I've been reluctant to discuss one of the findings from our eyetracking research because the conclusion is that unethical design pays off.
In 1997, I chose to suppress a similar finding: users tend to click on banner ads that look like dialog boxes, complete with fake OK and Cancel buttons. Of course, instead of being an actual system message — such as "Your Internet Connection Is Not Optimized" — the banner is just a picture of a dialog box, and clicking its close box doesn't dismiss it, but rather takes users to the advertiser's site. Deceptive, unethical, and #3 among the most-hated advertising techniques. Still, fake dialog boxes got many more clicks than regular banners, which users had already started to ignore in 1997.
Can't Hide Usability Findings
After much soul-searching, I've now decided to take a different approach and publish our new findings, despite their ethical implications. In reality, it's not possible to suppress research results because anybody who bothers to run the study will get the same findings. There are no secrets of usability any more than there are secrets of astronomy. If you point your telescope at Saturn, you will see that it has rings. And, if you conduct a series of usability studies, you will discover the same insights as we do — assuming you employ the correct methodology.
Many people without a grounding in behavioral user-research principles use bogus methodology and thus get misleading findings. Poor methodology is especially common for eyetracking studies, and thus most published studies in this area are wrong.
For example, unskilled researchers often ask users to simply look at a page, rather than have them encounter it as part of a task flow. Users naturally look at things differently depending on the context. For example, if you want to know how users look at the elements of a form, you can't just present the form on a stand-alone page and ask them to fill it out. Instead, you have to present the form in the context of a meaningful task that they might attempt in the real world. That is, users should encounter the form in response to particular actions, such as deciding to check out from an e-commerce site.
Still, even though most eyetracking studies are misleading, some studies do produce valid results. Trying to keep any results a secret is thus a lost cause.
Most of our eyetracking findings on Web advertising present no ethical dilemmas. For example, we know that there are 3 design elements that are most effective at attracting eyeballs:
Cleavage and other "private" body parts
I don't have a problem presenting details of these findings in the seminar on Top Web UX Design Guidelines; it's important that all Web designers know where users look on pages.
The most prominent result from the new eyetracking studies is not actually new. We simply confirmed for the umpteenth time that banner blindness is real. Users almost never look at anything that looks like an advertisement, whether or not it's actually an ad. (Indeed, banner blindness is moving beyond the online realm, for example into ballot design.)
On hundreds of pages, users didn't fixate on ads. The following heatmaps show three examples that cover a range of user engagement with the content: quick scanning, partial reading, and thorough reading. Scanning is more common than reading, but users will sometimes dig into an article if they really care about it.
Heatmaps from eyetracking studies: The areas where users looked the most are colored red; the yellow areas indicate fewer views, followed by the least-viewed blue areas. Gray areas didn't attract any fixations. Green boxes were drawn on top of the images after the study to highlight the advertisements.
At all levels of user engagement, the finding is the same regarding banners (outlined with green boxes in the above illustration): almost no fixations within advertisements. If users are looking for a quick fact, they want to get done and aren't diverted by banners; and if users are engrossed in a story, they're not going to look away from the content.
The heatmaps also show how users don't fixate within design elements that resemble ads, even if they aren't ads (and thus aren't shown within green boxes above).
Even when we did record a fixation within a banner, users typically didn't engage with the advertisement. Often, users didn't even see the advertiser's logo or name, even when they glanced at one or two design elements elsewhere inside an ad.
The following video clips show a gaze replay of one user's eye movements while looking for advice on how to invest for retirement. (The moving blue dot shows where the user is looking.) The page contains an ad for retirement accounts at Fidelity Investments, a site that offers good advice on the target topic and might therefore help users who click the ad.
Real-time speed (19 seconds — and yes, the eye really does move this fast when users view web pages):
Slow-motion replay (1 minute — slowing things down is the only way to really follow users' gaze patterns):
As the replay shows, the user did fixate once within the ad, but at that moment, the ad is obscured by a pull-down menu. In reality, the user couldn't see the message; the fixation was clearly a mistake that occurred while she was trying to reacquire the menu after briefly looking away from the screen. All of this occurs so quickly that you probably need to review the slow-motion replay to follow the action. (This is typical for eyetracking: the eye moves so fast that our best insights come from watching slow-motion replays.)
(Several readers have asked whether banner blindness extends to search engine ads. It doesn't: text ads on a SERP get a decent number of fixations. The other exception is classified ads. Finally, it's possible that commercials that are embedded within a video stream get viewed; we haven't researched this yet. So there are either 2 or 3 exceptions to the general rule that users avoid looking at ads on websites.)
The Fourth, Unethical, Path to Ad Fixations
In addition to the three main design elements that occasionally attract fixations in online ads, we discovered a fourth approach that breaks one of publishing's main ethical principles by making the ad look like content:
The more an ad looks like a native site component, the more users will look at it.
Not only should the ad look like the site's other design elements, it should appear to be part of the specific page section in which it's displayed.
This overtly violates publishing's principle of separating "church and state" — that is, the distinction between editorial content and paid advertisements should always be clear. Reputable newspapers don't allow advertisers to mimic their branded typefaces or other layout elements. But, to maximize fixations, that's exactly what you should do in a Web ad.
A specific ad may or may not be ethical, depending on how closely it masquerades as content. I caution against going too far, because it can backfire and mislead users. Unethical ads will get you more fixations, but ethical business practices will attract more loyal customers in the long run.
Now the truth is out. As far as I'm concerned, speaking the truth is my highest ethical calling, and it's better that the facts be known to everyone than that they remain a secret abused by a few.
Ultimately, the fact that online ads get viewed more when they match surrounding content is a strike against the tendency to build advertising networks. If advertising spots are simply auctioned off, then you can't design an optimized ad for each placement.
When you advertise through an advertising network, your ads will get fewer fixations than if you contract directly with the publisher for a specific placement and design your creative to fit that spot. As a result, you should bid less for network ads than for customized ads that you place yourself.