(For the last few weeks, I’ve been debuting a series of articles answering a question posed to me by The Ankler’s Richard Rushfield: Will The Irishman Make Any Money? It’s a great question because it gets as so many of the challenges of the business of streaming video. Read the rest here, here, here and here.)
The biggest uncertainty in The Great Irishman Project was figuring out how well the film did with viewers in the first place. I was all set today to parse Nielsen’s estimates from last week, but doing so meant tons of estimating based on a very limited data set. So I waited.
Specifically, I had a suspicion that Netflix would feel forced to tell us something. They couldn’t let Nielsen drive the narrative for their most high profile picture of the year. Sure enough, we got the results today, as Ted Sarandos spoke at the UBS Media conference:
26.4 million subscribers watched 70% of The Irishman in its first week.
40 million are projected to watch in the first 28 days.
Huzzaw! Now we can be a lot more confident in our estimates.
Here’s today’s plan. First, I’ll given you the “Bottom Line, Up Front”. The results and my model. Second, I’ll discuss a few specific estimates and inputs I still had to make. Third, I’ll answer what I assume will be the most commonly asked questions or criticisms of my model. In Q&A format.
(Also, look for my write-up in The Ankler if you’re subscribed.)
Bottom Line, Up Front: Netflix will lose $280 million The Irishman
As I wrote in Part I, the goal was to make a scorecard, and here it is:
For the full model, here you go:
In other words, if this were a big budget tentpole from Disney or Warner Bros–whose flops have extremely public numbers–I think Netflix would have to write down the costs for “only” getting 40 million viewers in the first four weeks. This film was extremely expensive, and it’s already decaying fairly rapidly in viewership. Even with a bump from a Best Picture nomination, Netflix will lose money on this investment.
The Model Details
Even having built the model ahead of time, I had to make some assumptions. Here they are.
Determining US versus International Split
One of the big assumptions of my model right now is that international viewership is much less valuable than US viewership. I do this based on their reporter lower international “Average Revenue Per User” and higher churn rate overseas (from what I’ve been told/researched). As a result, the more US customers for a film (for now) the better it is financially for Netflix.
We have two data points to triangulate the split for The Irishman. First, we can look at historical box office trends of mobster films. According to all the films listed as “Mafia” in The-Numbers database, about 61% of box office comes from domestic versus international box office. For example, a film like American Hustle did $150 million in the US/Canada vs $107 million in the rest of the world. Black Mass from 2015 was even more weighted to the US: $62 million vs $36 million rest of world. (The Departed did $132 million US to $157 million rest of world.) This would imply viewership was skewed to the US.
Second, Nielsen provided their estimates that 13.2 million people watched The Irishman during its first five days of release. That’s almost exactly half of the 70% completion Netflix claims. If we assume this would increase with two more days viewership, again we get closer to 55-60% of viewership being in US/domestic.
I decided to use 62.5% US viewership for Netflix. This is pretty beneficial to Netflix, but makes sense. In all, since US viewers pay more on average for longer, this change benefits Netflix.
Changes to Best Picture Bump
My initial model assumed 25% more folks would watch on Netflix if The Irishman is nominated for Best Picture. I decided to bump this up to 25% first window revenue, since that’s a more accurate reflection of the box office bump. (That’s also a benefit to Netflix’s bottom line.)
Adding in Box Office?
I wanted to add in box office revenue, but Netflix hasn’t released any since this film wasn’t released in the traditional theatrical system. (Netflix rented out theaters and then collected the revenue themselves.) Given the limited number of theaters, I think leaving this out won’t drastically impact the bottom line.
Frequently Asked Questions
I imagine a lot of folks have a lot of questions about this analysis. Let’s try to answer what I imagine are the most common.
What is the best case for Netflix?
The biggest criticism could be that I’m fairly aggressive on the cost of this film. I’ve seen estimates as low as $125 million production budget for The Irishman. If we put in that assumption, a 50% “Best Picture” bump and increase my attribution model, what’s the best case for this film?
Worst case isn’t much worse than what I have. While I could push up the production budget and marketing budgets, those would come directly off the costs, so you could imagine that analysis yourself.
How confident are you Netflix lost money?
Fairly confident. Again, I had to push every assumption to the max in the best case, and Netflix still lost money.
Moreover, the other method of triangulating value tend to say this film wouldn’t generate hundreds of millions of value for Netflix. Take “usage”, a method I recently explained. Or if it were a $300 million box office champion (it likely wouldn’t have been), the streaming window would be about $30 million.
In all, most of the metrics tend to support the idea that this film is popular, but not popular enough to justify its huge price tag.
Isn’t 40 million a lot of people?
It is and it isn’t. Mainly, it’s a number without context. Ask yourself, how many folks saw Jurassic World last year on HBO in the US? On any platform globally? (This has been my go to example because I skipped it in theaters and waited for HBO.)
You don’t know, do you? That’s because we don’t usually parse second window viewership since we have box office numbers. But I guarantee you more people saw Jurassic World on a platform after theaters than saw The Irishman so far. The former film is insanely popular globally. The point is while 40 million is a big number, it’s a global number, which means it’s size isn’t as impressive as it seems.
What about the “strategic” value of The Irishman if it wins a Best Picture?
This is the biggest criticism I see of criticism of Netflix. The idea that any losses are “strategic” and hence justified.
The simple answer is that “strategy is numbers”. Every strategic benefit has to end up on a company’s balance sheet at some point. This means to justify it, you have to ultimately point to some place on the business model that strategic investments justify their costs. Do they boost revenue? Cut costs? What?
While it may help Netflix to be known for prestige, we’d see that show up in increased subscribers. If we don’t, then the impact is small. Or maybe The Irishman helps Netflix secure top tier talent. My counter to that is that Netflix hasn’t had a problem signing up top tier talent because they pay the most. And spending a fortune on The Irishman won’t make it easier to lower costs.
Add it up and it’s tough for me to see how strategic value could make up hundreds of millions in losses.
What is your biggest source of uncertainty?
The attribution table. Which I explained here.
In other words, if say a million customers signed up for The Irishman specifically, then the film would get much, much closer to making its money back. The problem with that sort of thinking is how quickly you run out of “millions” of people to add. Ask yourself this, how many folks have been subscribed to Netflix for 5 years? A lot? Then they aren’t going to sign up for another five years for this film.
Why does it matter how well an individual film did?
Another complaint is that we should just ignore individually and look how Netflix is doing as a whole. This film is part of a portfolio of films, so it’s better to look at how Netflix does holistically.
First off, we don’t do that theatrically released films, though I’m sure studios would like us to. Second off, it doesn’t help talent to ignore individual film performance, so these type of models will continue to be used. Finally, frankly knowing how individual films do is how you build a portfolio model in the first place.
How well did Google Trends do at predicting interest?
Not as well as I would have hoped. Google Trends is pretty noisy. Films that get a ton of buzz (Bird Box, The Irishman) are matched in viewership by films with the same viewership (Murder Myster, Triple Frontier, respectively).
I still like this tool, it’s just a limitation to be aware of.
How well did Nielsen do?
Pretty damn well. Nielsen released that 13.2 million folks watched on average over the first 5 days. Which is about half of Netflix’s 26.4 million admitted viewers. Since the first is US only and the latter is global, well that’s pretty good.
In other words, I plan to use Nielsen data whenever possible. It’s definitely accurate enough for my strategy work.
Are there larger lessons from this model?
Absolutely. Without being too hyperbolic, I’d say that the theatrical vs streaming film models will explain the future of Hollywood.
One of the big debates in the entertainment press this year—and lingering for a few years now—has been the impact of Netflix on theaters. Should filmmakers bother releasing films to theaters when Netflix could get more viewers? This came up with Booksmart, Late Night and a few other others.
Mostly, I saw commentators jump into to opine one way or the other, but I rarely saw any numbers to help us judge. The models I’ve built should help us do that. And over the next few weeks I hope to impact those lessons.
And then I plan to tie these lessons to my series on “Intelligence Preparation of the Streaming Wars Battlefield”.
Can I see your data/model?
As a last point, I’ve gotten this question before and made a few of my models available. I’m debating rolling out a new monetization plan that involves downloading these models. (Right now, I’m a bit behind in monetizing this writing, so I’m exploring other options to keep this website running.)
If that’s something you’d be interested in, drop me a note so I can judge demand.
