0/5

So how do ratings determine simulcast figures anyway?

Here's how important the audio track is to ratings data measurement.

While TEN’s Commonwealth Games ratings are being broken out into separate data for TEN and ONE -and networks can choose wihether to simulcast or aggregate figures- it raises the question: just how is the data determined anyway?

OzTAM contracts Nielsen Television Audience Measurement to collect the metropolitan and national subscription television audience measurement (TAM) data.

As it turns out it is the Audio track that is a key factor.

In terms of coding, TEN calls the programmes the same thing, sending two separate logs to the software company – one for TEN and one for ONE. These are absorbed into their system and the ratings are then matched back to these logs.

The software matches the audio track of a show that a panel member watched with the audio track that the various networks broadcast.

When the software finds a match, it knows the viewer was watching a particular show.

But where a programme is being simulcast, the software can determine whether the programme was on a primary or digital channel by the different Network promos that aired.

Everything, including the commercials can be identical, but if the software detects a promo for Offspring it knows the viewer was on TEN and when it detects a promo for One Week At A Time it knows the viewers was on ONE.

14 Responses

  1. @Baggygreen – it’s the sample size of the population that is being measured that is important not how it compares with the number of choices. You can be assured that those people that rely on this data to tell them how to spend their money will employ or have access to people who’s job is analysing stats.

    @CK – that’s pretty much it. I know it seems like black magic that such a small sample is extrapolated to the whole potential audience but the sorcery is grounded in solid, well-tested maths.

  2. interesting article. What I don’t quite get is how they come up with the ratings figure e.g. 1.3 million watched the news, if there are only 3000 (?) boxes. Does each box represent a certain number of people e.g. a box = 100,000 people so if 10 boxes were on this would mean 1 million people??

    Thanks in advance to whoever can explain this to me!

  3. Interesting article David. For me a bigger question is how many boxes exist. In the last 2 years (ish) we have increased from 5 basic channels to 15. Has there been a 3 fold increase in boxes? If not is the sample still statistically valid?

  4. We have a ratings box – and a PVR. Anything we watch up to a week after it was broadcast counts in the new, “delayed” ratings the networks get. We skip ads all the time and I guess they know that, but there’s nothing much they can do about it!

  5. Oct 2009: OzTAM will measure the actual viewing of recorded television content, not what is being recorded or what has been recorded via a ‘state of the art’ metering solution called UNITAM. Only recorded material played back at standard speed will be recognised.

  6. IRT, you originally post that cutting out commercials, which results in you not counting, is a “flaw in counting viewers who watch using PVRs.” Uh, it’s not a flaw, that’s how the system is supposed to work.

    The viewing numbers only matter if people are watching the commercials. You know, the entire point of commercial television? Without ads, commercial television wouldn’t exist? Yeah.

  7. @Learning to Read is Fun! PVR’s let you watch the whole program at a slightly increased speed including understandable audio – so the audio won’t match, so you won’t be counted … sigh.. no where in the post did it say watch the commercials at fast speed – learn to read.
    If you cut out the commercials that the software relies on (as stated in the post) to distinguish between channels – you won’t count … read the article.

  8. @Rob – talk about Not being able to read the article and absorb what it says correctly – this is about simulcasting figures – since when do 7, 7TWO and 7Mate all have the Same show on at the Same time.

    Learn to read, then learn to understand.

    @IRT – why, if you cut out commercials or watch the commercials at fast speed do you Not count? Surely the Audio of the program is enough to determine what you are watching.. sigh… some people… the only thing this would do is skew the figure (most likely) to the primary channel if its a simulcast.

  9. Thanks David, wasn’t aware that was how the matching algorithm worked.

    @Rob – cross-promos won’t be a problem on 7 (or 9) network channels unless they are simulcast; ie have *exactly* the same audio track.

    @IRT – I don’t know if what you’re saying is actually the case, but I’d argue that if you’re not watching the ads, then perhaps you shouldn’t be counted!

  10. In this day and age, I am very surprised there is not a more sophisticated way of telling what channel is being watched such as an inaudible digital tag reporting the channel number every few seconds that the box picks up. Would work for both analogue and digital.

  11. This also highlights a flaw in counting viwers who watch using PVRs.
    The new “overnight ” and “consolidated” figures are supposed to include viewers who use PVRs either that night or the next week respectively –
    1. If you watch a recording of a show in anything but normal speed (DVRs let you play back at for example 1.3x speed) you don’t count.
    2. If you cut out the commercials you don’t count either

  12. May not work well for Seven since 7, 7TWO promos are on 7mate often enough.

    Might also explain to some viewers why cross promotion doesn’t seem to occur

  13. Thanks for this article David. 🙂 All very interesting. The difference in promo’s would be obvious between One and another station, but it seems flawed to me in other cases, where there may not be an obvious difference, or none at all. e.g. ABC2 and ABC News 24 when News Breakfast is on both channels. ABC News 24 has the crawler at the bottom, but there is no audio associated with that difference, it’s purely a visual difference.

Leave a Reply