SpotXchange, our friends with Otto the auto-optimizer, are expanding their offerings once again. This time they've gotten on the cost-per-view bandwagon and are launching the new format with a bit of hoopla. My major concern is, how can the industry launch a 'cost-per-view' ad format...when much of them can't even agree on what actually constitutes a view.
If you listen to Nielsen, a view is 3 or more seconds of a video play. I find that extremely simplistic and completely inaccurate, and open to interpretation, which also allows for a massive amount of fraudulent views being counted. SpotXchange defines a view as a user watching 100% of the video ad. With SpotXchange's CPV ad format, advertisers only pay if a consumer watches the ad from start to finish. So they've got a pretty good definition of a view, 100%. But that might not work all of the time and some may require some flexibility so I've got my own thoughts on what constitutes a view.
Chris` Definition of a View
Video View (in regards to online video) - A view is initiated when an Internet user clicks the play button on a video player and proceeds to watch the video play for more than 50% of its total duration.
The problem with that is many viewers might pause a video before the 50% mark and then pick up that video at a later time. If there are user accounts involved, it's not a problem. But if they switch machines during that process without a user account being attached, then the tracking problem comes into play. However, it's not as big of a problem as some might think. If the viewer watched 51% the first time and then the remaining 49% later, it still only counts as one view. Alternatively, if the viewer watches 49%, stops and resumes at another computer, it is still only one view.
If the views are part of a user account it is easy enough to track the view, pause, continue of a single piece of content as one single view.
Auto-play videos or ads are not views, even though many people in the industry would like us all to believe so. However, with an auto-play, there is no guarantee that the video was in fact viewed. This has been seen in many places with videos playing in the bottom right hand corner of a long web page with the sound off and the user never even scrolling down to where the video is playing. The only way to then be sure there is no fraud, is to not include any auto-plays as views.
However, if the video player is the first thing on the page, like at Hulu, then one could argue that clicking upon the Play thumbnail, which takes you to a page with an auto-play video, should then be considered a view. I would simply suggest that it not be an auto-play video purely for consistency.
What About Video Ads?
Now video ads are a different sort of beast. No one really clicks a video player to get to an ad, well, not exactly. It depends on the types of ads that are being played but I can make a new definition for video ads.
Video Ad View - A video ad that is shown as part of a standard Video View. The video ad is played within the video player as part of the standard video stream or as an overlay within the video player itself. The video must play for more than 50% of its duration.
Voila! See how easy that is.
What, Just in the Video Player?
Well, in order to play a video, you need a video player, do you not? Now there are video ads that play in their own video players and are not within a player that is made to show content. So then what do we consider those?
They are video, that's true and they're advertisements so they qualify as video ads. But those are mainly the video ads that auto-play outside of the viewers view now aren't they? So again, to cut down on fraud shouldn't we eliminate them from the cost-per-view calculations?
Perhaps what we need to do is make them their own sub-class of video ads or give them another name so as not to continue to confuse potential new advertisers. That's easy enough:
- Main-player Video Ads - Video ads that are shown in the main video content player
- Secondary Video Ads - Video ads that are shown in a player meant solely to show video ads
Now we've classified them both as video ads, since that's what they are, but we have broken them into their own sub-classes so that there is more transparency in their placement, their cost and their definition. That could also then be translated into the bidding and purchasing of those placements as well.
Now we can talk about cost-per-view.
Cost-per-view for Secondary Video Ads
For the places out there who aren't policing their network publishers accurately enough (I'm not pointing fingers nor need to as there are plenty of documented incidents) they'll love the cost-per-view because they'll roll in big piles of advertiser cash on these even though many of them will be out of sight, auto-play ads that are never seen by consumers.
Ideally, they would have such low CTR and brand recall attached to them that advertisers would become dissatisfied with the performance and move on to better video ads, like...
Cost-per-view for Main-player Video Ads
The great thing about this is that you know your ads are being shown to at least 50% completion and within the player where the user initiated the video content viewing. That means, theoretically, they will be seen.
There could be some expansion of the cost-per-view for Secondary Video Ads definitions as well. If there's a mouseover to start the video playing that might be considered as user initiation as well and then could be constituted as a view.
I can totally understand that many will find this too restrictive and that's why I thought it was time to publish, it's time to open the lines of communication so we can all figure it out, together.
Old Versus New
The main problem is that many people are trying to take a two-dimensional metric, impression, and turn it into what is essentially a three-dimensional one, view. For impressions it was all about eyeballs on the ad. But that doesn't quite work with video ads because there's a time element in there that wasn't being addressed. So many places, in my opinion, Nielsen included, are not taking into account this time element.
Three seconds constitutes a video view? That's about the average time it takes to flip from channel to channel on some cable boxes, does that then mean I watched every single show that I clicked past while channel surfing? If so, it might explain some TV show viewing audience numbers of late as well as why there are 400 reality shows on TV these days.
Seriously, I believe that the 50% mark should be the one used to determine a view. At that point, more than half of your creative has been shown to the viewer. If you haven't got your name or message in front of them by then, perhaps you need to look at creating some shorter ads that get to the point in a quicker fashion. I bet the user brand recall would benefit from it.
I also believe that user initiation needs to be taken into account. Granted, we didn't take that into account for impressions, but that's because the loading of the page itself was the user action that initiated the impression and again, lots of ads show where people sometimes don't see them. Many sites have four or more ads shown on a page in and around the content itself in an attempt to get the reader to 'see' the ads. In fact, ReelSEO puts an ad on articles that might be thought of as 'in-stream' since it's right smack in the middle of the test.
The equivalent of that in regards to video content is to put the ads in the main video content player so that the viewer sees them. Therefore, the ads outside of the ad player more closely resemble impressions and display ads than actual video ads, and yet are video. So perhaps to expand their appeal to video advertisers, they really do need to be given their own category. We don't want to deny them their place as video ads, but they are a very different animal than those main-player ads, aren't they?
You = View
Hey, you're still here! I totally count that as a read. I wish I could get paid on the amount of words that people read in my articles, since I'm long-winded at times, it might be financially more rewarding. Then again, like we tend to do with so many other things, you might just skip toward the end to find out what's going on and not have to read the rest. But if you read at least 50% of the words I've written (1386) then, it's a "view."
I've obviously been thinking about this for a very long time and I agree it's a sticky situation. It's definitely are more complex than the standard display ads and impressions and therefore we need a new way of looking at it. You might not agree with all of my thoughts on what constitutes a view and that's fine, let's open the floor to discussion and see if we can hash something out that we can then get others on board with. This is how progress is made, protocols are set and standards come about. Intelligent discussion about the topic at hand. Hopefully, this was an intelligent look at what makes a view and I welcome your thoughts. I know some people in the industry were talking about getting together to do this very same exercise. I jokingly said, "It would have to take place in pitch dark, so no one knew who else was there," because that is probably the only way that people might be inclined to give their uncensored thoughts on the matter.
So feel free to drop comments anonymously, I won't mind. Of course, feel free to tell me who you are or contact me in email and I'll take your thoughts on the matter into consideration as well.