Revisiting Growth vs. Optimization
Is Global Scale the Only Path to a Successful Streaming TV Business?
[Note that this essay was originally published on Medium]
Tl;dr:
I recently posted an essay arguing that in light of growing evidence that streaming will be neither as large nor as profitable as many expected, the streaming TV companies will need to shift their focus from subscriber growth to optimization.
I received a lot of feedback, including some thoughtful disagreement, namely:
1) My analysis was too U.S.-centric, while the real play is global.
2) The streaming market is likely to be winner-take-most and it will necessitate global scale to win.
3) Optimization — which definitionally means making different decisions in different contexts — is not a practical strategy because it will confuse and frustrate consumers and bog down companies’ decision making.
This pushback essentially frames the key debate in the TV business these days: Is “go-big or go-home” in streaming the only winning strategy, or is a more balanced approach viable? Call it the Jason Kilar v. David Zaslav debate.
In this essay, I address the pushback, where I think it is right and where I think it is wrong.
One of the reasons I publish online is to get (constructive) pushback. It highlights where my argument was unpersuasive, incomplete or flawed. I usually learn something new.
I got a lot of feedback to a recent post titled Media’s Shift from Growth to Optimization, including some very thoughtful disagreement. Before getting to that, here’s a summary of the essay:
For years, it has been obvious that linear TV profits will decline and streaming profits will grow. What has been unclear is the degree of each of these. Consequently, it hasn’t been clear how the transition to streaming will affect the aggregate TV (linear plus streaming) profit pool.
What’s new in the last six months is that it is becoming clearer that the total addressable market (TAM) of streaming in the U.S. will be smaller than hoped, in part because churn is higher than expected.
So, while the decline in linear continues unabated, or is possibly accelerating, streaming is likely to be less profitable than many expected. (For more on this, also see Is Streaming a Good Business?. It shows that — based on Netflix’s daunting unit economics — most of the other streamers can’t make money at current price points.)
This forecast of a smaller aggregate profit pool for TV presents more challenges for some streamers than others. Besides Disney, none of the media conglomerates are likely to retain the same share of the streaming market that they have in linear TV.
As a result, the streamers will need to increasingly turn their focus from streaming subscriber growth-for-growth’s sake to what I refer to as optimization: revenue (pricing, advertising, tiering, password sharing, etc.), acquisition and retention, cost (especially content optimization) and portfolio optimization.
The Pushback
No one disputed my characterization of the challenges in the U.S. streaming market. Here’s the pushback I got:
My analysis was too focused on the U.S. market, but the real play is global. Even if the pay TV/SVOD market is saturated in North America and Western Europe, the global TV market is still growing owing to rising broadband penetration and disposable incomes in APAC (excluding the ever-impenetrable China), CEE and Latin America. As important, as TV continues to shift to streaming, the total profit pool is up for grabs and global players will be able to take share from local and regional players.
The global TV business will be a winner-take-most market, making scale paramount. It’s critical to achieve global scale because only the scaled players will be able to compete for content and leverage global investments in product and data.
Optimization isn’t a practical strategy. By definition, optimization means taking different approaches with different subscribers, content assets, network brands, distribution platforms, marketing channels and territories. The concern I heard is that anything less than a full-throated commitment to streaming will fail, because anything else will confuse and frustrate consumers and paralyze employees.
This counterpoint frames the chief debate in the TV business these days.
In essence, this counterpoint frames the chief debate in the TV business these days. Is “go-big-or-go-home” in streaming the only option, or is it possible to succeed with a more balanced approach?
My Response
Here’s my response, point by point:
“The real play is global…”
Point taken, there is a big global opportunity… Admittedly, my analysis focused almost entirely on the U.S. market. There is no question that the TV market is larger outside the U.S. than within it. According to Magna Global, the TV ad market is roughly $100 billion outside the U.S. (vs. about $70 billion within it) and S&P Global Market Intelligence also pegs international pay TV subscription revenues at around $100 billion. It is also fair that streaming opens up new opportunities for the big U.S.-based media conglomerates. Many international territories have relatively low pay TV penetration and/or only a few (or even one), dominant pay TV distributor, so it is difficult or impossible for the big media companies to profitably launch and scale networks in these markets. The advent of D2C makes it feasible to disintermediate these dominant local distributors.
Figure 1. In 2018, ~59% of Netflix Content Amort Was International
Note: Netflix last disclosed this breakout between domestic and international in 2018. Source: Company reports.
…But reality intrudes. It would be nice to have the resources to compete on a global scale, just like it would be nice to be 6’4” with a full head of hair, 7% body fat and a trust fund. It is very expensive to build scale internationally. Besides the cost to staff and launch new markets, the requisite content spend is out of reach for most of the conglomerates. International audiences expect both high-quality foreign content and substantial amounts of locally-produced content. The last time Netflix disclosed the allocation of its content amortization between “Domestic Streaming” and “International Streaming” was in 2018. At that time, International represented ~59% of content amortization (see the bottom of Figure 1). That proportion has likely risen to 65% or more since then as the company has ramped up local content investments. With ~$18 billion of cash content spend last year, that equates to ~$12 billion internationally. This is multiples of what any other streamer (perhaps other than Disney) spends globally. Peacock plans to spend $3 billion on content this year; Paramount forecasts $6 billion of global content spend on streaming by 2024; and HBO Max is probably spending a few billion. These companies simply don’t have the financial wherewithal to spend on the same scale, or at least not without crushing earnings.
In the absence of global scale, media conglomerates should use the breadth of their assets. So, most of the media conglomerates cannot realistically compete on global scale. They are also burdened with a more complex portfolio of legacy assets. But they should try to turn the breadth of these assets, and long-standing relationships, into an advantage where they can. For instance, they should re-purpose content across SVOD, AVOD, broadcast and cable networks; leverage decades-long relationships with pay TV distributors, theatrical distributors and agencies; cross-market across their networks and D2C services; attract talent and projects by cutting more talent-friendly deals that provide some backend participation and, where appropriate, commit to theatrical releases; and even bundle multiple products and services, as Disney is reportedly considering. Is that better than having 250 million subscribers globally? No, but it’s something.
“TV is a winner-take-most market…”
No, it probably isn’t. It makes sense to “blitzscale” — i.e., to prioritize user growth above all else — in the formative stages of a winner-take-most market because the first business to achieve scale can establish an unassailable competitive position. But here’s why the TV networks business is not likely to be a winner-take-most market.
TV networks have limited network effects. Winner-take-most markets are generally characterized by very strong network effects, or demand-side economies of scale, and high multi-homing costs. (By “strong” network effects, I mean that the value of the network to all users continues to increase even as it approaches high penetration of the addressable market.) In other words, the largest network provides so much value that no one else comes close and it is difficult or costly for users to simultaneously patronize multiple networks. TV networks/SVOD platforms have neither strong network effects nor high multi-homing costs. Subscribers derive limited benefit as the number of users scales, other than some social benefit of watching the same programming as other people and perhaps the opaque benefit of better-trained recommendation engines or larger investment in the UX. It is also easy to subscribe to multiple services.
TV networks is not likely to be a winner-take-most market.
No one will realistically be able to corner the market for content. While TV networks don’t have significant demand-side economies of scale, because they are high fixed cost businesses, they have supply-side economies of scale: per unit cost falls as output increases. This is probably most relevant for content spending; Netflix’s $18 billion spend this year is lower per-paid subscriber than Peacock’s $3 billion. Even so, these benefits don’t foreclose competition in a market with a highly differentiated product like entertainment. It simply isn’t practical to corner the market for content for a bunch of obvious reasons (which I’ll enumerate anyway). It isn’t possible financially. (Morgan Stanley pegs global TV and filmed entertainment content spend at ~$150 billion this year and Ampere Analysis estimates it is $230 billion.) While money ultimately trumps anything else in Hollywood, it is not the only factor that talent considers when selecting a distribution outlet. They consider relationships, the quality of the development team, the cachet of the brand and the marketing support. No one provider is likely to excel in all these areas. It also isn’t possible to produce 1,000+ shows globally and maintain quality. (Netflix currently produces 700 shows per year and, arguably, that is already too many.) Also, successful, high-profile IP is spread among numerous companies (Star Wars, Lord of the Rings, Game of Thrones, the MCU, DC, James Bond, Mission Impossible, Pixar, Harry Potter, etc.). And, as William Goldman famously said about the movie business, “No one knows anything.” There is a high degree of luck in creating good content, so even having first dibs on the supposedly most promising content does not prevent competitors from succeeding.
“Optimization is not a practical strategy…”
This also raises a fair point. Optimization will be a lot more complicated operationally than prioritizing streaming above all else. As mentioned, by definition it means making different decisions with different subscribers, content assets, network brands, distribution platforms, marketing channels and territories. This will require very clear decision rules and analytical tools or decision making will bog down in infighting and confusion. For any given content asset, how do you decide whether to license or distribute it on your own platform? How does this answer differ domestically and abroad? How do you window content? What are you optimizing — licensing revenue, advertising revenue, engagement, subscription revenue, acquisition efficiency or retention? What goes straight to streaming and what goes to theatrical? And, in all these cases, who decides? A lot of media companies are not currently set up to process these decisions systematically.
A Few More Clarifications
From my discussions, there are a few other areas where I could’ve been clearer:
Optimization was Always a Matter of When, Not If
Businesses benefit from scale to differing degrees. For instance, Facebook has very strong network effects. Since each human is different, each person added to the network potentially has incremental value to everyone else (even if there is diminishing marginal utility). By contrast, Lyft has weaker network effects. Once there are enough drivers on the network in any given city that the wait time falls below a certain threshold (say, 5 minutes), then the marginal utility of the next driver falls to zero. Your local pizza place might experience supply side economies of scale, and lower costs for mozzarella, if it added a few locations.
Regardless, all of these businesses must eventually shift their focus to making money, even those with the strongest network effects and economies of scale. A16z partner Chris Dixon refers to this as a transition from “attract” to “extract” mode (Figure 2). You could substitute the words “growth” and “optimization,” respectively. My argument is that in the streaming business, that time is now.
All businesses eventually need to make money; I’m arguing that for streamers, now’s the time.
Figure 2. The Inevitable Shift from “Attraction” to “Extraction”
Source: Chris Dixon.
Optimization and Growth are Not Mutually Exclusive
When I referred to a shift in focus from growth to optimization, I didn’t mean to imply the end of subscriber growth altogether, just a shift in emphasis. One quibble I have with Dixon’s chart is that it implies the shift to “extraction” (or “optimization”) occurs when growth reaches the top of the S-curve. This isn’t necessarily the case; it happens when the need to make money exceeds the marginal utility to the business of adding additional users. Facebook is again a good, if extreme, example. In 2012, it went public and immediately felt pressure to prove it could monetize its user base. It launched about a dozen new ad products that year, prioritizing revenue growth over the potential impact on user growth. Its network effects were so powerful that it didn’t matter. Between 2012 and 2021, Facebook MAUs (monthly active users) roughly tripled, from 1 billion to 3 billion, even as ARPU (average revenue per user per quarter) increased from ~$1.50 to ~$10.
Optimization Isn’t a Silver Bullet, It’s a Necessity
At the end of the previous essay I touched on this, but probably didn’t emphasize it enough: optimization will be hard and unglamorous. The Street won’t pay for it until the results show up in the financials. And it won’t magically solve the industry’s problems. My point was simply that there’s no choice.