A Netflix Data Dive: What does their “annual” top ten lists reveal about their biz model?

Last December, I unveiled my theory for how big organizations use PR. Big entities—be they corporations, governments, non-profits, even news outlets—share their good information and actively hide bad information. It’s like the iceberg principle on steroids. Especially with digital companies like Netflix:

Slide2

(By the way, in government, the CIA is the absolute best at this. They have feature films like Argo win best picture, then have the gall to go on cable news and say, “You never hear about the good things the CIA does.”)

With this in mind, let’s draw some insights on Netflix’s (kinda) annual tradition to release a top ten “something”. In 2018, they released their top “binged” things. Now they’ve released for both film and TV across three lists in their most prominent territories. Sure, Netflix doesn’t give us much to work with, but I’ll interrogate these numbers to death in the meantime.

The Facts

Before the analysis, though, some facts to keep in mind. Whenever you see data, you should ask the “5Ws” of journalism. Most problems with data come from folks measuring it differently. (If you’re curious, I’ve tried to explain how to understand digital video metrics, and the distinctions, in this big article, which is one of my more popular.) If a news outlet buries these details, you should be skpetical.

– Who: Subscribers
– What: Watching 2 minutes of a given title
– When: During the first 28 days of release
– Where: Country-by-country. I’ll focus on the US, but they released it for a few major territories.
– How: Separated into content types, with all releases, by film/TV, scripted vs documentary.

Here’s a chart, with some additional details of the Top 10 Movies:

Top 10 tablesThat just leaves the why…

Thought 1: If this is the best “datecdote” Netflix could offer, that’s not great

Really, that’s what you think when you see a list that specifically changes the criteria from their previously announced metrics. Netflix had spent all of 2019 giving investors the “70% completion” metric for all their datecdotes. For this release, they dropped it down to “2 minutes of viewing completion”metric.

Using our iceberg principal above, what would the 70% threshold have told us that Netflix didn’t want to know? There’s clearly a narrative they’re deliberately trying to avoid.

Further, why not give us the “most binged” shows again as they did in 2018? Whenever someone changes the data goal posts, you should be very cautious. Yes, you see this all the time in Hollywood when development execs want to greenlight a project. If the numbers don’t look good, they change the measurements to get their greenlight. And yes, this happens all the time in business too. If leaders don’t like the numbers, change the measurements.

But it’s a bad habit.

Thought 2: This new metric doesn’t tie to Netflix’s self-stated goal for monetization.

If you’re looking for more red flags, this is it. In the last earnings call, CEO Reed Hastings said they care more about time on site than anything else. So why not give us that? They have the hours viewed data…they even could have limited it to new releases. (Which would have excluded Avengers: Infinity War, Black Panther, Friends and The Office.) What does the hours viewed tell us that customer counts don’t?

Or take the emphasis on acquiring and retaining subscribers. When Netflix execs speak at conferences, they downplay traditional viewership to focus on how well films bring subscribers to the platform, or keep them there. Clearly completed films would correlate more with sign-ups than only 2 minutes of viewing. (This also jives with my personal experience.)

Thought 3: Netflix Avoided Total Hours Because of Kids Content

I think Netflix avoided “total hours” for two reasons. Let’s start with kids content. Kids rewatch the most content. They don’t watch The Incredibles 2 once, they watch it a dozen times. That gives kids films an edge on viewership hours. Narratively, you don’t want to emphasize how valuable kids content is right after Disney+ launched. As Richard Rushfield has written, something like 60-70% of Netflix viewing may be on “family titles”. That’s a huge win for Disney+ if true.

It also means that if hours on site are the key metric—again as Hastings said in the last earnings call—then kids content seems even more valuable.

Insight 4: Licensed content still made it on.

Netflix also likely avoided the 70% completion metric because they wanted to downplay licensed content as much as possible. Netflix films have a dramatic marketing edge because when new seasons premiere, they get home page, search engine tinkering and top of screen treatment. This doesn’t necessarily drive completions—if shows aren’t good people don’t finish them—but it does drive 2 minute sampling. 

Still some licensed content made the list, even as it was deliberately curated out. Specifically, three of the top ten films and one of the top ten series. I’d argue this is bad for Netflix; even as they tried to weed out licensed titles a few prominent Disney films made the list.

This is more impressive than it seems because the biggest Disney films weren’t even released in 2019. Specifically, Black Panther and Avengers: Infinity War were 2018 releases. Meanwhile, Netflix was stuck with The Ant-Man and the Wasp—one of the lower grossing recent MCU films—and Solo: A Star War Story. Then the rest of the incredible Disney 2019 slate didn’t make it onto Netflix. 

Thought 5: Focusing on 28 days ignores films and shows with longer legs.

Licensed titles, especially big blockbuster films, also have longer legs than new releases. Don’t you think Avengers: Infinity War had some rewatching going on in the run up to Avengers: Endgame’s release? Absolutely. By focusing on 28 days as the time period, it narrows the window for licensed films to rack up viewership. (They also had a fairly crowded January 2019, with three Disney feature films being released in the same month.)

Thought 6: International Originals Still don’t play in the United States.

This might be the most important insight. Because it throws some cold water on this commonly repeated narrative in the media. As far as I can tell, we still don’t have any data that Netflix’s international originals play well in the United States. Not a single title originated outside the United States or was a non-English film/TV seres. The closest was The Witcher, that’s based on a Polish book series. 

In fact, perusing the international lists, they look perfectly ordinary. Which is pretty sub-optimal for Netflix. Netflix prides itself on paying more than anyone else because they buy global rights. This strategy only works if you can make money (which means drives subscriptions) in those new territories. And viewership is highly correlated with subscriptions.

Thus, If Indians don’t watch French films, if no one outside south Asia watches Indian originals, and if only some American films travel, then Netflix shouldn’t pay global rights for all those titles.

(I’ll have more to say on this in future articles.)

Thought 7: Netflix TV shows don’t grow season-to-season and 

I’ve written about this before, but Netflix data shows us the difficulty of sustaining audiences long term for TV series. It’s fairly easy for series to launch well, but then drop off quickly. From the initial data, it doesn’t look like Netflix has an advantage in driving season-to-season retention, and it might actually be worse than traditional networks and cable channels. In this data release, only 3 TV series in season 2 or beyond made it on the list. When you’d actually expect popular series to have an advantage in getting the most subscribers. 

The most notable omission in 2019? Orange is the New Black, whose final season didn’t make the list.

Thought 8: Some films are started, but not finished.

Specifically, the mysteries on the list are why certain films landed above others, when we know their actual 70% completion rate. Specifically, The Irishman had more initial viewers than Triple Frontier, but Ted Sarandos said it was only tracking for a 40 million completion rate through 28 days. (With about 16 days of data when he announced that, that 40 million estimate is fairly reliable.) 

Or take Extremely Wicked, Shockingly Evil and Vile. It never got the datecdote treatment, but it showed up on this list. Arguably, this list is more about marketing than the quality of the film.

Thought 9: I think Netflix had a strong Q4. 

That would seem to be the implication of 6 Underground, The Irishman, You season 2 and The Witcher all making the top 5 on these lists. Given that December is a historically good month for Netflix—people buy new devices/smart TVs and fire up Netflix—this is still a good performance.

My “think” caveat, though, stems from the fact that otherwise the films and TV series are spread out between all the quarters. Q2 had the most films/series on the list with 6, but the lowest was Q3 which still had 4.

  1. I think a looming issue might be the potential for advertising revenue.

    Penetration matters most in that case, of course, and Netflix may just want to emphasize their reach as opposed to the titles themselves. But that also rolls back to your comment on international broadcast rights…

    Are there really that many advertisers that would want global exposure. There are some perhaps (Coca-Cola, Nike?), but most are keen to have a geographic element to their marketing.

    Like

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: