Projects Too Much From Too Little
By ALAN JACOBSON
Brass Tacks Design
BEFORE CRITIQUING THE Stanford-Poynter eyetrack study, I'd like to express my appreciation and respect for the work done at Poynter. I'm particularly grateful for the pioneering work published in "Eyes on the News," which has been quite influential in all the work we've performed for clients since it was published.
That said, I'm particularly distressed by the equally influential, but possibly misleading, findings in your recently published research on the habits of people who use news websites. In particular, I don't believe your research supports your conclusions that (and I quote from your report)
- "The provider's first chance to engage the reader is through text."
- "Where do eyes go initially after firing up the first screen full of online news? To text, most likely. Not to photos or graphics, as you might expect."
- "Text seems favored over artwork for front-page attention."
My goal in bringing this to your attention is not merely to be critical, but to advance your research work so that it can be helpful to those of us in the field who are developing sites for newspapers. Poynter is one of the few resources for such help and we in the field depend upon Poynter to better serve our customers. My overriding concern is that well-meaning web content news providers may take your findings at face value, and alter their sites in ways that may be counter-productive.
At my company, Brass Tacks Design, we're big believers in research, making extensive use of both quantitative and qualitative studies in all work.
We've used research to save dying newspapers and bolster the circulation of others. But as good journalists, we are skeptical of any research that doesn't ring true. This skepticism took us to your website to examine your raw data. Here's what we found:
Who Was Tested and How Many Were Tested?
- Subjects were drawn from the limited universe of "frequent Internet news readers" rather than the Internet-using public at large. To be precise, it may have been wise to qualify your conclusions by stating that your findings were limited to this narrowly focused group.
- All results were based on tests of 67 people - hardly a large enough number to be sufficient to draw any conclusions.
- Of these 67 people, only 14 were examined for their behavior in front page attention. Of these, fewer than 7 pages viewed by these subjects contained a single photo. Again, this number seems stunningly small to make the assumption that text is preferred to photos considering how few photos were available.
- Subjects viewed sites familiar to them. But virtually all the graphics at these sites are static and never changing. Isn't it reasonable to assume that subjects will look first to the dynamic content, which in this case was limited almost exclusively to text? (I believe "Eyes on the
- News" found this to be true with regard to above-the-nameplate teasers.)
What the Subjects Saw
- Of the three sample pages published at the Poynter site, only two had photos and each of these pages was limited to a single photo. Neither of these photos contained an image of a person. Isn't it reasonable to assume that images of people are more likely to draw attention than images of things, and that multiple images are more likely to attract attention than a single image?
- Most web pages are text based, with only the front pages of sites offering the lion's share of images. But the front pages viewed by subjects were almost as text heavy as any other page. It doesn't seem as if a reasonable number of images were viewed by any subjects on any pages, based on the published samples.
- Unlike the front pages of most well-designed newspapers, which use roughly 50% of their space for images, most of the sample online pages were covered with text. Isn't it reasonable to assume that people will look at text first when it occupies most of the screen?
Although the sample pages were typical of contemporary online news design, they were hardly examples of good online design. To be precise, it may have been wise to qualify your conclusions by stating that your findings were limited to typical online news pages rather the most compelling information experience the Web can offer.
Ultimately, these criticisms are moot if you were merely trying to measure how a few atypical users view current news sites, whose design sophistication remains less than fully developed. But I suspect you wanted to achieve more.
To its credit, I think of Poynter as an institution that points to the future and helps provide direction to the industry. This study merely shows where a few people are now, not where most of us need to go. However, the newspaper industry is likely to react to your conclusions and alter their sites accordingly.
For instance, we're currently working with The Hartford Courant to improve the presentation of their online content. People responsible for their content have cited your findings as a potential source of valuable information, but upon inspection have come to agree with us that your conclusions may be counterproductive to our work.
For this reason, I hope you'll reconsider the context for your conclusions. Better still, I hope you'll consider further research based upon a broader universe of Internet users viewing pages designed to delight the eye in addition to informing the mind. I'll bet you dollars to donuts they'll go to the images first.
Thanks again for your time and interest. Let me know if I can help in any way.
-- Alan Jacobson is a principal in Brass Tacks Design, which provides editorial, marketing, research, design, and technical support to publications worldwide. Projects in 2000 include a website redesign for InfiNet, the leading provider of Internet services to newspapers; a web and print redesign for The Hartford Courant's classifieds; and redesigns of the Norfolk Virginian-Pilot, Omaha World-Herald, the Chicago, Ill., Daily Herald, the Everett, Wash., Herald, and Crain's Detroit Business.