Reading "The Streaming Book" by @ballmatthew

What We’ve Learned About Content (Development and Hit Rates)

Data also helped us disprove William Goldman’s famous quote “Nobody knows anything.” To be fair, Goldman’s full quote is more nuanced than those three words: “Nobody knows anything.…Not one person in the entire motion picture field knows for a certainty what's going to work. Every time out it's a guess and, if you're lucky, an educated one.” But data still suggests he was too pessimistic.

Over the four calendar years from 2019 through 2022, Parrot Analytics reports that HBO (inclusive of Max, which pulled down HBO's average) released five seasons of TV that ranked in the top 0.25% of all shows in the United States and 53 seasons in the top 3%. Netflix scored higher than HBO, with ten and 126, but then again, it released 1,017 seasons to HBO’s 145. Put another way, 37% of HBO’s output scores in the top 3% of all TV seasons in the United States, compared to only12% for Netflix. Competitors such as Apple and Paramount have broadly similar shares, 30%, but only 1% of their releases rank in the top 0.25%, versus HBO’s 3.4% (Netflix is also 1%). Showtime and Starz actually have a better batting average than HBO when it comes to their Outstanding or Better hit rate, with 46% and 51%, but they produce only a quarter of HBO’s volume.

Only 30% of HBO’s series rank as “average, compared to 71% for Amazon, 65% for Peacock, 62% for Netflix, 61% for Disney+, 46% for Paramount+, 42% for Apple TV+ and 41% for Hulu. Starz and Showtime continue to beat on percentage terms, with 21% and 25%. Overall, HBO released of 14% of "Outstanding or Better" and 17% of "Exceptional" seasons despite producing less than 7% of total releases.

How does HBO so deftly navigate output with hit rate—and with so many of its hits being home runs? The answer is, unsurprisingly, multifaceted.

One obvious driver of HBO’s success is its brand. Though not all consumers consider HBO to be “for them,” most understand it to be the “best” at scripted programming, and there’s widespread confirmation of this label. Since the premiere of House of Cards in 2013, HBO has received more nominations for Outstanding Comedy Series or Outstanding Drama Series or Outstanding Limited Series at the Emmys than any other network (48 compared to 40 for Netflix, this despite producing a sixth of the output volume) and been nominated for every single categories every single year save for one. Furthermore, HBO has won Outstanding Series a total of 15 times out of 30 categories (i.e. the entire rest of the market, combined, matches HBO's haul).

HBO’s long-earned and frequently affirmed reputation for quality produces a virtuous cycle. Consumers are more likely to try a greater share of the networks’ new shows than any other network, and HBO’s brand appeals to creative talent, too. The network is the preferred buyer for most writers and showrunners, which means that the network is typically given an opportunity to buy every competitive pitch and can win even when it falls short of the highest bidder. Similar advantages apply to actors, directors, and other below-the-line talent. The cumulative result is HBO has a far better-than-average shot of producing the next hit show, successfully launching this show, and keeping it affordable.

Supporting HBO’s hit rate is its approach to development. The network is extraordinarily selective, frequently passing on shows almost no other network would. In 2012, long before Hollywood A-listers starred in TV, HBO filmed a pilot for—but ultimately did not order a series adaptation of—Jonathan Franzen’s National Book Award–winning novel The Corrections. The pilot was written by Noah Baumbach and starred Chris Cooper, Dianne Wiest, Ewan McGregor, and Maggie Gyllenhaal. In 2019, HBO also shot the pilot for, but ultimately canceled the first Game of Thrones prequel, Bloodmoon. No matter how middling the first episode, it’s impossible to imagine HBO’s competitors making a similar move (at a minimum, they would redevelop the pilot). And the cost of the about-face was tremendous: Bloodmoon was set to debut in 2020, only a year after Game of Thrones had completed. In the event, both fans of the show and the network itself had to wait an extra two years until House of the Dragon was ready to go. This was particularly devastating for HBO’s new owner, AT&T, which was set to launch HBO Max using the blockbuster spinoff. Still, HBO pushed through the cancelation. In 2021, HBO surprised Hollywood when it chose not to pick-up a second season to the widely acclaimed, nominated, and hit series Lovecraft Country. The show had been greenlit as a limited series - and had exhausted its source material - but most imagined the first season's success would require a second season. HBO did invest in treatments for a second season, but, ultimately decided they did not meet its bar.

Another attribute of HBO’s development process is its infamously slow development process. On occasion, the network has lost previously bought options and passed on projects that went on to be hits for its competitors, because it wouldn’t order a project to production fast enough. Sometimes this was because HBO, which releases barely a dozen series per year, lacked room on its slate. Other times, the network wanted more time to develop/work on a prospective series before deeming it ready for production.

HBO’s willingness to pass on, cancel, and slow-roll projects helps with quality, but should also make the network less appealing to talent, not their top pick. But where HBO really thrives is the specifics of its development process, which cannot be rapidly replicated by competitors (even when they poach HBO’s staff) and is deeply respected by talent.

In contrast to other networks, HBO has a particularly low ratio of development executives to individual projects. These executives will typically stick with a series through its lifetime, rather than launching the series and handing it off to a colleague to steer thereafter. The average tenure of a development exec is also much longer. Accordingly, executives build greater trust with their colleagues and creative partners, better absorb (and pass on) HBO’s developmental approach, and enrich the series they help to shepherd. These many advantages convert into a final one—the development of longstanding relationships with top talent. One brilliant example is David Simon, who has worked exclusively with the network since the 1990 and across eight series including The Wire, Tremé, and We Own This City. When top talent does well and is treated well, some will never think of going elsewhere, let alone starting a bidding war.

Although Netflix’s hit rate receives considerable scrutiny, it’s helpful to contrast the company’s strategy, reality, and needs with those of HBO, not just their cumulative output. HBO essentially pioneered the high-budget serialized scripted format—and for years, faced only Showtime as a competitor for the next great series. Netflix, meanwhile, joined the fray several years after “Peak TV” had begun, fighting off not just brands such as HBO, Showtime, and Starz but also FX, AMC, and even the Discovery Channel as well as new entrants such as Xbox and Amazon Prime. Accordingly, many in Hollywood considered Netflix a back-up option, rather than a priority—which is why Netflix had to offer the producers of House of Cards a two-season commitment in order to win the series away from HBO and FX, which had offered only pilot commitments. Because Netflix was a new entrant, it also lacked a history with creative talent and had no development pipeline on which to draw, much less dozens of extant talent deals, a catalogue of adaptation-worthy IP, or filing cabinets full of options to adapt various books, articles, and the like.

Netflix faced other constraints. The service’s rapid growth, as well as the inevitability of its top suppliers turning into competitors, meant that Netflix needed to grow its original output rapidly year after year after year after year. Whereas HBO might need as few as 2–3 new series per year, Netflix needed to launch several times as many series as HBO offered overall. To do so, Netflix frequently picked up projects that had been partially developed by other networks but had failed to receive a series order. Another common move was to revive a series that another network had canceled. Often, these other networks made the wrong decision - AMC's success was possible largely because every other network passed on The Walking Dead (originally at NBC), as well as Mad Men and Breaking Bad; both You (cancelled by Lifetime) and Cobra Kai (orphaned by the end of YouTube Originals) have been big hits after their shifts to Netflix. But it is logical that, on average, these moves lower a batting average. At Netflix's volume, though, this was a winning strategy - especially since a breakout hit would attract millions of new subscribers that would go on to watch the rest of Netflix's series.

And where most networks, especially HBO, would only commit to developing a pitched series—or at most, filming a pilot—before ordering it to series, Netflix would jump straight to a series commitment. These moves saved time but they had the inevitable effect of reducing the network’s batting average. To appeal to the creative community and stand apart from traditional networks, Netflix also touted its relatively hands-off approach to production (“no notes”). There are countless examples of studio involvement harming a creative project, but most agree (as Netflix seems to today) that expert input can elevate even the highest-potential series. The magnitude of Netflix growth—not just in volume, but also geography—made it impossible for development executives to focus exclusively on a few projects, especially over their entire runs of their proliferating series. It was similarly hard to build a history working with specific peers—or even in a specific genre or market. Over an eight-year period, Netflix hired more than six times as many development executives as HBO had in total (each of whom averaged more than five years in tenure).

As growth in Netflix’s output has slowed (and it looks likely to shrink in 2023 versus 2022), team tenure increases, organizational models stabilize, and its development pipeline (inclusive of book/script options and talent deals) elongates, hit rates should increase. Indeed, they have; the share of Netflix’s new shows in 2022 which ranked “Outstanding or Better” was double that of 2019. Furthermore, we should recognize the smarts– and success – of Netflix’s aggressive output, rather than just the hit rate. Netflix knew it had a limited window before competition caught up to it, and that what it lacked in pipeline, IP, and track record, it could offset with spending its competitors couldn’t match. From 2019-2022, Netflix released a third of all "Outstanding or Better" and "Excellent" seasons of TV - twice as many as HBO, or as many as HBO and Amazon Prime Video combined, or HBO and Disney+ combined. At the "Excellent" level, Starz, Showtime, Peacock, Paramount+, Hulu and Amazon Prime Video combined barely equal Netflix. Note, too, that among all networks, Netflix ranks behind only HBO for Emmy nominations - and by only 20% - since the premiere of House of Cards in 2013. Third place, FX, is 54% behind.

Armed with this understanding of the dynamics that drive hit rates, it should be clear that content spend is a flawed way to assess a service’s offerings. Comping services based on spend treats content as a fungible commodity, which we know is not the perspective of users—nor does it track with common sense. Content is an art, development is a skill, talent doesn’t just pick the highest bidder, and marketing matters too, not just service size. When applied to tens of billions of dollars spent over a decade (sometimes $100B+), very minor differences in hit rates will produce very different libraries. For that matter, GAAP accounting does a terrible job of capitalizing original content spend. This model essentially values all content as having the same “useful life” (shelf life), irrespective of whether it was based on IP or an original concept, renewed or canceled, loved or ridiculed, forgotten or enduring, heavily marketed or quietly dropped. No reasonable person would conclude that the value of Service A’s library versus Service B’s library is primarily a function of money spent. To this end, GAAP accounting assumes that a given film or series is more valuable when its budget increases - and linearly so.

The differences in the long-term value of content have helped return incentives-based compensation to the video business. Historically, “talent” participated in the financial success of a film or TV series (typically known as “backend”). This was made possible because a title would have myriad different revenue events, each one individually valued, that spanned many geographies, terms, business models, and formats. For example, NBC might buy the rights to air new episodes of Friends—but not the syndication rights, which would be sold to a series of other networks and for decades after the original airing. The rights to Friends would also be sold abroad, and there were also VHS/DVD sales and rental revenue to think about.

For the first five or so years of Netflix and Amazon’s foray into originals, the streamers would often buy an original “outright.” This meant they paid a fixed markup (“cost plus”) for a film or season of TV, and the associated talent would receive no additional payments, irrespective of the title’s success. Though immediately controversial, there was logic to such an approach. Most of the early streamers had designs to be fully global and so sought global rights, thereby closing out one key ancillary window. And even when Netflix or Amazon didn’t buy out home video rights or pay-TV rights, the associated revenues would be modest—how many American households even considered buying House of Cards on VOD or DVD when the entire series, alongside thousands of other titles, was available on Netflix for $10 per month?). And while it was easy to differentiate a hit streaming title from a middling one, the title had no directly attributed revenue. As such, any payment to talent would be opaque and thus controversial.

But for years, more and more performance-based compensation has crept into SVOD deals. There are now bonuses for the number of seasons ordered, award nominations and wins, and the number of weeks of top 10 residency and rank throughout the overall year. Casts of hit shows often renegotiate their salaries well beyond contracts (Puck has reported that Winona Ryder, as an example, made $800K for Stranger Things Season 1, $2.8MM for Season 3, and will make $9.5MM for Season 5). While the creatives behind Squid Game doubtlessly feel that their bonuses pale when set aside the value the show made for Netflix, the series was originally envisioned as limited one—so all of its talent were “off deal” when negotiations for a second season began. Bloomberg reports that a “big raise” is due for Season 2.

Although performance compensation might be opaque – Disney has a complex system of “Series Bonus Exhibit” points that is used to determine how much “backend” a title makes and who reaps these rewards) – the rationale is not. Humans respond to incentives and streamers have found that the more aligned the incentives, the better the product and the more likely an ongoing partnership.” “I always think back to the days when I delivered BJs Pizza in high school”, Blumhouse founder Jason Blum wrote in a 2022 New York Times Op-Ed, “I sprinted to my car and rolled every stop sign to hopefully earn bigger tips. If I was delivering for Grubhub today, where the tip is often determined before the delivery, I would take my sweet-ass time. Same with movies”.

Writer/director/actor Ben Affleck shared similar thoughts with Hollywood Reporter earlier this year: I was talking to [cinematographer] Bob Richardson. He’s a genius. And I said, “Bob, what if I gave you a million bucks to save me five [million]? Could you do it?” And he goes, “Fuck, I’ll save you 10.” Affleck later added another example: “I was on [my 2016 movie] Live by Night, and they were dressing an extra — it must have cost $700 to dress this extra in the period. Five hundred feet away from the camera. And we were waiting while they did the touch-ups. And it was just like, “Guys. This is not meaningful, but it’s taking away from the time and the resources we have to do something authentically enough that it moves the audience. They don’t care if the curls are 1930 or 1920.”

While Blum and Affleck’s examples are focused on the easy-to-calculate costs of time and production, the same point applies to the less tangible aspects of effort, and which network/studio a performer, writer, and director chooses to work for in the first place.

Thus far, the “old ways” of releasing content, developing content, and paying talent have endured. But what about streaming-era experiments with budgets? No one could believe Band of Brothers cost $10 million per episode in 2000. And if it hadn’t been extraordinary—23 years and thousands of original series later, it still ranks 2nd on IMDb—it probably wouldn’t have paid off. To this point, the sequel, released in 2010 for twice the price, is not seen as a success and HBO passed on making a third entry.

When The Crown premiered on Netflix in 2016, many wrote about its financial excess. The seemingly-niche period drama cost a reported $10 million per episode, nearly 25% more than Game of Thrones. What Netflix CEO Reed Hastings, in turn, professed to be “interested in is, how do we expand…and figure out what $20 million-an-hour television looks like.”

In 2022, such budgets are commonplace. Disney+ releases a half dozen Marvel and Star Wars series per year at the $15–20MM per episode mark, while Stranger Things Season 4 cost $35MM and Lord of the Rings as much as $60MM. While HBO passed on Band of Brothers III, Apple picked it up without the Band of Brothers branding (it will air as Masters of the Air in 2023) and, in part due to COVID overruns, is estimated at $40MM each. Yellowstone spinoff 1883 cost Paramount+ as much as $24MM per episode in its first season

As mentioned in the data section, these budgets are possible because of the role these series play for a subscription service. For a significant number of viewers, Stranger Things is the difference between having Netflix and not having it. This means that the title delivers not just $15 monthly instead of $0 but also the opportunity for subsequent subscriber months, which affords those users the possibility of discovering (and thus monetizing) the rest of the service’s slate.

Finally, the era of content-based competition has revealed a familiar truth: IP is, yet again, more valuable than we imagined. The reasons here are many. The easiest way to produce a series that a subscriber absolutely has to watch is to base the series on a story they already love. And because the SVOD business model can now afford $20MM shows, the biggest stories in the world—typically sci-fi epics—can now be made for the small screen, not just the silver one. On-demand distribution, meanwhile, means that these series can be released nearly year-round—and even connect into one another.