Star Wars
Why Ninja's sudden defection to Mixer made Twitch streamers less productive
Platform Papers is a blog about platform competition and Big Tech. The blog is linked to platformpapers.com, an online repository that collects and organizes academic research on platform competition.
By Jens Forderer, Dominik Gutt, and Brad Greenwood.
It’s well known that superstars command an outsized share of attention and demand from consumers. Movie studios seek the biggest names for their casts, sports franchises compete to sign the best players, and music festivals fight to feature the most popular artists on their lineups. The logic is straightforward: beyond their acting talent, game statistics, or hit records, stars help draw attention to the entire movie, team, or event — effectively growing the pie for everyone involved.
This logic hasn’t been lost on digital content platforms either — and “battlefield” might be a better word than “arena” in this context. In January 2020, Friends was pulled from Netflix in a $425 million deal to stream exclusively on the soon-to-launch HBO Max. In a high-profile clash between Taylor Swift and Spotify, Swift withdrew her entire catalog from the platform, and Apple Music ultimately secured exclusive streaming rights for her 1989 album. Livestreaming was no exception. In a surprising move, Fortnite superstar Richard Tyler “Ninja” Blevins left Twitch to stream exclusively on Microsoft’s competitor platform, Mixer.
But what happens if we view this livestreaming case from the perspective of other creators? How does a Fortnite streamer on Twitch react when their biggest competitor suddenly leaves? This is the starting point of our newly published paper Star Wars: An Empirical Study of Star Performer Turnover and Content Supply on Multisided Streaming Platforms published in Information Systems Research.
Ninja’s presence concentrated Fortnite-loving viewers on Twitch, benefiting everyone by enlarging the overall pie. When Ninja left, many viewers followed him, shrinking the pie for those who stayed.
Before diving into livestreaming, let’s step back and consider what we would expect in a traditional industry. Imagine a local Ford dealership learning that a nearby Toyota dealership has closed. That’s likely good news: with less competition, the Ford dealer could expect more sales without increasing advertising or effort.
Digital platforms, however — and livestreaming in particular — work differently. At first glance, you might think that when a superstar leaves, other streamers would enjoy more viewers. But in practice, viewers can follow a star to a new platform with just one click. As a result, the idea that departing stars leave behind a pool of “excess” viewers waiting to be captured often turns out to be too optimistic.
So, how do streamers actually respond when a superstar leaves? Streamers are highly sensitive to audience fluctuations: they receive live feedback on concurrent viewers, follower growth, and engagement. Do they follow the star to the new platform? Or do they stay and produce more content to attract the remaining viewers?
Platform operators might be well-advised treating top-tier streamers as “critical infrastructure.”
We investigate these questions using daily panel data on Fortnite streamers on Twitch, covering two months before and after Ninja’s departure. The challenge is establishing a credible counterfactual: how would streamers have behaved if Ninja had stayed? To address this, we use German Fortnite streamers as a control group. Livestreaming is well-suited for this approach: due to time zone and language differences — Fortnite is particularly popular among younger audiences — German creators are unlikely to be affected by a U.S.-based streamer from Illinois. Their audience overlap with Ninja’s is minimal.
Using this framework, we find that Twitch streamers produced about 20% less content after Ninja’s exit. Surprisingly, this decline isn’t driven by creators leaving Twitch; rather, it reflects fewer viewers staying on the platform. On average, channels attracted 12.5% fewer viewers after Ninja left, making content creation less economically rewarding. Streamers with more diversified content portfolios — for example, those also streaming Call of Duty or Minecraft — saw smaller declines. Similarly, more popular creators were less affected. Altogether, the evidence paints a clear picture: Ninja’s presence concentrated Fortnite-loving viewers on Twitch, benefiting everyone by enlarging the overall pie. When Ninja left, many viewers followed him, shrinking the pie for those who stayed. As academics would say: Ninja’s departure triggered negative demand-side network effects that reduce other streamers’ content production.
To test whether this effect is unique to Ninja, we replicate the analysis for two other high-profile departures: Shroud and Disguised Toast. The results are consistent — the larger the departing star’s following, the bigger the drop in content supply. This strengthens the conclusion that superstar presence benefits the platform as a whole, while their exit harms it.
A single high-profile defection not only imports viewers but also depresses rivals’ creator output.
Several practical implications can be derived from these results. First, platform operators might be well-advised treating top-tier streamers as “critical infrastructure.” Since every minute streamed on the platform means a monetizable space through advertising, losing minutes (54,142 over 2 months in our sample, to be precise) implies losing potential ad revenue. Retention packages, flexible contract terms, and early-warning metrics for star dissatisfaction may thus become essential risk-management tools.
Second, creators with a narrow, single-game focus need to diversify their content mix. Streamers who already devoted at least a quarter of their content creation to other games saw only a dip in production of around 7 percent, versus ≈28 percent for tightly focused Fortnite channels. Building variety into the channel portfolio is a practical hedge against future star shocks. Similarly, recommending varied content to highly focused viewers might be worth exploring by the platform as well.
Third, smaller or mid-tier creators should invest in audience loyalty and multi-platform outreach. Popular streamers (upper popularity tercile) were essentially insulated from the shock, while those in the middle tercile lost nearly 40 percent of output. Cultivating a direct fan community (e.g., Discord, newsletters) and mirrored presence on secondary platforms reduces dependence on any single star or platform.
Finally, competing platforms (and talent agencies) can use selective poaching as a strategic weapon, but only if the price is right. A single high-profile defection not only imports viewers but also depresses rivals’ creator output, especially when the star is extremely popular. Acquiring stars therefore yields dual value: direct content plus indirect weakening of competitors. Yet, the payoff scales with the star’s brand power relative to acquisition cost and the case of Mixer — that was ultimately discontinued— teaches us that this may not be a silver bullet.
This post is based on research published in Information Systems Research and is included in the Platform Papers references dashboard:
Forderer, J., Gutt, D., & Greenwood, B. N. (2025). Star Wars: An Empirical Study of Star Performer Turnover and Content Supply on Multisided Streaming Platforms. Information Systems Research.
Platform Updates
When megastars drop, timing is everything (1) (2) - Releasing alongside superstars like Taylor Swift can trigger an algorithmic blackout, pushing smaller acts off the grid. Yet under the right proximity and playlist links, the “halo effect” can flip the script—related artists can gain more listeners and a popularity bump.
EA goes private in $55B buyout (1) (2) (3) - Saudi Arabia’s Public Investment Fund, joined by Silver Lake, JPMorgan, and Jared Kushner’s Affinity Partners, is taking EA private in one of the largest tech buyouts ever. The deal gives PIF a controlling stake, Kushner a 5% slice, and access to hit franchises like Battlefield.
Analysts split on EA buyout (1) (2) (3) - Supporters frame Saudi Arabia’s $55B takeover as part of its push into gaming and entertainment—and a chance for EA to regain creative freedom. Critics, however, question the lofty valuation amid slowing growth, warning that debt pressures could lead to further studio cuts and IP sell-offs.
ByteDance keeps rich slice of TikTok’s U.S. profits - Despite owning less than 20%, TikTok’s Chinese parent will capture about half of U.S. earnings through licensing and profit-sharing under a $14B Trump-backed deal. The joint venture, funded by Oracle, Silver Lake, and MGX, satisfies Trump’s divestiture mandate.
Spotify CEO Daniel Ek to step down after 20 years - Ek will exit the CEO role on Jan. 1, 2026, becoming executive chairman as Gustav Söderström and Alex Norström take over. The move comes as Spotify hits 696M users, hikes global prices, and doubles down on curbing AI-generated slop.
Meta to mine AI chats for ad targeting - Starting Dec. 16, conversations with Meta’s billion-user chatbot will shape ad personalization across Facebook, Instagram, and WhatsApp. The move turns shopping chatter into ad insights, widening Meta’s $46.5B ad engine while keeping sensitive data off-limits.
OpenAI’s Sora hits No. 1 on App Store - The invite-only iOS app for short AI videos outranks Google’s Gemini and ChatGPT, powered by the new Sora 2 model. Viral clips, including a fake Altman video, spark debates over safety, likeness rights, and the ethics of AI-generated media.
Platform Papers is published, curated and maintained by Joost Rietveld.