If you judged importance by following my Twitter feed, the most important story of the week is Netflix and Netflix and Netflix. For business leaders plotting the future of entertainment, though, remember to always look for the “signal” through the noise. A lot of Netflix news is Netflix noise. “Buzziness” may justify Netflix’s original programming goals, but it doesn’t tell us what stories really matter. (But yeah, I’ll have a Netflix take later.)
Most Important Story of the Week – Comcast NBC Universal Announces Free Streaming for Comcast/Sky Customers (and ads)
Sometimes, disagreements about the strategy of a company boil down to disagreements over who a company should be targeting with their newest products. For instance, at first, I was really skeptical about Quibi, the short-form, subscription video service. (This was a hold-over from my skepticism for Vessel.) My main criticism is I don’t think it will work on TV sets in living rooms. But that’s not Quibi’s plan: they’re focusing on mobile to reach even-younger-than-Millenials. In that sense, my critique of their distribution strategy doesn’t make sense.
That’s why I thought some of the criticism of Comcast NBC-Universal didn’t make a ton of sense either. (Beyond the criticisms that are just, “If you aren’t Netflix, you have already lost.” I can’t really debate that.) Instead, I think a lot of the criticism compared NBCU’s new plan to Netflix, when first you need to ask, who are they really going after here? Are they they same segment?
To evaluate a strategy fairly–and many times in business we don’t do it fairly–starts with understanding who they are targeting, then judging the tactics based on that plan. Or you explain why they shouldn’t target a given segment. The disingenuous way to do this is to assume a company should target a different segment, then evaluate their tactics in that vein.
With that mini-preamble, who is Comcast NBC-Universal (NBCU from here on) targeting with their latest offering?
This is where it gets tricky, as NBCU has both B2C (business-to-customers) and B2B (business-to-business) masters it is trying to serve. Starting with the customer side, the generous interpretation is that NBCU is trying to focus on customers who haven’t cut the cord yet. Essentially, get them used to streaming by offering it to them for free. (This could also be a different segment entirely, focusing on people who want a free streaming service.) In other words, making a streaming service for older-than-Millennials who already have cable.
In a lot of ways, this reminds me of the “TV Everywhere” push of the mid-2010s, just more centralized. TV Everywhere failed because it had too many offerings (an app for every channel and cable company), confusing offerings (5 rolling episodes), no guiding force (every channel was on their own) and lack of in-house technology and data analysis. This deficit extended from NBC Universal to Fox to Disney. That said, the purpose of TV Everywhere made sense. Even if this is just “TV Everywhere on steroids” or “alt-Hulu”, the focus on adding value to the traditional TV bundle could work.
Of course, the second set of masters for Comcast will appreciate this too. That’s all the MVPDs that Comcast risks offending by offering this new streaming service, including it’s own cable/satellite services. The problem plaguing the traditional studios is how to respond to Netflix while not trading streaming revenue (that is actually negative cash flow) while forgoing valuable subscriber fees (that is a huge free cash flow positive). The potential answer from Comcast seems to be a giant punt on the issue, which could be brilliant. If it works–a big “if”–then they’ve essentially cracked the most difficult nut of the whole “traditional studio with network transition to digital” piece.
Further, if “subscribers” are the only metric of performance that matters then with a stroke NBC-Universal can take a lead in the streaming wars. Of course, the skeptic could and will say, “Sure they claim 50 million subscribers, how many use the service?” But neither Netflix, Amazon or Hulu has released Monthly, Weekly or Quarterly users yet. Why should Comcast be the first? In the meantime, we’ll have to triangulate with device installs, Nielsen/Comscore measurements and new subscribers to triangulate. But we won’t know for sure.
(Final note: Using the (3C–STP-4P Marketing Framework for the new conglomerates streaming platforms is a tremendously useful way to look at this problem. That will be fun, and take me weeks to make. Expect it in March or later.)
Data of the Week – The Extremely Ordinary Content Economics of Netflix
Where are my thoughts on Netflix raising prices? Well, my rule of thumb is if I write 2,000 words on something, it becomes its own article. So tomorrow I’ll release my thoughts on the Netflix’s price increase. That would have been a candidate for “Most Important” event in many weeks, but the NBC-Universal announcement bumped it. Earnings reports usually don’t make it in, unless they have ground-breaking news.
Instead, let’s talk about the latest Netflix “datecdotes”.
If you aspire to make great decisions, you need to understand “Bayes Theorem”. Everyone from statistics professors to super-forecasters to Nate Silver would tell you this. This means, when taking in new information, you evaluate it based off your prior assumptions. If you believe something is true 95% of the time, then you need A LOT of contradicting information to overturn that assumption.
Take Netflix: a lot of press coverage treats their performance as otherworldly, as if their movies and TV shows that don’t follow the traditional release patterns of past movies and series. If we’re using Bayes Theorem, though, we would need a lot of information to justify that. Here are two examples from the latest earnings report that say that Netflix is just ordinary with its content.
Data Point 1: Netflix films follow the “logarithmic distribution of returns”.
In other words, Netflix is extremely average at producing filmed entertainment. Average in logarithmic terms, which means 1 out of 10 companies is extraordinary and the rest are just average or ordinary. This seems very bland, but is actually revolutionary when compared to the Twitter/critical coverage.
Anytime Netflix releases the data on a film release, inevitably it is hyped as a positive. Take Bright last December. Eleven million people watched it during the first three days. Was that good or bad? Many journalists said “great!”. In 2018, Netflix kept this train of announcements going, saying that 5 million watched The Cloverfield Paradox, 80 million customers watched a rom-com, The Christmas Chronicles had 20 million streams and Bird Box had 45 million customers in its first week. In each case, some journalist inevitably wrote, “Netflix is a juggernaut.” This was amplified on Twitter.
Yet, 2018 was really the first year that Netflix felt like releasing datecdotes quarterly. With all those datecdotes, we are getting close to a data set. Yes, it’s not perfect and to be clear, we need to make a ton of assumptions to fill it out. But we have some really good guesses we can make now.
So here goes: here’s a count of every Netflix original film sorted by unique (assumed) customers watching in the first 30 days. Netflix released over 80 films in 2018, but I’ll push my data set to include films released in November 2017 and beyond (about 86 films). For films with only one week of data, I’ll assume a 2x multiple for first 30 days. (See next section.) As we can see, Netflix released one blockbuster, and a few hits, the rest being (presumably) duds. Check it out:
I can hear the criticism: I had to guess and assume on a number of those shows. Fair, but the assumptions are fairly conservative. For instance, Netflix said that 80 million people watched one of their romantic comedies over the summer. Then, Bird Box did that by itself. Are we really to assume that each romantic comedy did a Bird Box in viewing in the first 30 days? That’s unrealistic. So it makes sense to assume two movies were “successes” while the rest mostly came in went, especially as the Netflix algorithm punishes movies that don’t take off quickly. (And had say 20-25% of viewing after the first 30 days.) Since we have week one numbers for Bright and Cloverfield Paradox, those are easy extrapolations to 30 days. Finally, I’ll assume Christmas Chronicles did about 40 million customers. (Assume that 20 million streams equates to 20 million customers, with some roughly doubling over the next 30 days. Even that, though, is probably generous to Netflix.)
Want to double check my math? To Google Trends…
Does one film stand out? So yeah, Netflix has released one blockbuster film out of 86 tries.
There is one other key assumption in this: if something did well, Netflix told us. The iceberg principle on steroids. Given that I believe anyone in power follows this rule–you tell the public good things; you hide all the bad things–this is a safe assumption. If you believe the best in corporations, well…I’m not finishing that sentence.
Back to the point that started this section: the logarithmic distribution of returns is WILDLY proven in entertainment. Repeatedly and in multiple parts (film, TV, music, podcasts, celebrity, TV channels, books, concerts, plays, etc, etc). This isn’t even an original thought by me. It’s literally the thesis of Professor Anita Elberse’s book. We’d need a lot of information from Netflix to assume that they have somehow short circuited this law that has held for most of entertainment. Given that the data they have publicly released definitely aligns with this model, it’s much more likely that Netflix isn’t able to make better shows, but to make a lot more shows.
So to say again, all the data Netflix supposedly uses to develop movies doesn’t actually translate into a higher hit rate for its films. They are very average.
(Further reading: TechCrunch had the best run down of data and Recode had a good summary from last week’s earning report. Shira Ovide had the best paragraph on the data she is looking for from Netflix.)
Data Point 2: The Decay Curve is Real
I’ll be quicker on this one, but Bird Box had over 45 million customers in week 1, then 35 million combined in weeks 2-4. Assuming the decay kept increasing over time, that means it looks like this:
In other words, the decay is real. But normal! This shape of distribution isn’t that unheard of in that most major theatrical releases follow the same decay. That’s why you can take the opening weekend and double or triple the numbers to get to the eventual box office. That looks to be exactly what Bird Box is doing.
To leverage Bayes Theorem, this discounts a point I have been making: Netflix films don’t decay any faster than traditional theatrical films, they decay at what looks to be (with one data point) a normal decay curve. They definitely have exceptions on either side, but that’s normal too. So I’ll admit a mistake on my part: I’ve been assuming their decay is faster than normal, and I don’t have evidence for that.
ICYMI – My Articles from This Week
Since some people come to this website for the first time from my weekly updates, here’s a summary of the articles I wrote this week:
I explain one of the “themes” of this website: always put your strategy in numbers terms. It’s easy to criticize or compliment a company if you don’t have to use numbers. But in business, like the army, strategy is numbers.
I set out to imagine a world where six companies controlled all media, entertainment and technology. As a hypothetical, it was just for fun. But I learned some lessons (while creating a map of the current entertainment universe I plan to leverage in future articles) that I explained in Part II.
Other Candidates for Most Important – Other People Enter the Fray Too
Was NBCUniversal the only company entering the streaming wars? Of course not. We had at least four new offerings, potential acquisitions, or tweaks to offering. Going rapid fire…
The key to this offering, in my opinion, is integration. If it is just another channel on Amazon Channels, integrated with all of Prime Video, then it could be another fine ad-supported offering. However, it looks like it is currently another app entirely, so I’m even more skeptical. (But I haven’t confirmed this as on FireTV is may be another channel a la Amazon Channels.) Either way, this is a company (IMDb) that has little experience with video. There was a time when IMDb was a website to look up information about actors in movies. Now it’s a webpage with tons of ads that you have to scroll to find actors. That’s a lack of focus by both IMDb and Amazon.
Put a giant “TBD” under his offering too. On the one hand, this may be a smart way to leverage Sinclair broadcasting across America via streaming. On the other, the streaming channels currently offered are fairly unpopular in a broad sense, but like many launches that could change as they bring on partners. I still need to download this to test its functionality, but for now, monitor the types of partners it brings on board.
Take this new approach with the biggest grain of salt. It really isn’t much of a new offering, but a marketing program to get customers to sample new programming. It does also another way to subscribe to SVOD channels in one place, so in the future have no doubt that we will have a ton of options for that.
Long Read of the Week – Value of SEC Football for CBS on Athletic Director U by Dr. Steven Dittmore
I’ve gone long again this week, so I’ll just say check out this great piece using Nielsen data and past college football TV rights deals to evaluate the monetary value of the SEC for CBS. Dr. Steve Dittmore article shows the value of “Theme 3: Strategy is Numbers” I wrote about above. Basically, you could write this entire article without numbers, but with them you have an actual makings of a strategy, as opposed to a narrative. Great read.