Stay marketing and tech-savvy. Get the latest in martech - subscribe to MarTech Today.
Eye-tracking study: Web ads need to viewable for at least 14 seconds to matter
New data shows that the current standard of ‘viewability’ -- half the ad on-screen for a second -- is far too low to have much value for advertisers.
According to the Internet Advertising Bureau (IAB) and the Media Rating Council, an online ad is deemed “viewable” if half of its pixels appear on the screen for at least one second.
It’s always been a curious measurement, given that no human could be expected to respond to seeing half an ad for a second.
Now, a recently-released eye-tracking study shows how inadequate this viewability threshold is for any advertiser wanting some impact on a website viewer.
The study, conducted by biometric tracking firm Sticky, ad tech firm InSkin Media, and market research firm Research Now Group, finds that viewers need about a second of “gazing” at an ad to have even the smallest amount of ad recall. In the study’s terminology, “gazing” means looking at an ad long enough and with enough attention to actually have the possibility of remembering it.
But, to get a second of “gazing,” the ad needs to be on-screen for an average of at least 26 seconds. That’s still just 50 percent of the ad, at minimum.
Fourteen seconds of viewability is required for gazing from zero to one second, on average.
Let’s say that again: Unless the ad is on the screen for at least 14 seconds, it has little or no value to the advertiser. Preferably, 26 seconds.
To generate at least two seconds of gaze time, the ad needs to be viewable for at least 33 seconds; for three or more seconds, 37 seconds. The median time a “viewable” ad is gazed at: 0.7 seconds.
The study found that a quarter of ads that meet the viewability standard are never looked at (“gazed”) in any way that is meaningful. A third have a gaze time of under a second, and only 42 percent are looked at for at least a second.
Gaze time results in levels of ad recall that differ by ad format. As shown in the report’s bar graph below:
For a takeover ad (which is the study’s term for ads that surround the periphery of a web page, not ads that take over the screen), the ad needs to be seen (gaze time) for 7.5 seconds for 52 percent of viewers to recall it.
Smaller ad formats, such as mid-page units (MPUs), need smaller on-screen or “opportunity to be seen” times, in order to have any gaze times that result in ad recall, or impact. But, the study points out, these gaze times are very low — such as 0.7 seconds for 18 percent recall for MPUs, or 0.9 seconds of gaze time for 21 percent recall for half-page ads — so the overall impact is low.
As the study noted, ad campaigns need to be assessed in three stages:
“… did the ad have the opportunity to be seen, was it actually looked at and what was the impact? It should be judged and optimised against the last stage (impact) but the focus on viewability [i.e, the viewability standard] means campaigns are increasingly optimised against the first stage (the opportunity) which can be counter-productive to maximising impact.”
Then there’s ad clutter. The report found that ad recall for three of the formats (excluding takeovers) can be heavily affected by the presence of other ads on the web page. Clutter can cause ad recall decreases of an average of 20 percent. The report didn’t specify how many other ads or what kinds constitute “clutter,” just that it was the presence of one or more other ads.
The study looked at desktop web sites and ads, using actual sites and ad campaigns, although Sticky declined to name the sites or advertisers. The ads themselves were all static, with no video ads included.
The research surveyed 4279 adults in the UK — where InSkin Media is based — about brand recall. Of those nearly 4300 users, 696 employed Sticky’s webcam-based eye-tracking technology in their homes, so the gaze time and other stats apply only to those 696. Technology from Moat was employed to monitor ad viewability.
(10.7: This story was updated to reflect Sticky’s revision of its original indication to Marketing Land, which had been that a “viewable” ad in its terms was 100 percent of the ad, not IAB/MRC’s 50 percent. Today, they corrected their original statement to say their “viewability” was, indeed, only 50 percent.)