Why Overnight ratings are not the full picture

2015-11-04_0054

OzTAM has amended rules for its new Video Player Measurement system following feedback from TV networks.

The new VPM number, which measures devices watching Catch-Up content online, was initially not to be added to Overnight and Timeshifted numbers, but OzTAM advises these can now be added.

“We’re eager to understand what our colleagues think about the new service, the ways they see themselves using it, and how they’d like it to evolve – in particular our moves towards a ‘hybrid’ measure that will include demographic data,” OzTAM said in a statement.

“It’s also clear that people want to see how VPM data relates back to the total audience for a piece of content across all platforms.

“OzTAM has discussed this with the MFA, AANA, commercial, public and subscription television broadcasters.”

VPM data may now be added to either:
an OzTAM Total People average audience figure OR
a combined OzTAM and Regional TAM Total People average audience figure.

The introduction of VPM reminds us that the way audiences are viewing content continues to evolve. Overnight data, whilst commonly reported as a measure of a show’s success (including on this site) are in effect only part of the picture.

Most Free to Air shows have a variety of playouts including:

1) Premiere episode*
2) Repeats*
3) Timeshifted* (the number of viewers who watched it up to 7 days later on their own PVR)
4) Video Play Measurement national figure (VPM: online catch-up views up to 7 days later)

* 5 City Metro markets

Another figure, those streaming Live, is yet to be counted.

For networks eager to put their best foot forward it’s not hard to see why they are keen to have full data reported. In the US, where 18-49 is the driver, the push has been towards Live+3 days and Live +7 days.

To demonstrate how a show improves beyond its Overnight figure here are three recent examples based on 5 City Metro*:

Here Come the Habibs (Feb 9):
Overnight: 1,258,000*
Repeats: 210,000 + 403,000*
Timeshifted: 223,000*
VPM: 54,000
TOTAL: 2,158,000

Home and Away (Feb 11):
Overnight: 1.177,000*
Repeats: 72,000*
Timeshifted: 75,000*
VPM: 48,000
TOTAL: 1,372,000

I’m A Celebrity Get Me Out of Here (Feb 8)
Overnight: 746,000*
Repeats: 113,000*
Timeshifted: 55,000*
VPM: 12,000
TOTAL: 926,000

Each of those samples climbs higher still if we add in Regional viewers.

Aside from VPMs, which represent all Catch-Up playouts, OzTAM metro data is drawn from 5 city metro markets where 3,500 homes are fairly representative in terms of their household demographics. Sadly we still can’t have a true national number because some regional markets are not covered by the RegionalTAM system.

There are other factors that come into play too. SBS has also argued its viewing audience for NITV is more populous in regional areas, and areas not covered by people meters. ABC3 also shuts down at 9pm, halfway through the primetime period a share is calculated. Until next week Nine’s network share also includes an extra channel over Seven.

2016-01-25_0054

Here are some other common questions that often arise:

Why does media report Overnight figures and not a final figure?
In part this is a reflection of our history, but also because the data takes up to 7 days later to be tallied. If last week’s Australia’s Got Talent episode climbs by another 120,000 viewers for example, that’s really hard to push as a story. Broadly speaking it’s the Dramas that shift the most, and these numbers are displayed regularly here.

I hear networks talk about Peaks and National figures. Are these correct?
Yes the figures are correct, but advertisers measure a show’s performance based on the Average in a 5 city metro, because all things are equal.

Why does an Overnight number that’s issued for a show sometimes change?
Adjustments happen all the time. The Overnight figure is preliminary, according to the time a show was due to play according to the EPG. An Adjusted figure will be the actual times, taking into account a late start.

Does that mean a network can get a better number for a new show if a popular one runs late?
You bet. It happens all the time. By the time an adjusted figure is confirmed media stories are already out there based on preliminary Overnights.

Why are some shows split into two halves such a the 6pm news, The Chase and The Project?
Networks argue it is to show advertisers competitive half hour figures. But some shows see dramatic increases in their second half hour, so by having two figures, one will be much higher.

What about simulcasts?
It stands to reason that a show roadblocked over 3 channels will enjoy a better figure than on 1. There is some argument it may decrease a network share because it offering less diverse programming. Networks are doing these less.

Haven’t viewers switched off in recent years? Shows used to get 2 million viewers all the time.
Yes and no. Free to Air viewing is down due to other entertainment alternatives, but it’s still very high. Fragmentation is the word. We can’t compare shows today to ten years ago when we had 5 channels and Foxtel. Maintain some perspective.

Are channels like Racing.com, SpreeTV, Extra and TVSN included in ratings?
No.

Does Foxtel differ from the Free to Air ratings?
Yes very much so. While FTA focusses on 5 city Metro, Subscription TV ratings are national. One should also remember STV is in around 30% of homes, and there are multiple playouts of titles. The number generally published refers to the premiere playout. Also not all channels are available in the basic packages.

Why don’t you publish more than the top 20 a day? With all these half-hours creatively coded, it’s not even 20 shows anymore.
Tell me about it. OzTAM limits publication of rankings to the Top 20. I try to include more in the daily wraps under News.

Should media reportage move from Overnights and 5 city metro to Live+7 days national?
In an ideal world yes, but the practicalities are far more challenging. We all love to know -sometime after 9am- how our favourite show performed last night. Would we all be willing to wait a week for it instead? Unlikely. While data is provided for Timeshifted & VPMs the perception of a show’s performance rides heavily on the Overnights. The argument for national rather than 5 City Metro is far more sound, and one OzTAM has said it is moving towards. It’s taken 2 years for the VPMs to get up and running.

It’s fair to say that when a network is looking at a show’s performance when deciding whether to renew or not, they are looking well beyond the Overnights and media stories.

8 Comments:

  1. Thankyou for this article, David. Explained everything very clearly, and even the things I knew about before, it was good to revise and clarify my understanding.

  2. Overnights are mostly the live figure, (but also include timeshifting (pausing and recording and watching) before 2am). Overnights measure what people choose to watch while sitting down in front of the TV that night. The live figure for 25-54 is where most of the ad revenue comes so that is also what is most important to programmers when deciding if a show will be remain in an early primetime slot, or get bumped or dumped.

    Since the percentage of timeshifting and catch-ups will similar for similar sort of shows, the total extra viewers will be roughly proportional to the overnights. So overnights will serve as a proxy for relative comparisons, until the final numbers, which count exact repeat, timeshifting and catch-up numbers are available (but by then people have lost interest in that episode anyway).

  3. I take issue with OzTAM limiting media outlets reporting only up to the top 20 shows per day on FTA and STV respectively. I get they are trying to increase the amount of companies paying them for the more detailed numbers, but surely it is in the media outlet’s interests to publish at least the total viewer numbers for shows beyond the top 20 on a regular basis.

    I also think there is something to say about moving from focusing on total people to the ad demo figure like the US do. Whether 18-49 or 25-54 is more relevent in Australia could be debated given our aging population, but either way we know advertisers don’t care about 55+ viewers. In the US, shows live or die solely on under 55 viewership.

    • Yes this is silly but it is what 7,9 &10 want. It means that only the top ratings shows get talked about. By splitting shows they can get a higher place in the list, extra spots in the list and squeeze others off the bottom of the list show that they don’t get reported. They may also be selling ads at different rates in different parts of the show so need the numbers to support that.

      The US mostly talks about 18-49 overnights, which are out the next morning, with adjusted figures out a day or two later. FTA advertisers pay for Live + 3 numbers for 18-49s or 25-54s (they will pay less per viewer to account for the ads skipped though). Cable uses overnights for premieres and +7s (since there are repeat airings for a week usually).

    • Exactly, I have asked before why those over 55+ don’t count in ratings numbers. There are so many more shows they should increase it to the top 25. I often wonder how the off channels shows rate as well but we never seem to hear anything about them….JAG,Dr Quinn etc.

  4. “Why are some shows split into two halves such a the 6pm news, The Chase and The Project?
    Networks argue it is to show advertisers competitive half hour figures. But some shows see dramatic increases in their second half hour, so by having two figures, one will be much higher.”

    And this should be banned – if you’re showing the same program it should be reported as one, not split into 2, 3, 4 or 5 parts. It’s absolutely ridiculous.

Leave a Reply

You must be logged in to post a comment.