Streaming Analytics: Key Points to Track

Learn more about streaming analytics at Streaming Media East.

Read the complete transcript of this clip:

Steve Geiger: Key points to track in making your decisions about video preparation. Screen resolutions and those being presented based on the playback experiences are a core piece of this, and as I talk about this, the thing that I will really want everyone to focus on is that it's best to look at your data. Industry benchmarks will give you a good idea of directionally where to go but when you collect that via your own analytics tooling, when you pull in your own data you'll start to see the deviations that you have from those industry standards. And that can mean a big difference in your performance, in your savings ultimately when you start to make decisions based off of that.

So the core thing here is don't encode resolutions that your users can't benefit from. So here we don't see really any 4K utilization so let's not produce any 4K renditions or serve those up to these customers. Scale factor is another thing that a lot of people don't consider but is a relatively important one. Really what we're looking at here is what resolution was delivered versus the detected resolution of the playback. And again, you're optimally going to match this and have no scale factor. If you're scaling video down that means that you're potentially wasting bits, you have a higher resolution than you needed to deliver. Scaling video up means that your users aren't getting the ideal experience. Now, there are caveats to that, and sometimes you do want to be in a situation where you're scaling down video. But as a general rule, you want to really try and aim for this whenever it is possible.

Lastly we're going to take a look at bit rate in two different ways. One is, don't generate rates that your users can't watch. You obviously have a pretty even trend here around that 900 kilobit mark here, so producing a three or four megabit rendition isn't going to do your users very much good; they'll never adapt up to it. You should also look at what we call a heat map or a density graph of what people are watching over time and look at the distribution of your users across those renditions.

So here in a different situation we have about a quarter of our viewers hitting the highest rendition of 4.8 megabits, and we have folks that are going to be down into the 800 to 400 kilobits range, but there's this kind of odd stripe in the middle of that in this case we can track back down to zero rated or cap data plans. So there was a gap there in which we weren't going to have any viewership of that 1.21 megabit so maybe that's not something that you needed to produce.

So, ultimately I want to come back to one of the things that I mentioned earlier which is video codec support, another thing that you can get from your analytics. And here again the pervasiveness of AVC so great everybody can receive that. We have 19% at HEVC and 74% at VP9 as well as a 55% for AV1. So the interesting fact here is if you remember what the plans were and what people were implementing again in a general rule is that they were doing way more HEVC than they were VP9. But when you start to peel back the data and look into it, the actual addressable market for those codecs might be a little different.

Free

for qualified subscribers

Subscribe Now Current Issue Past Issues

Winning the Data War: Accessing and Leveraging Streaming Analytics

This article explores the topic of data rights versus technology capabilities as it impacts media companies' ability to leverage accrued data and analytics to target relevant audiences, expand their reach, and monetize their content offerings. If the intersection of data rights ownership and technology capabilities access were captured in a Venn diagram, the space in between would be very small for some media companies.

09 Feb 2023