Australian arts jobs, news, industry commentary, career advice, reviews & data

What's On

Parrot does data science to track the future of television shows

David Tiley

Ratings tell us about eyeballs, but can we deep dive into the sea of conversation about shows and genres? Get some feeling for what people really want?
Parrot does data science to track the future of television shows

The fight for entertainment supremacy is brutal. Image: Game Of Thrones. 

Parrot Analytics is an international data analysis company which provides audience analysis services for exhibitors and program makers.

It is promoting public papers utilising its data, which are pretty interesting. They come from the quarterly Global Television Demand Reports 'based on Parrot Analytics' global TV analytics dataset, which is comprised of 3.5 trillion data points across 60+ languages in 100+ countries' according to the company. The papers are a PR tool for a much more detailed paid service. 


The latest, covering the first quarter of 2018, looks at TV industry figures compared to the OTT and subscription service in Australia and nine other territories. It looks like a snapshot of the battlefield as it rages. 

How did they get to share these numbers? Ratings and audience breakdowns are pretty restricted except on a general level. 

The answer is, they use a different key metric which they have devised themselves, called Demand Expressions® and is based on a Demand Measurement Methodology. If we only use actual usage statistics, then we can't explore approaching changes. So Parrot wants to find and measure the waves of interest and intention before they break onto the walls of traditional platforms. 

According to an email from the company, 

In the most basic sense, demand for television content is what drives transactions and consumption on all platforms, such as linear TV, TVOD or SVOD services.  

Consumers express their demand for television shows in a variety of ways, including video streaming, social media, photo sharing, blogging and micro-blogging, fan and critic rating platforms, peer-to-peer protocols and file sharing sites. 

These activities are what we track and measure, on a weighted average basis. 

We capture the empirical expressions of demand from all of these sources and combine them into one distinct, global standardized measure of demand that we call Demand Expressions. 

This allows us to compare the popularity of TV shows in all countries, for any platform or device. 

Here is one use of the idea of intention. It turns out that around 45% of the population does not want to sign up to any streaming subscription service at all in the United Kingdom and the United States. In Brazil that figure is slightly lower but in Italy it is down to 18% while over 70% will pay for a subscription service, which would make the accountants in terrestrial broadcasters very unhappy.

At the other end of the scale, about 7% of Italians will pay for four or more, most over 45 with a big group over 65. In the UK that figure is 10% but the 25-44 category much larger and probably the same as the 55+ group. In the US, the 4+ group is down to 3% with the great majority in 25-34. Men outnumber women in willingness to pay. 

This is a measure of the market available for paid streaming services, from which individual companies can calculate the potential scale of their businesses and how much they need to take off their streaming competitors. They can cross-check the figures with their own sales staff.

In general, companies will hire Parrot as a weapon in the marketplace.  Sellers can provide comparative data on a market and estimate how much attention a show will get. Buyers can do the same for a production coming into their marketplace. Turning attention to ratings is the responsibility of broadcasters or streaming companies and varies according to resources, membership and competition. 

The document also contains data about Demand Expressions for individual streamed series in nine different countries. Here are the top eleven for Australia on an average day in the first three months of 2018.

The numbers don't mean that much by themselves but they do work in comparison. The company can break these numbers down much more to provide detailed day by day graphs and a pile of demographic information. Clients can track their shows against competitors and measure the effect of advertising and public relations. 

They do show us that Stranger Things has nearly twice as much more attention than anything else except the venerable Star Trek brand - and almost all the shows in the chart have been running for a while.  In our world we mostly talk about Crown and The Handmaid's Tale, which are way down the list. Netflix dominates this chart by a humungous amount, there is only one reality show, and Science Fiction Rools. There is very little horror, and the highest Australian show is STAN's Wolf Creek at 38 with 0.204m. 

In the US Star Trek handily beat Stranger Things beat Black Mirror; the UK was similar but not as extreme; ditto Switzerland; in the Netherlands Stranger Things just beat Star Trek just beat Black Mirror; ditto Belgium. In several territories like Brazil, Star Trek was not on offer and Stranger Things beat Black Mirror. 

Black Mirror is near the top in every territory and The Grand Tour, a reality show from Amazon is in the top half dozen. But it is pretty well the only reality show and the only Amazon project in the top twenties.

This data is about streamed services. What happens when we look at the wider world of television? Parrot is not so interested in those comparisons because it is tracking the relative success of shows on the particular delivery system.

We of course are tracking the life and death of the entire system. Parrot has supplied figures for the previous six months in which it has compared all the television shows. Game of Thrones was on Foxtel cable. It ran just under 6m demand expressions each day, followed by The Walking Dead at 3.73m and then Stranger Things at around 1.8m.

A daily demand figure of 1.69m might sound a lot, but the previous six months, the second of 2017, had Game of Thrones among the options. Game of Thrones ran just under 6m each day, followed by The Walking Dead at 3.73m and then Stranger Things at around 1.8m. And GoT was on Foxtel.

So Stranger Things was relatively constant and probably tells us that viewers will happily talk about both shows - they are not trying to push each other out. These numbers also show us that one fantastically popular show can upturn the relative popularity of genres completely.

Nevertheless, the general figures tell us that sitcoms come out on top, followed by crime drama. Here is the graph of monthly Demand Expressions:

Parrot reckons this is caused by the 'continued popularity of sitcom superhits The Big Bang Theory, Brooklyn Nine-Nine and Modern Family.' 

Australian drama Wentworth pushes Crime Drama into second place. 

These are not audience figures; FTA TV wins those because it has access to the whole market so moderately enticing shows can clobber great shows on subscription. However, the numbers are about cultural attention and a production like Game of Thrones becomes a meme which spreads much sider than viewing audience, which was only 887,000 on the first screening of the season seven finale. Different genres will also generate different kinds and scales of conversations. Sitcoms may be easy to talk about. 

But the fight for attention hardly features any Australian shows. We watch them, we like them enough, and they are there. But the blanket cliche that Australians love local shows is maybe not so true, which helps to explain why the local FTA companies want to get rid of them. 

We love them but we don't love them to bits.


How does Picnic at Hanging Rock fit into this story? The ratings for a good Australian drama on Foxtel can reach 150,000 for the first screening of an episode before timeshifting. Multiply that by three to get an equivalent figure for free-to-air broadcaster, and we have a Network Ten average for drama.  Foxtel can add to it with repeats but they don't rate very highly either. 

Foxtel is running Picnic as an On Demand box set at the same time, so the figures will be better than the ordinary national ratings. The first ep on Sunday night ran around 112,000, but the second ep dropped to 45,000 on Monday.

David Knox on TV Tonight has a discussion about these numbers. I would add the word sad. 


With not a feather in sight, Parrot Analytics CEO Wared Seger gave a presentation at the recent Business of Media Summit, which is on Youtube.

About the author

David Tiley is the editor of Screen Hub.