Facebook To Roll Out New Video Metrics
- by Mark Walsh @markfwal, May 5, 2014
Facebook on
Monday announced new metrics for videos (paid and organic) uploaded to brand pages on the social network.
The step comes a couple of months after the social network launched premium video advertising on the site with 15-second spots that autoplay and run in the newsfeed.
Currently, page owners on Facebook can only see information such as the number of views a video has received. But with the new metrics rolling out in the next few weeks, marketers will be able to see views, unique views, the average duration of videos viewed and audience retention.
“These new metrics are designed to help you learn what’s resonating with people and determine how to more effectively create and promote your videos on Facebook,” stated a company blog post today.
Video views are defined as views of three seconds or more, and unique views indicates the number of people who watched a video for three seconds or more. Marketers can also see the portion of views driven by paid advertising.
In terms of retention, the Page Insights dashboard will graph viewership throughout a video, showing where audience interest spiked or tailed off. Marketers will also be able to see the number of views that reached certain points in a video: 25%, 50%, 75%, 95% and 100%.
Facebook will highlight views to 95%, showing that stat in both the Page Insights and Ads Reporting tools. Users will be able to see how many people made it to the final seconds of an ad.
To paraphrase Yogi Berra: What's the metric du jour? I don't know, but they got it every day!
So it's basically a reach metric and not an 'average minute viewed' so not directly comparable to TV ratings.
The graph shows seconds viewed (top left).
This is more of an exact science, right? Can one really gauge TV commercial viewing? Last I checked, most people get up to go to the bathroom, get a drink or fast forward through the commercials if they have the show recorded.
Bill, most FB video would be video on a computer, mobile or tablet - not a traditional TV. While I agree that most people would have a 'nature break' during commercials when watching TV ... most people do not have 'nature breaks' EVERY commercial break. Here in Australia the TV ratings system attributes no credit for FFWD during commercials (as there is no sound and we rely on audio matching), so while very real, it is not a major issue in a commercial sense. However, TVs are different to the other devices (especially computers) in that when a commercial plays you can quickly jump to other content (e.g. another tab on your browser, another application, minimise the video player etc.). The big issue is that server-side data simply has no idea that while it is streaming content, there might be no-one consuming that content on the device. In essence it reflects that maximum potential audience, akin to a reach estimate.
The plain fact is that electronic measurements of TV/video usage have no way of knowing whether anyone is present and attentive on a second by second or, for that matter, a minute by minute basis. As has been said, above, all that the data tells you is that something might have been seen. Our best estimate is that approximately 40-45% of the reported average minute TV "commercial minute audience" had already stopped watching the show before the break began, has left the room during the break or, if present, is totally inattentive. Moreover, there are sizable variations between highly involving and less engaging program genres that remain to be uncovered. Similar evaluations need to be made for online videos and, especially, their commercial content.
I take you point Ed, but I think your 40s is a low-ball. Here in Australia we measure compliance using co-incidental surveys and the range from low 80s and to the low 90s. That is 80-90% of the time they are in the room with the TV on and the sound up and logged in. When we look at the prime-time viewing data (and I must admit it has been a few years since I have repeated the exercise) the average 'turn-off' during ad breaks was around 6%. It was closer to 10% in the middle-break and low single digits in the first/last in break. I suspect that may have slipped but if we were pessimistic and said that compliance was only 80% and that the 20% were INCORRECTLY logged in (it's nowhere near that) and crediting a false rating, and then that 10% of those left the room on average in an ad-break we're still talking low 70% being picked up by the system as logged in viewers. Remember this is a 'worst-case scenario'. I agree that this does not take into account. However, that is really the job of the AD to capture attention - not the programme. Neuro-analysis here has shown that sound is important (esp. in the first few seconds) to hold attention. I have problems with the logic that a programme can deliver 10m viewers and that can drop to 7m (my %age) or even 4.5m (your %age), and that the broadcaster has to take a bath on its rates because of crappy ads. Surely the solution is to make better creative. Use sound better is a good start.