deliver on that promise”.
Your job as the content creator is to cater to EVERY type of reader – the scanner, the deeply invested, the guy who gets bored halfway through and hits eject.
“Content” is not homogeneous, neither are the readers. For example, when writing a landing page for CRO, you try to account for the multiple types of readers – those who will skim with a specific solution or question in mind, and those who are actively engaged in problem solving and want more than a lightweight comparison. You try to make the content appeal to both, with the full knowledge that formatting for scanning is necessary because people will always try to protect their most valuable asset – their time.
“Time to attention” ratio is important to consider – how much time are you demanding of a person in exchange for the information they need? When that balance shifts, scanning happens – and you don’t get to demand the acceptable wait time beforehand because it’s dependent on the user:
Where are they? What are they reading on? What is their intent in reading the content – to get a fast answer, entertain themselves or help evaluate a serious decision?
In a blog post that promises 5 tips, can you blame people for cutting to the meat of the article and skipping the intro, where they know nothing is waiting for them?
Do we really blame people for skimming that eBook you put out for the most actionable bits?
On a website, can you really expect people to read every last line when they’re on a mission to accomplish something?
But – when reading a book – almost nobody will skip over chapters unless the writing is truly dreadful, because their motivation is to be captivated by the story.
Your job as the content creator is to cater to EVERY type of reader – the scanner, the deeply invested, the guy who gets bored halfway through and hits eject. Every line needs to be captivating, every section scannable – and the “so what” can never be buried, or it will never be found.
I like the idea of hypotext and allowing the user control over what they read/don’t read; I’d be interested to see the stats on how they engage with the copy. This level of interactivity could also be conversion fuel – if you see a ton of people opening an internal link, chances are good that information is valuable to them; you may be able to draw some conclusions on this kind of behavior.
That said, I don’t think you can ever solve this content “problem” by improving your content. People are distracted and fickle – they skim because they’ve learned that behavior after years of reading content or even just studying from textbooks. Consuming content can be a transactional experience (“just give me the info!”), and immersive experience (“tell me the details!”) or both – and as a writer, you don’t always get to decide which that will be. But if we look at the kinds of content people DO stay tuned into, there are clues for at least retaining the majority of those who are willing to be retained. Some percentage – probably a substantial one – will always be in a mindset more conducive to scanning, and you’ll never change that entirely.
It reminds me very much of the early days of the Web when we were guided to write in short sentences.
The research was really very interesting. It reminds me very much of the early days of the Web when we were guided to write in short sentences, and very short paragraphs for the “skimming” nature of the web. It, quite literally, blew away all the notions of the minimum 5 sentence paragraphs I grew up with in my English classes. But, then again, they call me grandpa because I tend to put double spaces after periods, so there’s that as well – Ha!
In any event, I thought the conclusions and the research were really interesting. But I wonder if you’ll run into challenges trying to call the accordion type interface (which is pretty common these days in HTML 5 UX Design) “Hypotext”. I absolutely do think the concept is a valid one – but not sure that other than the specific application you’ve got in your article (e.g. your iceberg approach to deeper layers of content) that it’s unique enough to segment off from the other types of HTML 5 accordion type of interfaces that are out there.
In short the question I have is:
Is it “Hypotext” because it’s referring to the KIND of content underneath the interface (e.g. more detailed, nuanced content)
Is it Hypotext because of the HTML interface trick.
Both need to be true to be Hypotext
If it’s the former – I think you may run into purists who will debate your naming – based on the origins of Hypotext (though you can make a good argument for this I suspect). Though this is really where your pyramid comes in – and this may be your whole point. In other words – it’s not the simple link and expansion that is “hypertext” – it’s the act of building your interfaces in the refined pyramid that means you’re using Hypotext.
If it’s the latter – I think you may find that there’s just already enough names out there for expanding panels of content in HTML 5 interfaces that people will say “do we really need another name for this?”
If it’s both – then Hypotext is a technique for ui design. It includes a design pattern, a writing style and a technology application.
‘Hypotext’ involves the background or expansion text appearing within the main copy when you click on a link.
The problem is that we need to present content in such a way that “skimmers” can get the information they need quickly, without neglecting people who want to read every last word of background.
The idea of hypertext (links to other documents within the text) has been a fundamental part of the web since its inception. However, while using links for background information streamlines an article, clicking on them takes people away from it. The idea is really no better than footnotes in books, which you can read but then awkwardly have to try to find where you were in the main text.
The concept of hypotext being put forward involves the background or expansion text appearing within the main copy when you click on a link. It’s not difficult to implement, and if done right, could work well for both publisher and reader.
Additional Comments & Reactions
I speed read and skim. And if you don’t use proper headings to break things up and help me get to the nugget I’m looking for, I usually bail.
I will often skim read most things and if I enjoy the post will re-read it again. I really prefer articles with images and data represented in graphs. If you just have a single block of text I’ll probably not read it.
One thing that I think has dramatically worsened this problem is inline ads. People are now accustomed to skipping over content because the “content” isn’t actually content anymore. If you have a bunch of “scantily clad girls living in your area” in the middle of an article about conflict in the middle east, of course I’m going to skip around.
Dan Petrovic, the managing director of DEJAN, is Australia’s best-known name in the field of search engine optimisation. Dan is a web author, innovator and a highly regarded search industry event speaker.