Hacker News Reader

Comment by codehusker | Hacker News Reader

On my programmer hand, I can see myself making a similar technical choice. We have auto-playing videos, we shouldn't count the views less than 3 seconds because that wasn't really an intentional view. But that means we should also exclude them from the overall view count. I don't know if FB did that as well.

On my shareholder hand, this seems slightly like fraud.

On my advertiser foot, this seems slightly like fraud.

View on HN

See just get rid of the auto-playing videos problem solved, and make the world a better place while you're at it.

I bet Facebook gets a decent enough amount from people who scroll, look away for a second, video starts playing, then user pauses video, but not before the 3 second mark.

What really should be done is getting rid of auto playing videos at the HTML spec level. While there are legitimate uses for autoplaying videos (YouTube for example), it has too much potential for abuse by ad networks.

Advertisers will find a workaround and it will inevitably be much worse for the user. In the following talk, someone from the Google Chrome team describes how, when they tried to not autoplay videos on Android, video advertisers worked around the restriction by compiling video codecs to JavaScript and then rendering their video-like ads to an HTML canvas.

https://youtu.be/wQGa_6CRc9I?t=23m52s

That's where NoScript comes in

Or blocking 3rd party scrips with uBlock.

Don't ad networks charge for ad size as well? (or they should start charging for this - would cut on abuse)

If your ad is > 50k then you're paying a fee per display, even if you're "pay per click"

I disagree. I see small auto playing videos all the time in reviews or on feature pages, explaining little concepts or details that are much easier to grasp in a few seconds of (silent) video than in a paragraph or two.

The alternative is going back to large, low quality and bandwidth-heavy gifs, which isn't the better option in my opinion.

That's a legitimate use of video, but not of autoplaying video.

I'm not sure, I actually enjoy the autoplaying videos on FB. I probably wouldn't click 99% of videos I see but silently autoplaying it I get to see a few seconds without having to interact and then decide if I want to keep scrolling or click on it to get sound.

> hat really should be done is getting rid of auto playing videos at the HTML spec level. While there are legitimate uses for autoplaying videos (YouTube for example), it has too much potential for abuse by ad networks.

Most of Facebook's video traffic is in their mobile app, getting rid of it in the HTML spec level wouldn't make a difference for FB.

Wait... get rid of auto playing videos at the spec level? I'd rather have the autoplay flag made explicit, rather than some hacked up soup of code that makes sure the video autoplays.

When intent is captured (e.g. specific "autoplay" flag) my browser client can interpret the flags according to my choosing. Capturing intent > not capturing intent.

You mean an attribute on the video tag? That is a thing. It looks like this:

    <video autoplay>...</video>

> What really should be done is getting rid of auto playing videos at the HTML spec level.

Or, let autoplay be a an explicit flag in the spec that is advisory as to publisher's intent, and establish a norm of user agents providing at least "big switch" configurability (Autoplay video on/off) and, ideally, optional per-site override of default on/off configuration for autoplay.

It's not a problem but a feature that increases Facebook's revenue by not insignificant multiples.

That is an important question that wasn't clear in the article. If they counted those sub 3 sec views in the count but not in the duration then it would definitely be fraudulent. If they didn't include sub 3 sec in the count or in the average then it is a valid technical decision. I am guessing they were included in the count but not in the duration.

They should have a histogram of durations.

That distinction makes all the difference to me. IMO one answer is fraud and the other is borderline valid / no big deal.

It looks like they didn't count it in either metric. Peoples' complaint is that they excluded it from all averages and all reports, which inflates their average duration of video watched because videos less than 3 seconds aren't included at all.

> If they counted those sub 3 sec views in the count but not in the duration

It sounds like this was exactly what was happening.

From the article: "Video views of under three seconds were not factored in, thereby inflating the average. Video views of under three seconds were not factored in, thereby inflating the average. Facebook’s new metric, “Average Watch Time”, will reflect video views of any duration."

One wonders whether this was an intentional filter to avoid scans of the video that shouldn't count as watching, or just a bug – some deeply nested, vestigial filter that everybody forgot about until some random engineer stumbled across it...

On my programmer hand, if I made the cutoff 3 seconds for the metrics, I would make it 3 seconds to count a view.

If the metrics and the view counts match, I don't see a problem.

If they calculate the average for > 3 seconds, but count EVERY view, I agree--fraudish.

This sort of Windsorizing will inflate the average anyway.

Making decisions based on "average" view time is still silly.

Why would it be silly? If I have an ad that has double the average view time than other wouldn't it make sense to invest more in it and less in the second one?

The distribution is what they really need to include. The average view time doesn't mean much when the distribution is super skewed, which it's likely to be for these ads.

If 0.1% watch all 30 seconds, 0.9% watch to 28 seconds and quit at the credits, and the rest all watch 0.2 seconds as they scroll through, that tells you something that "Average watch time = 0.5 seconds" doesn't tell you. Especially if last week, no one watched the whole thing, but a bunch of people watched for 2 seconds and then quit, and your average was 0.9 seconds.

Reply on news.ycombinator.com