How robust is our Ratings system?

2013-11-21_2233Television is a numbers game and at its heart is the OzTAM ratings system.

This week at the Screen Forever conference there was much discussion, in several moderated sessions, about the accuracy of the OzTAM panel and the desire to count views on tablets, phones and catch-up services.

TEN CEO Hamish McLennan cited the example of The Bachelor.

“We have as many people on any given week watching it on TenPlay as they are on terrestrial broadcasting, which is good and bad for us because we can’t monetise the show on digital the way we can on terrestrial,” he said.

“The market was very quick to say The Bachelor has only been a moderate success. But if you take the 800,000 people who watched it last week on linear TV and the 820,000 I think it is on TenPlay, together with catch-up service in the week, you’re looking at 1.8m odd. Which is extrordinary.

“So it’s an absolute smash but we’ve got to redefine how we judge these shows.”

Asked whether TEN was disadvantaged by having a younger audience who may not have a landline phone (the traditional method for sending data between households and ratings service provider Nielsen TAM), he was careful with his words.

“Whenever I talk about that it looks like I’m being defensive because our ratings aren’t where (they need to be),” he said. “But I would just pull it back to looking at TenPlay it’s not reflected in OzTAM. The very fact that so many people are consuming premium content, and doing it in different ways, we need a more robust measurement system so that we can measure and monetise it.”

McLennan’s final sentence is bold, given that TEN is a one-third shareholder in OzTAM along with Seven and Nine.

It should also be pointed out OzTAM does indeed sample ‘mobile households’ too.

But not everybody is in agreement about video plays, either in definitions or measurement methodology, which is part of the problem about why they aren’t yet in the OzTAM panel survey.

Doug Peiffer, CEO of OzTAM later noted, “Those numbers are usually Stream Starts. That’s when you go into the player and you start viewing something, that counts as a viewer. It’s a Start. If you come to the next Pre-Roll, that’s another Start. You Start the content, Stop the content. That accumulates, it’s not a new viewer or a unique viewer. That number needs to be discounted and that’s what we’re working on.”

John Grono, CEO of Gap Research representing advertisers, agreed with Peiffer.

“Think of it like this: a stream is an opportunity to serve an ad, so me as a media buyer is very interested. But as a Broadcaster, a Stream is not equivalent to the OzTAM ratings. It’s nowhere near it,” he said.

“Every time you serve an ad there’s another stream break, so when Doug talks about 2 million people he actually means the average minute of that programme was 2 million. Sometimes 1.8m, sometimes 2.2m -but the average is 2m. Streams is nothing like it.

“It’s an extremely misleading statistic in my humble opinion.”

Streams can include up to 5 Starts in a one hour programme, distorting the number of actual viewers who may be watching.

Lisa Walsh, Head of Audience Research, ABC TV, spoke favourably of the current system but was looking for improvements.

“The TV ratings are still by far and away the best measure of television content and far and away capture the vast majority. Most particularly for adult content. We think there is probably a larger, discreet children’s audience viewing increasing amounts of content through tablet devices,” she said.

‘We’re blessed in this country, (OzTAM is) very robust. But we’re eager and working with Doug and other industry partners to include that bigger universe that’s out there and is growing, and how to include that.

“We’d like you to include Program Plays as People,” she told Peiffer.

While television sets are large, static appliances, the mobility of tablets and phones creates new challenges. It’s impractical to attach an iPad or iPhone to a People Meter. And with trying to monitor data in terms of Plays, there is still the problem of working out the demographics of the viewer, especially when devices are shared in households.

“We’re heading from a hardware solution to a software solution,” Peiffer explained. “Lisa talked about the data she gets back at the station. Hamish from Channel TEN talked about it. Everyone has data. So what we’re working on in the industry is to pull all the data together and harmonise it and put out official reports. That’s what we’re working on at the moment. But the tricky part is who do we assign it to?”

John Grono agreed: “We know you might have watched a video in one particular app and then you watch it again in another app, you count as two completely different people. So all of a sudden you see the audiences being inflated again and again and again.

“Who watched on television, who watched on a smart phone, who watched on an iPad? Who watched at work? It’s about that de-duplication of those numbers. It’s a huge black box of mathematical mumbo jumbo -which I actually love- but that’s what we’re trying to build. Trying to get that quantum right. If we don’t get the quantum right we’re stuffed.

“The first thing we have to do is make sure the numbers are comparable. The word ‘harmonise’ was used. So we have a lot of work to do. The problem is this horse has already left the stable and it’s bolting down at Caulfield. How do you stop it?”

There was also discussion over the drop of numbers in viewers from 9:30pm. Do we all retire early these days? Not necessarily.

“The number of people sitting on the couch together watching TV has gone down to 1, except for the top rating shows like X Factor and Talent and they’ve got more people. They’re rating big because they’ve got more people in front of the TV set. The more niche a programme goes, the less people are sitting in front of the TV,” Doug Peiffer explained.

He also elaborated on the way numbers are tallied, reminding us that Overnight ratings include Live and As Live numbers -the number of viewers who watch content on the same night it broadcasts. If a viewer watches one 8pm show then records another, but watches it at 9:30, the data shows both figures as an 8pm view.

“The Live ratings start to drop at 8:30 or 9:30 because you’re shifting the content. So that’s why the 9:30 programmes are falling back a little bit. Because people are shifting the data, shifting the evening and shuffling things,” he said.

“If you have 2 big competing programmes on then you’ll get that shifting. That takes the Live down later in the evening.”

Lastly, Peiffer was asked why OzTAM does not include Regional audiences with 5 City Metro numbers.

Regional viewers are measured separately by RegionalTAM (owned by Regional broadcasters) and there is some cross-over in some geographic areas.

“Technically there are some math reasons why you can’t do it. But I think we’re heading towards a national panel eventually when the Reach Rules change and consolidation happens in the marketplace.”

27 Comments:

  1. With the roll out of IPTV and HBBTV over the next 5 years won’t terrestrial be less significant anyway? HBBTV will give much more accurate ratings as well. Surely that is the way TV networks want to go?

  2. I am a 42 single mum who has not watched a tv show live since Foxtel bought out their IQ service. I don’t watch ads as I will purposely wait until the show is twenty minutes in so I can avoid them. Now I know you are thinking that I am someone you don’t care about but the thing is that every one I know does this. The people who have their ratings boxes do not give an accurate picture of how people watch the shows. In fact streaming shows at a later time via the catch up services are best for the advertisers because on those you are forced to watch the ads. I think the question shouldn’t be how many watch the show but how many actually watch the ads and how can we come up with different ways to imbed advertising that can’t be avoided.

  3. cynical old codger

    Of course we all want to have good shows to watch, and the accuracy of ratings systems can play a major role.
    So I ask how can any network question the ratings results, by drawing attention to their own data gathering such as tenplay, which leads me to wonder if tenplay was introduced primarily with the viewer in mind, or in fact was it another way of counting viewers.
    But if accuracy is the crux of the issues, perhaps some networks get exactly what they deserve, after all these networks who want more accurate numbers that reflect accurate and actual data etc., could start by cleaning up their own acts, and get their EPG’s and Untrue Promo’s to actually and accurately reflect something they the networks have complete control over, that being advertised starting and finishing times especially with pre-recorded programmes.
    EG. Where when there are eliminations each week, but with…

  4. Surely you can just measure a ‘complete’ stream like this: e.g. 5×2/3 ads in each stream. Then count views only for those streams in which 5 ad breaks played, therefore making it a complete stream.

  5. OzTAM/Nielsen have GSM peoplemeters for use where there’s no landline.

    The networks each run their own online / catchup systems, so it’s up to them to either do their own statistics or pass the numbers on to Nielsen. But since the demographics are practically unknowable (e.g. respondents to online surveys will self-select, if the survey is forced then an unknown number will lie just to get straight to the content, etc) and the number of people watching unknown, there’s no way of incorporating that info accurately into the ratings data.

    If you want semi-accurate ratings, then it’s probably best that they are reported separately. Though ideally they should be audited by Nielsen as well, not the networks alone…

  6. Brett, I remember there being similar talk re:landlines during the ellection. As all the pre Election polls were done Cold calling landline numbers.
    I’m the same as you, no one i know besides my mum still uses a landline. We have one connected for the internet, but no phone hooked up to it.

    I’d love a ratings box. Because i dont, i get really vocal on twitter when i like a show. Doing my little bit to show the producers that there are passionate people watching, even if there aren’t a Million of us.

  7. I’ve always wondering and thougt that it would be flawed, I would love it know if they get a full mix of people, gay, straight, black, white, asian, mixed, retired, working full time, unemployed, single parent, etc etc the list would go on as everyone of those people would watch different things, it wouldn’t just be aged brackets.

    I also wonder about the use of a landline for the reporting, as not ones of my friends has a landline, which I feel would skew the results very favourable to an older family demo, which is probbably why channel 7 does so well.

  8. I’m sure David will correct me if I’m wrong.

    OzTam doesn’t do rating surveys, the Nielsen company does the hack work. OzTam was set up by the 3 commercial channels to provide data (not information) sourced from Nielsen that could be used to drive sales. Hence, there are so many ‘options’ for every channel to proclaim that they are the ‘winner’ for any particular scenario for a given timeslot.

    Unfortunately, the Nielsen system was designed back when Jesus and I were in grade school and is being made redundant by new age technology. Something has to give, but no-one wants to change.

  9. The online viewers are hurting not helping Ten. They are not a public broadcaster.

    Advertisers won’t pay for online ads unless they can know precisely who is watching their ads. That is why Google and Facebook have flexible attitudes to privacy and are always getting into trouble.

  10. David,

    If ten owns Tenplay, then he has statistics of the amount of clicks and unique viewers ….. and this may not be overnight, it could be over a week or so …..he should be more accurate with his statements …..

    • Randwick sorry but that’s absurd. A CEO giving a speech to a forum of producers does not need to drill down to the exact number of plays on a single show when they have 3 channels and various platforms. He was giving an overview of ratings vs online plays. While I don’t buy into the clicks = people theory, his knowledge of 820,000 plays demonstrated he was across it. I have clarified the context for you, and you are misconstruing it. Moving on.

  11. “and the 820,000 I think it is on TenPlay” …… Hamish is guessing the figures? …. “I think it is on tenplay”…?

    Is he just talking up the figures?

    • Not at all. The 820,000 was a ballpark figure and the “I think” was referring to digital exec in the audience for affirmation. It may not translate well here in text so apologies for any confusion. The fact he knew the figure was actually a sign of level of detail. Trust I cleared that up, thanks.

  12. The main reason Ten gets hammered with the current ratings system is that Oztam require you to have an active phone line to be part of the survey. Now, how many people under 40 (or 50 for that matter) do you know that only have mobiles and their ADSL connection? The people that do have phone lines, broadly speaking may be a little more conservative in their thinking or live with older influences?

    It is an outdated system and you cannot underestimate the affect on Ten’s ratings.

    I wonder if the buying agencies have a handle on this and have been getting some great value out of their slots on Ten…

    • Many points to this complex problem.

      As noted above OzTAM also includes ‘mobile households’ to address the diminishing landlines. Advertisers do not accept 800,000 plays = 800,000 people. I also believe everything is proportionate. If The Bachelor were tallied at 1.8m all up, then what sort of numbers would we seeing for Seven and Nine content? Nudging 3m? We would all have to move the goalposts about what we discern is a hit, a good number, a fail.

      That said, The Bachelor was absolutely a hit for TEN, by coming home with a wet sail.

  13. jezza the first original one

    The monitoring of online clicks/views is already a huge sophisticated industry and advertising rates are charged accordingly. You have to remember OzTam produces ratings so advertising can be priced for fta.

    Any well managed, easy to use online portal will have its own figures and produce its own revenue stream. There is nothing to be gained by merging the figures. Keep them segmented as it creates more opportunities to sell.
    If the Bachelor pulled 800,000 on tenplay, well why isn’t ch10 deriving as much income from this as their normal broadcast. It sounds like someone is half asleep and not doing their job.

  14. @Bass_A- i think this is the point McLennan is trying to make. You have the impression that Ten may be doing a lot worse than it actually is. If a disproportionate number of people watch their content online instead of terrestrial TV it isn’t reflected in OzTam ratings and gives a poor picture of the consumption of their content.

    You may think they’re doing terribly because you don’t watch and the ratings you see stink, but that might just be your view point in the same way as no one i know watches 7’s content and we watch in disbelief as they achieve triple the ratings of Ten. You have the impression from a series of bad media releases about ratings that Ten’s is performing woefully. No one can argue that it’s a success at the moment, but the picture may not be as bad as you believe

  15. OzTam getting defensive of their flawed rating system… who would have thought?

    I’m sure C7 will be the ones least resistant to any ratings system movement

    Time to start reporting catch-up views and stream starts – the audience is moving on, time OzTam catches up or works out a better way to report numbers

  16. ryan: “How can 3-5000 people be any guide to how 23m people watch tv.”

    Very easily. It’s called ‘inferential statistics’. You may not have heard of it; it’s only been around a few hundred years…

    (Sorry for the sarcasm; you may have missed the dozen or so times myself and others have tried to give a concise explanation of OzTam’s accuracy here…)

  17. I think a multi purpose ratings system could be introduced that could track tv, radio, print, cinema going etc so advertisers can get a true sense of where the eyeballs are at. Use phone apps or computer etc. results would be fairly instantaneous

  18. I think the system is outdated, it doesn’t take into account regional viewers or online options in the total numbers until later when networks release them to highlight a particular show.

  19. Please…. if anything this just shows 10 are in bigger trouble than we know. If it concerns him that much make the Finale a Live one – duh

    TenPlay isn’t 10’s issue, the brand is a mess, I barely even watch 10 at all, in fact there are nights I don’t even click through it, usually start at 7 Digital (71) and go through the stations that way, avoiding 10 and 11 as there is nothing on there worth watching

  20. jezza the first original one

    Talk about attempting to moddy the waters and create a very confused argument.

    You charge one price for adverts via the traditional fta viewing measured the traditional way and charge another price for hits/views via online/mbl streams measured in the way these things are measured. Youtube seem to manage it ok….durrr

    Still stuck with stone age thinking some of these execs

  21. I don’t like the system what so ever! How can 3-5000 people be any guide to how 23m people watch tv. We now live in 2013 and it amazes me we haven’t sorted this out through technology. I don’t agree that online views should be counted but I think there needs to be a more accurate way of doing it. It could be as simple as oztam asking the public who wants to participate, then oztam can develop an app that you can download to your tv set and the app either records all your viewing info of you enter it yourself.

Leave a Reply

You must be logged in to post a comment.