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 homepage — here's an Eyetools Heatmap of a group of 19 new visitors viewing the new page and 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.

Advertising

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.

Comments

[Bud Gibson]

So, I can have an SEO-aware copywriter crank out 10 pages for $2000. I can get an SEO analysis of what it takes to drive people to my site for $1000.

Analyzing one page with this technique costs me $1900, and I only get 19 subjects.

The SEO copywriter's claim is that he will INCREASE my conversion rate.

I like the analysis, but I would like to see the following story. We saw web page X, before testing, web page X was getting a conversion rate of Y (say what conversion means). Our analysis led to a conversion rate of Z, justifying the price of W.

Right now, your pricing is targeted at high sales sites (that's what would potentially justify the cost). I might be willing to pay $100 to get a page analyzed for a site. Are you able to take your research, extract some heuristics, put those in an algorithm, and turn it into a higher volume package?

[Greg]

Bud, thanks for your comments here. It emphasizes that I need to tie things to revenue and can't just talk about interesting findings (darn... :-). I've put up a new post at ( http://blog.eyetools.net/eyetools_research/2005/02/is_the_washingt.html ) that talks about the Washington Post revenue impact.

Also, our clients have had:
- 60% lift in PAID conversions (on a sales site --MarketingSherpa-- selling $150 reports -- as reported by Anne Holland, editor of Marketing Sherpa.

- 10x times increase in homepage click-throughs ( http://blog.eyetools.net/eyetools_research/2005/02/how_eyetracking.html )

- 47% lift in ad click-throughs on the main sponsored ad of a niche publishing site.

Eyetools' services are geared toward websites that generate revenue, because we can help optimize that process so much, and it's not $100, but it _is_ $1000 for a homepage, which in the bigger picture isn't expensive for most companies.

[David Temple]

Are you kidding me? This is probably the most cost effective analytical tool yet. I would like to see a larger sampling but maybe 19 is enough. At what point do you get diminishing returns? Anyway I still think it's a great tool. Let me know when you adjust pricing for blogs. dt

[Website Design]

It's a great blog,I am impressed by your writing style.Great tips and resources.Thank you Sir.