I like JN, but these kinds of studies don't tell the full story. If you did A/B comparisons of the Yale page with and without images, which would convert better? I have a feeling that it would be the one with the picture.
The eye tracker finds the images ignored, but it isn't meant to be scrutinized, it sets a tone for the page. pictures have very different roles in an ecommerce setting vs. an aspirational info page. it isn't an apples to apples comparison. I wouldn't care so much except JN does such a good job marketing himself that studies like these become gospel in design reviews because they seem scientific even though they are built on poor assumptions.
I think JN's choice of the Yale photo was a bad example. It shows a collegial, multi-ethnic, communal environment that probably was meant to convey an image. By contrast, I've been to software sites where the stock photos of smiling people all gazing at a single laptop screen look positively asinine. Even worse are the glass and steel building photos that are meant to convey "business". Those types of images are worthless and ignored.
Does the average user of page X feel reassured by the stock photos? By demographic? By other categorizations? (web literacy, education level, income, whatever).
I agree. The photo on that page serves the same function as white space. It also helps the page feel less like a wall of text and more easily digestible. Plus, one can get the gist of the image without ever actually focusing on it.
I had similar thoughts when reading his eyetracking book. In many ways it would be more problematic if these decorative images were getting a lot of attention; it would mean they were distracting from the content.
Accepting that big "useless" images are indeed ignored (as per the Yale School of Management example), how does adding them or removing them affect the conversion rates?
To rephrase: what if they are completely ignored by the eyes but they add to the website in a way that increases the conversion rate? What if having these images made the websites be perceived as more "professional" and thus increases sales?
In short: what if eye-tracking is not the correct and only way to measure the effectiveness of these images?
I think what Jakob misses in his critique is the relationship of images and their context. Of course people spend more time looking at product photography and even avatars on a Facebook page -- it's because the images have a direct relationship with the textual content the viewer is reading. One is directly related to the other.
Does that mean photos without a direct relationship to textual information are somehow bad? I hardly think so. A well positioned, sized and selected photo contributes to the tone and message of a site. A user may not spend a lot of time looking at a decorative photo (which is a good thing) but the time they do spend looking at one (even for a split second) helps quietly generate an overall impression of quality, mood and professionalism.
Of course I don't have a paid usability consultancy business or any hard data to back that up, but that's my impression.
The thing I love about Jakob Nielsen, as opposed to Edward Tufte (though I also love Tufte's stuff) is that he never just tells you "this is how to present this information"; he backs it up with real studies with real people that track eye movements and actions and records what people learn and retain. It's more than just truthy; it's science.
I believe the broader point Nielsen is making is that "filler" photos provide nothing of value to the user, even if they are supposed to set a tone. However, detailed product shots, photos of actual employees or customers using a product can be valuable and relevant to the user. Stock photos used to "jazz up" websites are the scourge of the internet and the bane of any designers dealing with clients or internal stakeholders.
The next time someone suggests that a stock photo of smiling people huddled around a glowing computer be added to the homepage, send them a link to JN's post and suggest an actual product shot instead.
The eye tracker finds the images ignored, but it isn't meant to be scrutinized, it sets a tone for the page. pictures have very different roles in an ecommerce setting vs. an aspirational info page. it isn't an apples to apples comparison. I wouldn't care so much except JN does such a good job marketing himself that studies like these become gospel in design reviews because they seem scientific even though they are built on poor assumptions.