Viewability is an issue that arises time and time again in the field of digital advertising, particularly in mobile.

What makes an ad viewable has at times been up for debate and revision, and has been used by some detractors as an argument against over-investment in digital. Numerous working groups and adtech companies have championed and made strides toward greater viewability. But just because an ad is viewable, does that mean it’s having the intended effects? Do viewers remember it?

A new study by adtech company Kargo tracked the eye movements of 500 participants when using desktop and mobile devices, analyzing how people reacted to several different ad formats, including desktop banners, ad units in mobile games and Instagram ads. It also surveyed respondents after, testing them with questions about their recall and impressions of the ads.

Banner ads within games have potential to reach many users – according to GlobalWebIndex, 83% of Canadians engage in “casual” mobile gaming. Generally speaking, gaming banners are cheaper buys and can be purchased programmatically, but the study rules that these ads might be more-or-less invisible to the users they’re reaching.

Most gaming banners cover an average of 5.25% of a phone’s surface area, and because of their placement are considered 90% viewable. However, during the session, the ad was not being looked at most of the time. Among participants, the ad unit’s average share of attention during the session was 1.5%. To add to that, ad recall was extremely low, with participants struggling to name the companies or products featured during the game. All in all, these banners generated the lowest rate of ad effectiveness of all platforms at only 1.3%.

So what happens for larger screens? On desktop computers, sidebar ads were displayed, taking up 9.8% of the page. While it was in-view approximately 80% of the time (considered to be “highly viewable” by the MRC), the attention it drew wasn’t much higher than mobile gaming ads. Only 1.9% of session time was spent looking at the ads, making them the second-least effective ad formats of all those measured at 2%.

Instagram ads proved that success can be had on smaller screens; the ads proved to be 8.3x more effective than gaming ads and 5.4x more effective than desktop. The ads may be on smaller screens, but at larger dimensions occupy a much larger portion of the screen (between 56.6% and 62.4%). While the share of session attention is very low (1.9%), that is also because users tend to spend much more time on Instagram than other platforms.

Due largely to recall and retention, Instagram’s effectiveness score came to 10.8%.