Last week Netflix published its most detailed set of viewing data to date.
Prompted partly by the deal terms emerging from the WGA strike negotiations (which include a clause on ‘Streaming data transparency’) and partly by expectations from advertisers, who are unlikely to agree with Reed Hastings’ 2019 remark that viewing data “doesn’t matter to anyone” (as I’ve said before in relation to Netflix: it’s not the strategy until it’s the strategy).
The report provides viewing hours for every title watched for more than 50,000 hours in the first 6 months of the year. This equates to 93 billion hours of viewing of over 18,000 titles - 99% of all viewing on the service.
Whilst viewing hours is a somewhat frustrating metric, skewing the rankings in favour of titles with more episodes / longer running times, it’s still nice to have a proper dataset to play with, versus Netflix’s opaque Top 10 and Most Popular lists which “take into account…premiere dates” (presumably in order to foreground more recently released titles).
Talking of release dates, the dataset has a column for Release Date although only 27% of titles have one, which feels like an odd omission for such a supposedly data-driven company.
One column which is fully populated is the ‘Available Globally?’ column, leading to my first takeout from this dataset:
1.) Netflix is a global platform but only a quarter of its catalogue is globally available
And it’s not just the long-tail that varies by market - 46% of the top 200 shows aren’t globally available.
It’s partly a reflection of the fragmented nature of the global TV market, a hangover from decades of divvying up secondary broadcast rights to the highest bidder on a territory-by-territory basis.
For example, AMC’s The Walking Dead, which clocked up a cool 768m viewing hours across its 11 seasons in the first 6 months of 2023, is available on Netflix in most markets but not in the UK (where you can stream it on Disney+ or Amazon Prime Video).
In fact, the ‘Globally Available?’ column serves as a reasonable proxy for identifying Netflix-produced titles (which would have been another helpful column) as the vast majority are globally available, whereas the majority of acquired titles aren’t.
2.) Evergreen comedy and drama series are the gift that keeps on giving
I remember discussions when I was at the BBC about the value of availability beyond what was then referred to as the ‘catch-up window’. Default on-demand availability of programmes on BBC iPlayer started out at just 7 days from broadcast, then got extended to 30 days in 2014 and to a minimum of 12 months in 2020.
As unlikely as it now seems, there were voices then querying whether there was really going to be that much viewing beyond that initial ‘catch-up window’. Surely most people who were likely to want to watch a given series would have done so within those first 30 days of availability?
Of course, the data tells a different story. Waterloo Road, originally broadcast between 2006 and 2015 was the fifth most watched programme on BBC iPlayer last year, whilst Doctor Who, whilst had just 3 new specials and no series in 2022, took the #2 spot.
Netflix has an even stronger track-record of giving old comedy and drama series a new lease of life. Friends is the poster-child of this phenomenon. First made available on Netflix in 2015, 11 years after the final episode aired, it has done incredible business for Netflix. Despite coming off the US version of Netflix in 2020, it still generated 448m viewing hours in the first half of this year in other territories.
Suits appears to be cut from a similar cloth. Produced by Universal and broadcast on the USA Network cable channel between 2011 and 2019 and on Dave in the UK, until they dropped it in 2017. Despite all series not being available globally, Suits still clocked up 599m viewing hours in the first 6 months of this year (and is likely to have generated much more in the second half of the year due to seasons 1-8 being added back onto Netflix in the US).
Largely unknown outside of South America, but coming in at #45 in the global rankings is Chiquititas, a Portuguese-language children’s musical soap opera, released every weekday between July 2013 and August 2015. Its 545 episodes generated 162m viewing hours in the first half of this year, despite only being available on Netflix in Brazil.
3.) Some of the shows you most associate with Netflix are less viewed than you might imagine
Whilst this dataset confirms the enduring success of some of Netflix’s most iconic shows (e.g. Bridgerton, Stranger Things, Better Call Saul - none of which had new episodes released in this reporting period but all of which exceeded 250m viewing hours), it also suggests some of the streamer’s flagship titles have less enduring appeal.
Season 5 of The Crown was released in Nov 2022 but all 5 available seasons generated 215m viewing hours in the first half of 2023. Not a small number but less than Peaky Blinders (which finished airing in April 2022 and yet clocked up 244m across 6 seasons in H1 2023) and with a high cost per viewer hour (£407m had reportedly been spent on the production of The Crown by the end of Season 5).
Meanwhile Squid Game and Sex Education, which hadn’t had new episodes since Sep 2021, generated 87m and 77m viewing hours respectively in H1 2023. No doubt partly a reflection of how many people binged them when they were first released but also potentially a reflection of them being less rewatchable than some other titles.
Ignoring Netflix’s liberal use of its ‘Netflix Original’ label (which it had the audacity to apply to Peaky Blinders), here’s a snapshot of H1 2023 viewing hours for some of Netflix’s most iconic homegrown shows:
4.) You don’t need exclusivity to rack up big numbers for kids shows
CoComelon began life in 2006 as a YouTube channel and its content has been widely licensed by its current owner Moonbug. However, that wide distribution didn’t stop it from racking up 601m hours across 8 seasons on Netflix in the first half of this year. PAW Patrol did 392m hours across 8 seasons and 3 movies, whilst Peppa Pig did 276m hours across 6 seasons.
5.) Netflix’s promotional muscle (plus some star power and escapist genres) can make a lot of people watch some not very good films
Perhaps not news (*cough* Adam Sandler) but this dataset confirms that Netflix has the ability to persuade a lot of people to watch some pretty mediocre films.
Whilst the spreadsheet unhelpfully doesn’t indicate which titles are films and which are TV series, making analysis of all films in the dataset tricky, a quick scan reveals 6 movies made it into the top 50.
You’d be forgiven for assuming that All Quiet on the Western Front, released in Oct 2022 and winner of best international feature film at the 2023 Oscars in March might be one of those 6 top 50 titles. Likewise Guillermo del Toro’s Pinocchio, released in Dec 2022 and winner of best animated feature film.
However, you’d be wrong. The 6 films troubling the top 50 are J.Lo action thriller The Mother (with a Metacritic score of just 45%), Idris Elba crime thriller Luther: The Fallen Sun (53%), Chris Hemsworth action thriller Extraction 2 (57%), Jonah Hill rom-com You People (50%), Jennifer Aniston/Adam Sandler action comedy Murder Mystery 2 (44%) and Reese Witherspoon/Ashton Kutcher rom-com Your Place or Mine (49%). Between them these 6 films generated 1.2 billion viewing hours in the first half of this year.
Meanwhile, All Quiet on the Western Front comes in at #150 with 80m viewing hours. Whilst Guillermo del Toro’s Pinocchio comes in at #392 with 42m viewing hours.
Star power is clearly a factor. Don’t-make-me-think escapist genres feels like another.
It reminds me of some research I saw a few years ago on the difference between the TV programmes people tend to say they watch (at dinner parties, in focus groups, on social media) which tended to skew towards the latest high-end drama or Attenborough doc, and the TV programmes they actually spend most of their TV time with, which tended more towards the likes of Death in Paradise and Pointless.
As with most datasets, it will get more interesting once the next couple of biannual reports are out and you can start to look at trends. Here’s hoping Netflix treat us to a full set of release dates and other metadata (format, genres, running time, no. of episodes) to help slice and dice the data (*doesn’t hold breath*).
The BBC used to do (and probably still does do) research into what people wanted to hear on the radio and watch on TV. When asked what they wanted on the radio, the answer was always “more music”, but when asked their favourite radio show the top answer was always (at that time) “Terry Wogan” - who was famous for his loquacious chat, interrupted by the occasional popular tune. In other words, people can only tell you what they think they want when you survey them, but their actual behaviour will often be very different to that and is a better reflection of what’s actually “popular” at any point in time. None of this, of course, helps people trying to survey their way to a new format (in any medium).