Is the Washington Post Wasting Money on the Bottom Half of Their Homepage?

Comments Off on Is the Washington Post Wasting Money on the Bottom Half of Their Homepage?

Washingtonpost_eyetools_new_homepage_1My recent analysis of Washington Post’s homepage points out that the weak design of the bottom half of their page reduces reading. I received a comment urging me to tie the findings to actual revenue impact, so here goes!

Content costs money for the Washington Post to provide — they have to pay their reporters to generate it, their IT people to put it on the site, and their ISP for bandwidth usage.

If people decide not to read a piece of news because they’re not interested, then that’s fine. But if potentially-interested people don’t see entire sections of content because of a design flaw, then they lose money, and their brand is negatively impacted — it will appear to people that the site offers less content than it actually does, and they will spend less time on the site.

Do they realize they are suffering financially? Probably not… they probably think it’s “normal” to have low click-throughs from content below the fold. Just looking at their click logs, they wouldn’t be able to realize that potentially-interested people never received the opportunity to click.

February 24th 2005 Uncategorized

The New Washington Post Homepage Design — an Eyetools Eyetracking Analysis

Comments Off on The New Washington Post Homepage Design — an Eyetools Eyetracking Analysis

Eyetools Heatmap of people reading
the new Washington Post homepage
(launched February 15, 2005)


Last week, the Washington Post announced a new homepag, and here is the data showing how it performs: an Eyetools Heatmap of a group of 19 new visitors viewing the new page; along with what we can learn from its design.

Fast summary:

  • Top half of page — good readable design.
  • Bottom half of page — bad example of line-spacing and white-space discourages reading.

Note: This entry is not an exhaustive report, instead it’s a quick sample of things! Despite that, this entry is still too long by most weblog standards — there’s always so much to be said about the data… alas.

Learn from Washington Post’s successes

  • Main content area — good readable design. It is heavily viewed and read (more so than some other news sites). Good use of line-spacing and white-space. People even scroll. Job well done!

Learn from Washington Post’s mistakes

  • Bottom half of page — ineffective line-height spacing and lack of white-space reduce reading. Most of the content is being missed and there is no consistent guidance of eyes to section headings.

    Opportunities to communicate value to visitors is greatly reduced in this area. We’ve seen other websites do a better job.


The ads changed for each person, so numbers are averaged and are not tied to specific creatives. The numbers reported are pretty typical, in our experience, for sites of this kind. It’s worth noting that we measure what is happening with advertising and photos with statistics rather than relying on the heatmap (email me if you want to know why, and maybe I’ll write-up an entry about it).

  • Masthead ad — 32% of visitors saw it, looking at it on average 1.2 times, for a total time of 0.7 seconds. On average, it was seen 28.5 seconds after the page loaded.
  • Cube ad — 79% of visitors saw it, looking at it on average 2 times (each time for approximately 1 second), for a total time of 2.1 seconds. On average, it was seen 18.7 seconds after the page loaded.
    • Clearly there is a more we could say about advertising effectiveness, but not today — the main point of this entry is about design.

      A final note about this type of research

      Getting this type of data isn’t difficult or expensive. Giving direct visual eyetracking feedback to designers is a great thing because it completes the creative design loop — designers already utilize visual design, fonts, background colors, and spacing with the intention of effectively guiding visitors’ eyes, and eyetracking data introduces feedback into that system.

    February 23rd 2005 Uncategorized

    How Eyetracking Helps Website Redesigns: an Eyetools Case-Study

    Comments Off on How Eyetracking Helps Website Redesigns: an Eyetools Case-Study
    Client realizes 10x increase in click-throughs after redesigning a homepage guided by three Eyetools mini-studies.
    Test #1 of Existing (old) Page Test #2 of Prototypes Test #3 of (new) “Best practices” Page Launch


    • Most content wasn’t seen.
    • Most people preferred the competition.
    • 3 prototypes generated based on data from test #1.
    • Prototypes were quickly generated as images (can test image-mock-ups or HTML).


    • Most successful elements of each prototype from test #2 pulled together into a final prototype.
    • Most content now seen.
    • Most preferred this website over competitor.
    • An entire column of paid-content was added, and was economically successful.
    • Click-throughs increased by 10x times.
    • Satisfaction increased by double.
    • Faster redesign — entire process took less than 45 days from start to launch (including the three tests) because all the parties involved agreed on what was working and what needed fixing.

    A PDF of this homepage redesign using Eyetools eyetracking can be found here (4 pages, 905 K, images of a site before/after, plus Eyetools heatmaps).

    Guiding a website redesign with Eyetools eyetracking data offers these advantages:

    • Objective, visual feedback to your designers and copywriters about what works and what doesn’t,
    • Test before launching — don’t launch a mistake that loses traffic,
    • Remove “opinion-based discussions” (and guess-work) about what is seen.

    February 22nd 2005 Uncategorized

    Blog Analysis and Optimization With Eyetracking — Or "Oh, Oh, That Blog’s Writing Needs Fixing"

    Comments Off on Blog Analysis and Optimization With Eyetracking — Or "Oh, Oh, That Blog’s Writing Needs Fixing"

    A historical note regarding this post

    A special note about this blog entry and the first comment on it from ‘Micro Persuasion’ (Steve Rubel): his comment, made back in February 2005, and the resulting traffic that came from his site are in many respects responsible for all of the blogging that occurred in 2005 by Eyetools, and by extension, most of the buzz that got generated around eyetracking as a field in 2005. The spike in traffic based on his one comment and his link to our site made us realize that we had something special we could do with our blog. That year we posted as many interesting eye tracking heatmaps as we could think of and wrote about “eyetracking heatmapping.” Now in 2009, “heatmapping” as a term seems standard, but back then it didn’t exist, and there wasn’t much buzz or discussion about eyetracking back then in the marketing world. 2005 was a great year for Eyetools, and Steve Rubel deserves a lot of credit for starting that, and so it is with a special feeling that we express gratitude to Steve Rubel and his blog. If you haven’t ever read
    Micro Persuasion,
    then we highly recommend it.

    — Greg Edwards, founder, and CTO in 2005, now CEO since 2007

    March 3, 2009 at 05:57 PM

    This is what happens when you have easy access to eyetracking testing: you start wondering how well you’re doing writing about it… and after wondering very briefly, you test it and find that it

    needs improvement.


    An Eyetools Heatmap of people reading one of our blogs


    Our first blog: All that red at the top shows that everyone reads our title and the first 1-line paragraph, but that

    they start dropping off in the middle of the “story”


    (guess the writing wasn’t as good as I thought 🙂



    “statistics” section held good information, but only 60% of people read it.

    Can I do better?


    the “punch line” certainly is getting wasted down there

    (50% and mostly skimming).

    No big surprise. Shouldn’t have been so egotistical to believe that I could actually hold people’s attention ALL the way to the end

    , but that final point was important, so I’m moving it.

    So, I’m going to tweak the text of the original blog and post it as a new one. I’ve still got a bunch of people who will be coming through for this study so I’ll test that as well!

    Crossing fingers… I’m going to go write the
    new version

    February 11th 2005 Uncategorized

    Eyetracking a Navigation Bar — How Many Elements Are Read? Well, It Depends…

    Comments Off on Eyetracking a Navigation Bar — How Many Elements Are Read? Well, It Depends…

    A historical note regarding this post

    Back in 2005, people were talking about how “important it was to have a blog” and so we started one, writing this first entry you see below, which was originally posted on February 2, 2005. But as we posted it, we realized that we (along with most of the world at the time) didn’t have any particularly good idea about how blogs should be written. And so we decided to do “eyetracking analysis” on it (it wasn’t yet popularly called “eyetracking heatmapping” back then, since the term “heatmapping” was only used by us with clients and didn’t start to grab hold in the marketing world until July of 2005 when
    Jeanniey Mullen wrote about us in AdAge
    (thank you, Jeanniey)). You can see the blog entry where we
    posted the eyetracking results of how people read this entry
    (which back then didn’t have this comment on it, of course). And, if you check out the
    on that blog entry, you will see the original comment and link by Steve Rubel at Micro Persuasion that truly launched this blog into the public eye, and resulted in eyetracking becoming a common word and tool in the marketing world. Now, jump back to 2005, and read the blog entry below…

    — Greg Edwards, founder, and CTO in 2005, now CEO since 2007

    This right-hand navigation bar was the final motivator that got this blog started. When I first saw it, it wasn’t doing so great.

    People interacted with this nav differently depending on other page elements.


    It was on the (old) San Francisco Police Department website, which, some might say, was "cluttered" before the homepage redesign. I know that context plays a big part in how web elements are viewed, but when I saw it in a follow-up study on the (new) re-designed website, I at first thought it was a completely different right nav!

    Old Page Design New Page Design



    Eyetools Heatmaps™ showing group viewing trends on each web page


    Eyetools Heatmap Legend


    The behavior on these two identical navs on two different pages was strikingly (and statistically significantly) different: the nav on the new page was clicked by 64% of our test participants as opposed to only 14% on the old website. People looked at the new site’s right nav longer, more often, and read more — despite there being no change to the design of the right navigation bar at all.

    The moral of the story: A change on one part of the page can impact other, unrelated elements on the page.

    The right navigation bar was used completely differently on the new re-designed website because the content to the left of it changed.

    February 3rd 2005 Uncategorized

    Oh, the Joy!

    Comments Off on Oh, the Joy!

    I love what I do and I believe that understanding visual interaction data (what people look at) can really help people design better websites… yet, other than talking directly with clients, I haven’t had an opportunity to talk about all the really cool things that we discover — until now!

    This weblog is going to be my vehicle to talk about surprising data I’ve seen, “rules-of-thumb” that we discover and interesting directions in the research.

    I’ve got some catching up to do (we’ve done a lot of research so far!) so I imagine that the posts on this site will be relatively frequent once I get a hang of this.

    So… here goes!


    February 2nd 2005 Uncategorized