Long Pages Rule!

Jared Spool

December 26th, 2006

Over on the ClickTale Blog, there’s new data to support what we’ve been saying all along: people like to scroll and, when they do scroll, they are more likely to succeed.


These statistics demonstrate that the vast majority of web designers are designing pages with scrolling, that the majority of users do scroll and that a significant portion of them scroll all the way to the page bottom. While 22% may seem low at first, it is actually quite high as many page-views are repeat views where the visitors have previously scrolled all the way to the page bottom and are already familiar with the page. In addition, visitors often find what they are looking for near the beginning of the page and may not bother scrolling further down.

Distribution of Relative Scrolling Reach From the ClickTale blog

The 100% bar clearly dominates all others and shows that 22% of the visitors scrolled all the way to the bottom.

Long pages rule!

7 Responses to “Long Pages Rule!”

  1. Ruben Timmerman Says:

    I understand (and agree to) the point that pages being long isn’t bad in itself. But where (in ClickTale’s data) do you get the point that users like to scroll, and that they are more succesful if they do so?

    The graph you picked to show, can also mean “78% of users do not scroll to the end of the page”. Does that sound scary? I find that statement just as useless as saying 22% scrolls to 100% of the page length.

    Some data that I find worth discussing is that every page length has roughly a same amount of scrollers:
    http://blog.clicktale.com/wp-content/uploads/2006/12/scrolledto90_sm.gif
    What could that mean, looking at more qualitative sources of data? Some people are “scrollers”, others are not? Page length is not a variable influencing scrolling, but page design is?

    That can also mean a lot of things. Mostly, I think it’s bad to draw any conclusions on ClickTale’s data. It offers some insights to keep in mind when analyzing other web analytics data, but every conclusion can also mean the opposite, as far as I can see.

    I have a little Dutch wiretup on this too:
    http://www.usarchy.com/2006/12/vouw-scrollen-webpagina/

  2. Jared Spool Says:

    Ruben, I didn’t mean to imply that ClickTale’s data said users liked to scroll. I only said it supports our research that says they do.

    The graph you picked to show, can also mean “78% of users do not scroll to the end of the page”. Does that sound scary? I find that statement just as useless as saying 22% scrolls to 100% of the page length.

    Scrolling to the end is not the requirement. Scrolling until you find the link which serves your objective is. What’s interesting from the graph I picked is that they scrolled at all. (And that the largest bar comes from people who scrolled all the way.)

    Some data that I find worth discussing is that every page length has roughly a same amount of scrollers:
    http://blog.clicktale.com/wp-content/uploads/2006/12/scrolledto90_sm.gif
    What could that mean, looking at more qualitative sources of data? Some people are “scrollers”, others are not? Page length is not a variable influencing scrolling, but page design is?

    Since the ClickTale data comes from many different sites, it could also suggest that some sites are more inclined to induce scrolling than others. That would support our data. (See the Scroll Stoppers section at the end of this article.)

  3. Doug Says:

    I think there is some bad methodology of deriving user intent/user accomplishment/user satisfaction from simply analyzing log results, which is what ClickTale did. So what if people scrolled? There is no mention by ClickTale that users indicated through some sort of self-initiated action that they actually found what they were looking for, or found something of tangiential interest. Were they actually successful in finding the information that they were looking for? Were they equally or more successful in finding the information at the bottom of the page versus the top of the page? The analytics do not tell us that.

    The only way you can determine that is through some combination of direct observation (and analytics are an indirect observation) and “think aloud” by the users. Conceivably some self-reporting could be used as well, but there there is a greater chance of self-reporting bias sneaking in there.

    Implying that analytics supplies any kind meaningful indication of successfully finding information on a page as perceived by the user goes a long way to invalidating the testing methodology of “think-aloud.” If analytics can read people’s minds, why bother with inviting people into a lab to actually observe them and listen to what they have to say?

    (Caveat: I think, though, that you can measure some success through anayltics when you combine it with e-commerce sales or some other user initiated action that strongly implies “this is the thing I was looking for.”)

  4. Arik Yavilevich Says:

    Doug, in our research we didn’t try to derive user’s intent, accomplishment or satisfaction. We merely tried to see how often users scroll. The result was that 76% of the page-views with a scroll-bar, were scrolled to some extent and that 22% were scrolled all the way to the bottom. We just collected some data and tried to give it as little interpretation as possible. Everyone is welcome to derive their own conclusions and debate. You are also welcome to suggest ideas for a more sophisticated research and if they are interesting enough we might carry them out.
    Inviting people to a lab gives better insight to what goes in a user’s mind, but it is also more complex and costly. Accomplishment (as you have mentioned) can often be defined algorithmically as ‘clicked on x’, ‘viewed y’ and so on. Therefore, unlike intent and satisfaction, accomplishment can be measured using passive analytics tool. Even when accomplishment can be defined algorithmically it is still difficult to analyze in a mass study as it is different for each site and page.

  5. Doug Baker Says:

    Arik,

    Thank you for your thoughtful response. As I reflect on each of our comments, I recognize my reaction is as much to Jared’s comments at the beginning of the blog entry: “there’s new data to support what we’ve been saying all along: people like to scroll and, when they do scroll, they are more likely to succeed.” While he later backs off this original statement (“I didn’t mean to imply that ClickTale’s data said users liked to scroll” — more accurately, Jared, you didn’t imply; rather, you explicitly said it), that is not much of a retraction for his original statement. After all, this blog entry was entitled in the spirit of celebration and coronation of “what we’ve been saying all along” that “Long Pages Rule!”

    Jared, I only mean to be playful here, but perhaps you should have titled the blog entry “Jared Spool Rules!” if only for the fact that it trips off the tongue so nicely. There has been more than one day when I’ve felt that “Doug Baker Rules!” But part of the reason why I originally posted was an argument to myself to carefully think about the data I collect, and how I present it to my clients. With the most sincere intentions I have presented some conclusions that I later felt was not adequately supported by the data available. Yet on a gut level I felt those conclusions were right on the mark.

    Back to the issue at hand. Jared, I’ve clearly come in on the tail end of your research in this area, so I don’t know how to interpret it. It would be helpful for me to have access to a link or two to any previous reports/entries on this issue, or any of the related items things that you’ve been saying for sometime now.

    As I think about long pages, I want to know how is it that long pages actually rule? How do longer pages accomplish usability goals of efficiency, functionality, learnability, and memorability? How do long pages vs. paginated content/function impact the cognitive load of users? What is the user experience of people who use long pages? Do long pages trump paginated pages in every one of those categories?

    Arik, your comments clearly indicate that you were able to statistically monitor a particular behavior, but did not attempt to “derive user’s intent, accomplishment or satisfaction.” For those of us who draw conclusions from data like that which you have kindly offered, we need to remember that just because people exhibit a certain behavior doesn’t mean that it is a particularly helpful or useful behavior. I’ve seen more than a few people repeatedly attempt to make broken functionality on a web site, as if simply trying again could broken link work. I’ve also seen people search for a link on a long page with a willingness to scroll to bottom and back to top again, with the pointer hand passing directly over the link they sought, but their eyes did not see. I saw the behavior — they scrolled. I also saw the congitive overload and failure of that particular task.

    Arik, I apologize if I heavy-handedly castigated analytics. A good research tool should provide certain types of interesting and hopefully valuable information, and analytics certainly does that. The analysis and presentation that some of the most sophisticated e-commerce analytics tools provide is stunning.

    A single research tool almost never gives us all the valuable information we need. While I explicitly said that analytics can’t give me everything I want to know, out of fairness (and direct experience) I will also say that usability observation observation labs also falls into that same category. They each tell me part of a much larger picture, and so I’ve learned that it takes utilizing a combination of tools. Usability engineers can use analytics as a helpful discovery tool to help identify more precisely which issues should be investigated in the lab. After all, using a lab, or even doing gorilla testing at people’s desks, is resource intensive (as you mention, Arik), so it is good to have some preliminary data that hopefully will make that investment of resources pay off.

    And certainly, Jared, you and your group have contributed significantly to the usability community, so I say with appreciation that Jared Spool does indeed rule.

  6. Information scents rule! « Simpler is better Says:

    [...] 2007 by simplerisbetter I spoke at the OZ-IA conference yesterday and added my thoughts to the Long Pages Rule! and associated discussions (such as Millisa Tarquini’s article Blasting the Myth of the Fold [...]

  7. Design by Elle » Blog Archive » 5 Rules to Writing Online Says:

    [...] Spool, 2006—Long Pages Rule!, on User Interface [...]

Add a Comment