The idea would be that advertisers would pay for, say, an eight or ten-second view -- and perhaps longer -- and would be charged less if the engagement did not last that long. Alternatively, advertisers could buy a set amount of time and have each ad contribute to it. For example, brand x went for one hundred hours of exposure -- to pluck a figure out of the air -- each ad that appears for longer than, say, eight or ten seconds would count toward that final figure.
I think it is fair to say that this was foreseen to happen because in isolated incidents, it is already happening. The FT.com charges on a time-based system, for example, and it is usually seen as a success story for where the industry is likely to go.
So it was interesting today to see that the latest Meetrics figures show that viewability in the UK is holding steady at 42%, or 58% for banners. It's pretty much the same story in Q2 and Q1 but this time Meetrics pointed out that at 26 seconds, UK ads generally remain in view for nearly five seconds longer than the international average.
The news this morning coincided with an interesting talk I just had with Andy Evans, CMO at Sovrn. He developed technology several years ago to tell whether someone was actively engaging with a page, so publishers could prove to advertisers that someone was there to be advertised to on in a unit that was in view. The technology can also count up how long an ad remains in view by looking for signs of activity. The technology basically looks for signs that someone is still in the same tab and engaged with the content by looking for signs of activity.
The premise is that there is no point paying extra for time if that person has since moved on and is engaged with content in another tab or has popped off to put on the kettle.
This leads Evans to wonder aloud whether we need to think of a maximum or optimum time that an ad should be delivered for. "We have the floor metric of half an ads pixels for a second but what about a ceiling measurement," he suggested. "Our research shows that brands get very little additional impact once an ad has been viewable for twenty five seconds or more. So there is no point in having the same advert for that amount of time. If you reach that point, we'd suggest the ad is reloaded for another brand to be seen in. Advertisers aren't getting anything more out of being on view and so publishers could receive extra advertising opportunities which they can sell to brands as being demonstrable proven to be in front of someone who's already highly engaged in the content."
So how about it? We've got a bare minimum, which even the IAB UK agree is pretty much just a bare minimum ad hygiene metric, so are we going to need a ceiling? If so, who will set it?
My thoughts would be that advertisers will take the lead in this and discover that, say, an eight-second exposure gives just enough engagement lift to be worthwhile, and that lift drops off after a given amount of time. This will prompt publishers to offer time-based slots in ad units that will reload once the desired exposure has been given.
This is pretty much where Evans see things headed too, but points out that we would need an update to Open RTB to make this possible in programmatic. This would be unlikely to both happen and be rolled out for at least another 18 months.
Sean, it all depends on the kind of ad and its placement. While the IAB's "standard" is clearly unacceptable, any one definition of "viewability" is also doomed to failure. For, example, suppose you run a banner ad, perhaps, based on time-on-screen, one might accept a 10-second rule as representing an effective chance to be seen. But what if it's a 15-second video commercial---does 10-seconds cut it? Of course not. Time spent, must be evaluated with other measures---verified ad recall, one would think---to develop norms for a wide range of ads and clutter situations. It's a difficult and expensive task and I don't expect to see any movement in this direction until the digital media titans finally make a concerted move to clean their swamp.