Stop Pagination Now
slate.com
slate.com
As much as I hate it and would love to think that pagination is just a publication pipedream, it's not.
Some online publications have spent (and will spend if they're interested in making $) significant time and $ A/B testing various levels of pagination such that their sites are optimized so that there are just fewer pages than their readership is unwilling to tolerate.
The end result is that their sites are optimized to make the most amount of money. If it didn't improve their revenues, they wouldn't be doing it.
Of course, there are copycats that don't do the testing, don't optimize for their niche of readers, and are probably losing both money and readers because of it.
We can argue about publications adopting different business models, but that's a different story. Advertisers pay a premium for engaged, targeted niche audiences that match their products/services. That's the reality of the online advertising market.
I can assure you that this is absolutely not the case. If you look at Plotz's comment, it's exactly what most editors will say:
Pages that run too long can irritate readers. We run stories of 2,000, 4,000, even 6,000 words, and to run that much text down a single page can daunt and depress a reader. So pagination can make pages seem more welcoming, more chewable.
They really believe this.
It's important to realize that despite all the buzz about A/B testing in the industry, editors pretty universally believe they don't need to validate their opinions through testing. Why not? Because though school and their careers they've been told that their opinions matter, and, in fact, that's all their job really is: voicing opinions and making judgements, because they believe they have special insight into what readers want.
Look no further than Poltz, a writer and editor with no software design or development experience, who somehow believes he's qualified to make that statement.
(hopefully a wider content-well after an upcoming redesign will help)
I'm curious to hear what people think of a hybrid pagination system, like that currently used by Computing Now magazine. I like that it gives the user a sense of the article length without interrupting the flow of reading. http://www.computer.org/portal/web/computingnow/careers/cont...
It's completely justified. For example: show me where in Poltz's long career as a writer and editor that he secretly squeezed in the years of UX, design, and development experience that qualifies him to make a statement about designing user interfaces.
The problem is that editors simply don't believe there is special expertise here or that it requires anything beyond their insight.
I admit that I'm not always happy with the result ... http://spectrum.ieee.org/aerospace/astrophysics/the-cosmolog...
This design doesn't incorporate many of the base standards for displaying long-form web content. Font size is way too small (should be 16-18px, 14px absolute minimum), column is too wide ("ideal" is 66 characters, though there's a wide range). A good starter article: http://informationarchitects.net/blog/100e2r/ There are many, many more.
Until that page is completely redesigned and returned to a readable baseline, it's difficult to discuss or test anything.
>The fact is that editors simply don't believe there is special expertise here.
I agree that UX and design are often neglected when traditional publishers work in digital mediums. But I find it odd that you can appreciate the experience that leads to good UX and design but not the editorial and reporting experience that leads to good copy.
Writing, content strategy, curation, copywriting, etc are obviously all important skills. However, being an editor does not remotely qualify someone to run a multimillion dollar business, design software, dictate engineering decisions, etc. Unfortunately, the industry is structured such that they are doing exactly that.
Is it any surprise that slate's division is operating at a loss? It's the norm in a broken industry.
> However, being an [insert pre-digital mgr/executive]* does not remotely qualify someone to run a multimillion dollar business, design software, dictate engineering decisions, etc. Unfortunately, the industry is structured such that they are doing exactly that.
This is everywhere, I think. I've experienced first-hand the same problems in non-publishing industries where companies who rely on the web to do business are staffed with people who hardly understand the web making technical and software decisions, overriding software engineers and programmers making suggestions that are in opposition to the random thoughts in their heads. This is where the trump card of being higher up the ladder is played, and typically with detrimental results of varying degrees.
It's a serious problem in any business when web & technology decisions are still being made by people who don't understand the web & technology, and can't build it themselves.
Outside of the software industry, most businesses I've run across think being able to open a web browser or being in a mgmt/exec-level position qualifies them to actually make good software decisions.
* By pre-digital I mean old businesses that have moved to the web for various things, but are staffed by people who don't build software making key software decisions [e.g., electrical engineers, marketing execs, etc. making decisions on platforms, languages, and implementations while ignoring the advice of actual developers].
Somewhat ironically, I find that page a little hard to read because the background is so bright. I find reading text on HN much easier with a darker grey background.
I think a potentially nice implementation of this idea would be more like reading on a Kindle - text is broken into screen-sized chunks, and then there are low friction ways to go forward and backward "chunks".
By adding a 'next page' button, you create an incentive to keep reading. By spiking the curiosity of the reader to find out what's behind the bend, they linger instead of leaving.
While perhaps the last part of the deal is measured in ad impressions, all the other facts and figures are important: time on site, pages per unique visitor, where the site's traffic is coming from, etc.
I think this is much more common than sophisticated A/B testing.
At our publication, we were able to make the business case that the added pageviews from pagination didn't make much of an impact on the bottom-line. Because we generally run out of premium ad inventory each month anyway, most of the extra impressions were just for extremely low-revenue, Google-served remnant ads. Again, some publications thrive on volume, but it's a tough way to make money as CTR on Google ads is very low.
>If it didn't improve their revenues, they wouldn't be doing it.
The publishing industry has not always been about maximizing profits. The weird hybrid of advertising and reader-supported has created a different mentality in the business. There are plenty of things that publishers still do- like supporting foreign correspondents- that are driven more by the editorial side of than the business side.
but it must be said: users of that site special, they actually want to find something, so they are (mostly) not casual users.
i remember a case a few years back on a travel site where we had a different experience, time on site went up after we ditched pagination (but that site also had horrible page load time, so every page view was a great opportunity to leave the site, so well, it's all interconnected)
Advertisers want to appear near the top of the page or near the top of the content - that's why ad networks have rules about the placement of ads. Even Google (which is more performance than brand) limits the number of units that are displayed on a single page.
If your ads are at the top then there's a good chance that they will appear on the user's screen. If your ads are at the bottom and the user found the article boring then the browser was closed before seeing the ad, yet the impression was still paid for (hence, much lower rates).
I'm guessing the issue is that if you use paged content you know that the user has looked at that part of the article. If the adverts are all on the same page then an impression is not proof that a user has had that advert appear on the screen. Having said that if bootstrap's scrollspy can detect when you've scrolled content into view then I don't see why you shouldn't be able to track advert views in the same way.
Typically, it makes for a horrific reading experience. And I believe it will become bad for business, too, if enough high-quality content sites emerge and offer a far more compelling experience to draw readers away.
There's a hurdle preventing that from happening, however, and the hurdle is that users seem fairly site-agnostic in their content preferences. They consume information without a great deal of thought as to where it came from. And it's easier than ever to receive content now, through social channels, and never even have an inkling as to its original source or site. This liquidity and disposability of content means that content farms still have a big advantage over sites that attempt to make themselves into interesting, repeat-visit destinations.
Now, savvier and more intelligent readers -- the kind who probably frequent this community -- tend to avoid content farms, and probably don't click on articles entitled "10 Totally Weird Diet Tricks That Will Burn Away 10 Lbs in One Day!!!," requiring a 10-page click through to read. But sadly, a lot of people do click those things. A lot.
If you are selling views, that will not wash with the advertiser as you can't reliably tell if the user has scrolled to see the lower adverts (whereas you can easily see if they clicked to see page two).
Of course even now your stats are going to over-count already (the ad may stil count as a view if there was an error on the ad server, or if the user had a proxy or plugin that blocked it (depending how that blocking is done), and so on) but that particular source of uncertainty is already accepted as an unavoidable consequence of the way the tech works, where the advertisers will see the "can't really tell if they saw the lower one" issue as a solved problem: just put the lower content and adverts on a new page.
The user might click through to all the pages, but obtain a general disdain for your site. Not saying that pagination enough to make users hate a site, but for me, personally, it doesn't exactly help.
If I'm simply tolerating some aspect of a site, it means that as soon as an alternate source of or means for accessing the for the same content shows up - I'll leave and never look back.
Its not so much the reading. Its the slow page loading which gets me down. If people did pagination right with ajax preloading the next page then I would be happy. 99% of sites do not do this. Instead their sites are slow because they have to load a bunch of third party ad's. Moving from page to page when each page takes 2-3 seconds to load kills the reading experience.
A Wired article featured on HN recently. I don't remember what it was about but it had a gallery on it. You would think in today's day and age when you click the thumbnail the main image would change. On this wired story no.. the whole page reloaded. Due to third party ad's each page load took ages. The page didn't bounce down to where the image should show. While I thought the topic was interesting the user experience was so terrible I abandoned it.
If you are going to force pagination on readers then have a single page option. Preferably at the top of the article. This way you cater for everyone.
I almost never read long articles to the end, unless I've added them to readability. If I can't open them in the background so they're loaded when I want them, I usually don't bother
If there's no Print option, I'm not bothering reading the thing, unless it contains info I absolutely need.
BUT, from a web design perspective, there are actually some good reasons behind text pagination, the main three being:
1) It helps you keep your place -- you can bookmark one of the 15 pages to continue reading from later on, or forward to your friend ("check out the third paragraph" instead of "84th paragraph"). Obviously, you'd never show a 100,000-word article on one page (except for printing). At some point, pagination is necessary. Where to draw the line is obviously depends on a multitude of factors.
2) Feeling like you're "in" the site, not in no-man's land. From a web design perspective, you always want people to be able to navigate easily to other parts of the site -- you want stuff in the sidebar, never be too far away from header/footer, etc. You don't want people to be in a narrow column of text, with total white on both sides and nothing else. (Obviously users might often prefer this, like I do, hence Instapaper etc., but other users also like to be able to skip around to the rest of the site easily once they decide this particular article isn't for them, etc.) Of course, "floating" sidebars, headers, etc. can help alleviate this, but this requires the sidebars to be very small, which is not always possible.
3) "Chunking" large pieces of content is just a good thing. A huge wall of text is just no good from a usability or legibility standpoint. This, of course, is why articles are broken down into sentences, paragraphs, sections, parts, etc., so people "know where they are". Now I'll be the first to admit that paging is a particularly inelegant sledgehammer way of chunking, but when you're dealing with lots of articles of different lengths, formats, etc., it's a kind of lowest-common-denominator that is seen as better than nothing.
I don't think the current way most sites handle pagination is really that good, and it's a particularly bizarre combination to have pages that are longer than the viewing area, so you have to scroll and have to page. But given the way browsers are put together, it's a solution that basically works, given all the design constraints.
As a final note, I also think sites which don't provide any single-page or print option at all, also deserve a special place in hell. :)
If you really believed in 1, 2 & 3 then why start off saying you hate it too? And you think everyone else hates it too?
The reasons you give seem like building for the edge cases instead of the main use cases.
Also if this were the case they'd have ajaxified these articles long, long ago so it doesn't require a full page load to paginate to improve retention.
We all know why it's done, it's for more ad impressions.
Saying s/he hates it doesn't make the other points wrong (I actually agree with them). The dislike for pagination might come from waiting for the page to load. It doesn't feel nice to see the whole page flash and wait for it to load again. It's a poor experience. If loading the next page is as quick as it is on Amazon's cloud reader, maybe people wouldn't hate pagination as much.
Put permalinks next to sections headings and/or change the fragment in the url bar with javascript as the user scrolls. Links work, and no need for pagination
>2) Feeling like you're "in" the site, not in no-man's land. From a web design perspective, you always want people to be able to navigate easily to other parts of the site -- you want stuff in the sidebar, never be too far away from header/footer, etc. You don't want people to be in a narrow column of text, with total white on both sides and nothing else. (Obviously users might often prefer this, like I do, hence Instapaper etc., but other users also like to be able to skip around to the rest of the site easily once they decide this particular article isn't for them, etc.) Of course, "floating" sidebars, headers, etc. can help alleviate this, but this requires the sidebars to be very small, which is not always possible.
Use position: fixed (as you mention).Having submenus that expand on click alleviates the space constraints. You can also swap out content in your header or sidebar as the user scrolls (but please use tranitions, or you'll give everyone eyestrain).
> 3) "Chunking" large pieces of content is just a good thing. A huge wall of text is just no good from a usability or legibility standpoint. This, of course, is why articles are broken down into sentences, paragraphs, sections, parts, etc., so people "know where they are". Now I'll be the first to admit that paging is a particularly inelegant sledgehammer way of chunking, but when you're dealing with lots of articles of different lengths, formats, etc., it's a kind of lowest-common-denominator that is seen as better than nothing
What's wrong with section headings?
A lot of users don't understand section permalinks, fragments. Position: fixed severely constrains your design possibilities, and swapping out content just involves a lot more design and programming. And section headings are great, but getting your CMS to implement a mini-table-of-contents at the top/bottom of each page may involve a lot of programming too, and changing editorial processes, etc.
That's why I called pagination a kind of lowest-common-denominator solution that isn't great, but works. Clever programming can come up with better solutions in many cases, but they can be a much bigger pain in the butt to implement well, and there are often much more important priorities.
Taking existing websites and just getting rid of pagination isn't necessarily always an improvement.
I hate to be contrary, but no, I really don't want anything in the sidebar, at least when it comes to a piece of text that requires a certain amount of concentration to consume. That's why it's actually easier to read something from, say, the new york review of books in a physical form then it is from the website; there's just too much potential distraction, full stop, while the physical page tends to be more focused, more clean, for whatever reason.
Anyway, I think it's ESR that does the best text/article presentation on the web, for example (http://www.catb.org/esr/writings/homesteading/cathedral-baza...)
If you're browsing on an (arguably flawed in its implementation) iOS device where page state gets unexpectedly dumped all the time (e.g, tab switching, browser crashes, "multitasking" switches out to an IM client), this style of "pagination" is utterly pants. I end up having to manually edit the URL to advance through content like Boing Boing and Tumblr just to be sure that when Safari or Chrome inevitably next dump my page state I won't have to start from the very top again.
Perhaps it's irrational, but it also frustrates me that such pages have no bottom. I think people got used to a model of a page where it has a top and a bottom, and now it's a bit weird to interact with. One of my favorite examples is Facebook's implementation, which leaves several useful navigation links in the page footer. These links taunt you as you tried to scroll to reach them, only to have more news content appear in your view and push the footer off screen again. (I think they've since duplicated those footer links in other locations so you can reach them, but it doesn't change the fact that this broken UI model presents options and then sucks them away before you can click on them.)
I'm perfectly able and quite willing to approximate my progress with reading an article by observing the scroll bar. However, since the advent of social, I find I can rarely trust that information, since the comment section adds an additional multiple to the page length. I often find myself going on a pgdn adventure to figure out the length of the article I've been reading for ten minutes.
Would it really be so awful if the dozen pages of commentary were collapsed behind a "show comments" or "show more comments" button? I want my progress indicator back.
It's not obvious. It depends on the situation. At Etsy, we experimented with infinite scroll in search and pretty much every metric we cared about got worse. But we do use it successfully in other contexts.
The thing that annoys me about this article is that it could mention A/B results in some way, but it doesn't. He's just countering some editor's opinion with his own.
http://www.slate.com/articles/technology/technology/2012/10/...
I am surprised by the differences in opinions here; it's kind of like there were two completely different worlds. For example, I have an exactly opposite opinion to ps2000 [0] - single page long articles are the best and I never get lost in them, because, well, every browser has a visual position indicator, also known as "scroll bar". And I love infinite scrolling of Google Images; I'd hate to have click "next page" several times.
Similarly, I completely disagree with mcpie [1] - for what I saw, it's the "next page" button where people lose their interest. One-page articles have one huge benefit: you can skim them before reading. For example, it takes me about 15 seconds to skim an article and decide whether or not commit another 10 minutes for reading it. And for that sole reason, if I can't skim an article because someone cut it into 10 pages, I usually decide it's not worth reading.
Then there are A/B tests [2] [3] which completely blow my mind, because I feel like I'm living in a completely different world, and there is this HUGE mass of people out there, people of whom I saw or know no one, who actually click on web ads, and love paginated articles, etc. and they are the majority, they are the income source of the Internet, and I have never seen a single representant of this huge group. So I'm confused. Are we really so different?
[0] - http://news.ycombinator.com/item?id=4606143
[1] - http://news.ycombinator.com/item?id=4606388
Cite: http://www.theatlantic.com/technology/archive/2011/08/crazy-...
For regular articles, I use the Autopager extension for Firefox.
There's a few websites I run into when looking at news on my phone that have a pagination option plus a "Full article" button. I don't understand why it doesn't just load the full article. Getting to the end of the page and having to WAIT over 10 seconds to find out what the next paragraph says is frustrating.
As others point out, I rarely click past the first page when I see pagination because I hate waiting for the load files. I read a heck of a lot faster than a browser can go fetch and render a new page. I'd rather take that time to read than wait.
1) isn't used to generate page views,
2) does not insult my capacity to read more than 50 words at a time (FOOORBES!),
3) is not used on a website that loads a lot of shit on each page
I don't use it too much, but unless the site provides anchor links to save your progress like bookmarks, it's going to be a pan to save your progress. Have fun reading that 15-page New Yorker article in one go.
Pagination sucks, when there isn't a need for it. Sometimes there is.
But the plugin didn't seem to have any effect.
edit: Within minutes it has shown to be useful on other sites. Thanks for the suggestion.
Pagination is better than no navigation, this is why Google Image Search is weird to navigate. The absence of proper pagination may be even the reason Google Wave has found no friends beyond its believers.
Pagination sucks when text length per page is too short and even more when 50% of the content consists of advertisements. I avoid such websites.
That's why some people tend to select and highlight random bits of text while reading. It's not always because they're just fidgety :)
"Free" isn't really free when I'm paying for a bunch of stuff to be downloaded that I neither want nor need. And putting myself at risk from unscrupulous or incompetent ad networks.
If the goal is more ads, why not just put more ads on longer pages?
Is this fair to the publishers? No. But I'm glad a browser came up with a solution.
> P.S. Of course I didn’t divorce my wife just because she likes multipage articles. I divorced her because she types two spaces after a period. (Not really.)
Don't know what's that about though..
Perhaps ironically, but still.