VPM's and match length

[NoobPie]

So it turns out OzTAM have has a measure of internet streaming based on data provided by "participating" outlets

https://reports.oztam.com.au/#/live

The guidelines which set out what is gathered are provided here

https://oztam.com.au/documents/Other/OzTAM_Guidelines_to_VPM_Reports_1.pdf

OzTAM’s Video Player Measurement (VPM) Report provides Australia’s first official figures for viewing of internet-delivered TV content.

Launched in February 2016, the VPM Report is separate and complementary to OzTAM TV ratings.

OzTAM’s VPM service captures minute-by-minute data on participating broadcasters’ TV content played to connected devices such as tablets, smartphones, smart TVs, games consoles and PCs/laptops.

Here is a cut and paste of relevant information set out in the guidelines in regards to the report

What is a ‘VPM rating’?

A VPM (Video Player Measurement) rating is the total minutes of a program played across all connected devices divided by the content length and rounded to the nearest thousand. You can also think of this as the average number of connected devices playing the content across the program’s duration, which is similar to the way television program average audiences are calculated, albeit for devices rather than people in the VPM Report.

What is meant by ‘stream starts’?

Stream starts are the total number of requests to play the video content lasting at least 15 seconds.

How do VPM ratings and stream starts compare?

Stream starts represent the total number of times the program video content began to play for a minimum of 15 seconds; this count provides a device reach–like number. VPM ratings are the average number of connected devices playing the content in any one minute across the program’s duration.

Can a program’s catch up VPM rating be added to its TV average audience to get an overall figure?

The calculation of a VPM rating is in line with how an average audience for television is calculated, though the two are derived through different measurement methodologies:
o VPM measures the connected devices playing internet-delivered TV content, and those devices may be located anywhere in Australia.
o OzTAM TV ratings are estimates of people watching broadcast television, broken down into a variety of demographics, in both the five mainland capital cities, and nationally in subscription television homes.
o In VPM ratings, OzTAM cannot yet ascribe a user demographic back to a device, or determine whether more than one person is watching.

OzTAM does recognise however that some data users wish to see how VPM data relates back to the total audience for a piece of content across all platforms. Given the different methodologies and coverage areas for TAM and VPM, users wishing to combine the data should follow these protocols:
o VPM data may be added to the audience for an individual program episode (or series average) using either:
o an OzTAM Total People average audience figure, OR o a combined OzTAM and Regional TAM Total People average audience figure.
o No demographics other than Total People should be used.

So streaming data exists at a national level and you can multiply it to the OzTam "total people average audience figure" to get a number you can add to the national ratings for FTA or STV

BUT

What is clear to me is that you cannot add up average audience figures a make apples with apples comparisons of shows of different length. This is true of VPM and so must also be true of the normal ratings for FTA and STV

This is clear from looking at one of these reports



Here we can clearly see that two AFL preliminary finals on foxtel had more minutes streamed than either of the fox league games (obviously this excludes the channel 9 figures....not clear whether channel 7 is "not a participant" or its AFL streaming numbers are lower) but they end up on lower figures

The Richmond Collingwood game had almost 40% more minutes streamed than the Rooster Rabbitohs fox minutes but ends up with a VPM 2K below.

So the average viewer figures might be relevant if you are buying general advertising space and wondering what the expected audience will be when your add may be on but you certainly not add up averages of AFL and NRL matches Row Masters style and declare that the NRL ratings exceed the AFL ones

 

[NoobPie]

Apologies to anyone else that erroneous possessive apostrophe has annoyed....I can't edit the thread title unfortunately

 

[Rob]

You're right, but what you're saying isn't revolutionary.

Years ago ASTRA used to publish ratings figures for pay TV and include audience reach numbers - a similar figure to the 'stream starts' in the above report. It was common for NRL games to have higher average ratings, but AFL games would generally have far higher overall audiences.

For example (back from 2015):

http://www.astra.org.au/pdf/news/STV_Ratings_Report_Week_3215.pdf

The top 2 highest averages were NRL games, but both of those were dwarfed in audience reach by the AFL games coming in at 3, 6 and 8. The AFL game at number 10 had pretty much the same audience reach as the NRL game in number 1.

If you make the assumption that AFL games go for 180 minutes and NRL games go for 120, you can work out the number of viewer minutes for each show as well.
For example: NRL Sharks v Cowboys at number 1: 120 * 249k = 29,880,000 viewer minutes. Divide by the reach figure gives you the average viewing time - 54 minutes.
AFL Adelaide v Richmond at number 3: 180 * 234k = 42,120,000 viewer minutes. AVT - 66 minutes.
And just for the lolz - AFL Brisbane v Gold Coast at number 10: 29,520,000 viewer minutes. AVT - 54 minutes.

So basically the number 1 rating program for the week - an NRL game - got pretty much exactly the same number of viewers watching for pretty much exactly the same amount of time as the number 10 rating AFL game. Yet the ratings would have you believe 1 comprehensively outrated the other.

Which I guess proves your point.

 

[Gigantor]

You're right, but what you're saying isn't revolutionary.

Years ago ASTRA used to publish ratings figures for pay TV and include audience reach numbers - a similar figure to the 'stream starts' in the above report. It was common for NRL games to have higher average ratings, but AFL games would generally have far higher overall audiences.

For example (back from 2015):

http://www.astra.org.au/pdf/news/STV_Ratings_Report_Week_3215.pdf

The top 2 highest averages were NRL games, but both of those were dwarfed in audience reach by the AFL games coming in at 3, 6 and 8. The AFL game at number 10 had pretty much the same audience reach as the NRL game in number 1.

If you make the assumption that AFL games go for 180 minutes and NRL games go for 120, you can work out the number of viewer minutes for each show as well.
For example: NRL Sharks v Cowboys at number 1: 120 * 249k = 29,880,000 viewer minutes. Divide by the reach figure gives you the average viewing time - 54 minutes.
AFL Adelaide v Richmond at number 3: 180 * 234k = 42,120,000 viewer minutes. AVT - 66 minutes.
And just for the lolz - AFL Brisbane v Gold Coast at number 10: 29,520,000 viewer minutes. AVT - 54 minutes.

So basically the number 1 rating program for the week - an NRL game - got pretty much exactly the same number of viewers watching for pretty much exactly the same amount of time as the number 10 rating AFL game. Yet the ratings would have you believe 1 comprehensively outrated the other.

Which I guess proves your point.

Good post.

The other thing about the streaming numbers, and this comes as no surprise either, is that the reach equivalent figure is generally more multiples of the average rating than what occurs in traditional TV, especially for the AFL games shown above, I guess because the online attention span is less. While that remains a significant factor, traditional TV continues to have something attractive to offer advertisers (but like everyone else, I don't expect that even that will save it).

I have to admit, even that high reach figure of 107k above, which is some 8x the multiple of the average rating, is still a pretty low figure when compared with traditional TV. You keep hearing that online streaming has achieved parity with traditional TV viewing, but this seems to suggest otherwise.

 

[NoobPie]

You're right, but what you're saying isn't revolutionary.

Years ago ASTRA used to publish ratings figures for pay TV and include audience reach numbers - a similar figure to the 'stream starts' in the above report. It was common for NRL games to have higher average ratings, but AFL games would generally have far higher overall audiences.

For example (back from 2015):

http://www.astra.org.au/pdf/news/STV_Ratings_Report_Week_3215.pdf

The top 2 highest averages were NRL games, but both of those were dwarfed in audience reach by the AFL games coming in at 3, 6 and 8. The AFL game at number 10 had pretty much the same audience reach as the NRL game in number 1.

If you make the assumption that AFL games go for 180 minutes and NRL games go for 120, you can work out the number of viewer minutes for each show as well.
For example: NRL Sharks v Cowboys at number 1: 120 * 249k = 29,880,000 viewer minutes. Divide by the reach figure gives you the average viewing time - 54 minutes.
AFL Adelaide v Richmond at number 3: 180 * 234k = 42,120,000 viewer minutes. AVT - 66 minutes.
And just for the lolz - AFL Brisbane v Gold Coast at number 10: 29,520,000 viewer minutes. AVT - 54 minutes.

So basically the number 1 rating program for the week - an NRL game - got pretty much exactly the same number of viewers watching for pretty much exactly the same amount of time as the number 10 rating AFL game. Yet the ratings would have you believe 1 comprehensively outrated the other.

Which I guess proves your point.

Yeah my key point is that aggregate comparisons should build up from a common time unit

Average ratings are useful for advertising transactions which is the key commercial purpose of ratings but they are not an appropriate basis of comparison for between the popularity of programs / sports that run over different lengths of time

 

[NoobPie]

Good post.

The other thing about the streaming numbers, and this comes as no surprise either, is that the reach equivalent figure is generally more multiples of the average rating than what occurs in traditional TV, especially for the AFL games shown above, I guess because the online attention span is less. While that remains a significant factor, traditional TV continues to have something attractive to offer advertisers (but like everyone else, I don't expect that even that will save it).

I have to admit, even that high reach figure of 107k above, which is some 8x the multiple of the average rating, is still a pretty low figure when compared with traditional TV. You keep hearing that online streaming has achieved parity with traditional TV viewing, but this seems to suggest otherwise.

I think this is all correct but need to factor

1. the VPM is a measure of number of feeds not number of viewers. I don't know what the normal average viewers per unit is but presumably it is some multiple of this figure.

2. it does not include telstra. I'm not sure what this would add but it would presumably add something.

But you are correct, you'd highly doubt that the streaming options could increase the number of viewers substantially. I'm pretty sure IIRC the Richmond Collingwood final averaged over 2 million on FTA and fox....the most generous estimate of the number that may have been watching a stream on average surely couldn't get near 10% of that figure (and almost certainly less)

Bare in mind though that, as at least a "competitive tension" option in the next negotiations, the current streaming figures themselves are less important than the capacity to do so by 2023. Obviously the bulk of people watching the whole game will continue to do so on FTA or STV HD while that option is available