The Full Season Drop Is Quietly Taxing Platform Revenue by 39 Percent

Based on the research ofZhao, Mehta and Shi, "An Empirical Investigation of a Digital Platform's Release Strategy for Serialized Content," Marketing Science, 2026

The experiment that measured what bingeing actually costs

Clarice Zhao, Nitin Mehta and Mengze Shi got access to something most platform researchers never see: six months of 2017 transaction data from a leading Chinese online book platform that holds a 70 percent market share in the romance genre and charges readers per chapter. Their sample covered 575 readers and 138 bestselling titles in the "Pure Love" category, and it let them watch the same two forces that every serialized content business argues about play out in real purchase and visit data.

The first force is binge consumption. When a full book is available at once, readers consume chapters faster, and faster consumption in one period raises consumption in the next. That momentum raises the odds a reader gets hooked and finishes the whole thing, buying every chapter along the way. The second force is exploration. When chapters arrive on a schedule instead, readers have to come back to the platform to get them, and each of those return visits is a chance to browse other books too. The paper's own numbers make the exploration effect concrete: a notification that new chapters are available for a book a reader already follows roughly doubles that reader's probability of visiting the platform that day.

Those two forces point in opposite directions, so the authors ran the comparison the industry rarely runs on itself. Starting from the platform's actual practice, which is sequential release, they modeled what would happen under a full simultaneous release instead: revenue fell 39 percent, driven mostly by a 60 percent drop in platform visits, because once a reader has the whole book there is no reason to come back for it, and fewer visits means fewer other books get browsed and bought. A stricter, cleaner sequential policy, one chapter released per day for every title, outperformed the status quo by about 5 percent. But the strategy that won outright was a hybrid: release the first 40 percent of a book's chapters simultaneously, then drip the rest at one chapter a day. That combination delivered 16 percent more revenue than the platform's existing practice, because the opening burst still hooks the reader into finishing, while the sequential remainder still pulls them back to the platform again and again. "Faster release gets people hooked on a single book, increasing the chances that they will finish reading it," Zhao said of the finding. "Slower release gets them back on the platform. Those two effects pull in opposite directions, which is why an all-or-nothing content drop is not the revenue-maximizing choice." Mehta put the hybrid logic plainly: releasing the first 40 percent at once is enough to hook readers, and releasing the remaining 60 percent sequentially ensures they keep coming back and exploring other titles.

Netflix proved the binge thesis, then quietly hedged it

Streaming video did not need a structural model to discover the binge side of this tradeoff. Ted Sarandos built Netflix's entire original-content identity on it. Profiled in late 2013 as one of Hollywood's "rule breakers," Sarandos was described as committed enough to all-at-once release that he "likely wouldn't move forward on a series that was not on board" with binge viewing, breaking with a television industry built around weekly episodes and cliffhangers. Years later, with real viewing data behind him, he kept defending the model on its own terms rather than retreating from it. In an October 2019 comment reported by The Hollywood Reporter, Sarandos pointed to roughly 35 Netflix shows that, for licensing reasons, release week to week in some markets, and argued that the all-at-once version of the same show still generated more total viewing and more cumulative social media buzz than the weekly rollout, just concentrated into a shorter window.

What has shifted is how absolute that commitment stays in practice. Industry analysis published by NewscastStudio in mid-2025, drawing on BB Media and Fabric Data tracking of U.S. streaming releases, found that Netflix still releases 68 percent of its original series as full-season drops, but is now routinely splitting the rest. "Beauty in Black" premiered its first eight episodes in October 2024 and held the second eight back until March 2025. "Wednesday" season two followed the same logic, releasing in two four-episode parts roughly a month apart. Neither matches the paper's specific 40-60 split exactly, and a Chinese romance-novel platform is not Netflix, but the structural move is the same one the model identifies: give the audience enough up front to get hooked, then make them wait, and come back, for the rest.

Disney+ and HBO Max bet retention on the opposite instinct

Other platforms arrived at the exploration side of the tradeoff from the opposite direction, as a retention tool rather than a hook. When Disney+ launched in November 2019 with only a single episode of "The Mandalorian" available, Ricky Strauss, then the service's president of content and marketing, told The Hollywood Reporter the choice was deliberate: "We're just getting started, so it made sense to build our fan base with episodic content coming out once a week in a more traditional model. We felt that if we put everything out at once it would be harder for us to build the audience." HBO Max's Kevin Reilly made a similar bet for the same reason, telling investors in October 2019 that the service would offer a "weekly pulse" of originals, arguing that HBO hits such as "Succession" and "Chernobyl" became part of the cultural conversation through a weekly release schedule rather than "fading quickly after a binge and burn." Media analyst Jeffrey Wlodarczak framed the mechanism in almost exactly the paper's terms at the time: "The hope is weekly releases will help reduce subscriber churn, which for these players is likely to be quite high given the lack of switching costs." A subscriber who has to keep opening the app to get the next episode is, every time, a subscriber newly exposed to everything else on the platform, which is the same visit-driven exploration effect the paper measured through chapter notifications.

The real question a platform should be asking is not binge versus drip. It is how much to give away up front to earn the right to make people wait for the rest.

What the paper's numbers say about getting the split wrong

The size of the gap between the extremes is the part platform operators should sit with longest. Going from the status quo to full simultaneous release did not just fail to help, it erased more than a third of platform revenue in the model, because the exploration channel collapsed along with platform visits. Going the other way, to a stricter pure sequential policy, only gained 5 percent, well short of the hybrid's 16 percent, because a reader who never gets a meaningful chunk to binge is more likely to lose interest before the hook takes hold. Both pure strategies leave money on the table relative to the strategy that treats bingeing and exploration as complements instead of substitutes.

The 2025 streaming data suggests the industry is converging on that same curve empirically, title by title, even without the structural model behind it. Prime Video raised its share of all-at-once originals from 33 percent to 67 percent in 2025, while Max cut its own all-at-once share from 50 percent down to 11 percent over the same stretch, according to Fabric's tracking, each platform moving toward the mixed middle from a different starting extreme. Neither is running the paper's exact counterfactual, and no platform has published a revenue breakdown as clean as the one the authors built from six months of chapter-level purchase data. But the direction of travel, and the plain economic logic connecting a return visit to a browsed catalog, both point the same way: the release calendar is not a creative afterthought, it is a revenue lever, and setting it to either extreme is the one choice the data says not to make.