Stop Fencing In Your Creators. Their Second Platform Feeds Your First.

Based on the research ofXinzhi Rao, Jingyan Dai, Yulin Hao, and Liangfei Qiu, "Content Creator Multihoming and Attention Spillovers Across Platforms," SSRN Working Paper, 2026

Every incumbent platform treats multihoming as leakage. If a creator who built a following on your service starts posting the same clips on a competitor, the intuition is obvious and grim: split attention, divided loyalty, an audience being taught the exit route. So platforms build fences. They watermark downloads, suppress rival-branded uploads in the ranking, and dangle funds and exclusivity to keep talent from wandering. The logic feels airtight. It is also, according to a new working paper, backwards.

In "Content Creator Multihoming and Attention Spillovers Across Platforms," Xinzhi Rao, Jingyan Dai, Yulin Hao, and Liangfei Qiu track thousands of creators across a seven-year panel spanning a major Chinese short-video platform and the lifestyle network RedNote, watching what happens to a creator's performance on the original platform after they adopt a second one. The result contradicts the leakage story cleanly. Following adoption of a second platform, creators saw a 23.5% increase in followers and a 25.4% increase in engagement on the platform they started on. The rival did not drain the pool. It refilled it.

The Spillover the Incumbent Was Told to Fear

Read the mechanism carefully, because the direction of causation is the whole story. The authors do not find that already-surging creators happen to expand; they find that expansion itself generates growth back home. Their mechanism analyses attribute the gains not to cannibalization but to audience expansion through cross-platform attention spillovers: exposure on the new platform sends fresh viewership to the creator's existing account. A viewer who discovers a chef on RedNote goes looking for their back catalogue on the short-video app. The second platform functions less like a competing store down the street and more like a billboard the creator rents in a neighborhood the incumbent never reached.

The behavior around adoption sharpens the picture. Creators in the panel tend to add a second platform after their growth on the incumbent stalls, and they cross-post selected content rather than dumping everything indiscriminately. They also raise their overall content production and shift topics toward the styles characteristic of the new platform. In other words, multihoming is not a passive hedge against decline; it is an active growth strategy that a creator reaches for precisely when the home platform has stopped delivering, and it works well enough to reverse the stall on the very platform they were supposedly abandoning.

The counterintuitive part is that the incumbent captures this upside for free. It spent nothing to acquire the new audience, ran no campaign, bought no billboard. A creator, chasing their own growth, imported an audience from a rival's user base and deposited a chunk of it back on the incumbent's ledger. The platform's only job was to not get in the way, which is exactly the job most platforms refuse to do.

Why Platforms Wage War on the Very Thing That Feeds Them

The defensive playbook is not hypothetical, and it is expensive. In August 2022, YouTube began adding watermarks to Shorts specifically, as TechCrunch and The Verge reported at the time, to discourage creators from downloading their clips and reposting them to TikTok and Instagram Reels. TikTok has long stamped downloaded videos with a bouncing, hard-to-crop watermark carrying the creator's handle, engineered to pull viewers back. Instagram has stated its ranking will suppress content bearing another app's watermark. The entire apparatus exists to raise the cost of the exact behavior Rao and colleagues find is accretive to the home platform.

This is the friction that the anchor paper reframes as self-defeating. Every watermark that makes a creator's cross-posted clip look worse on a rival, every ranking penalty that buries competitor-branded uploads, is a tax the platform levies on its own future spillover. The incumbent is spending engineering effort to suppress the billboard that would have sent viewers back to it. The war on multihoming is a war on inbound attention, fought by the party that stands to receive it.

The watermark is a tax the platform levies on its own future spillover. Every fence that makes a creator harder to follow off-platform makes them harder to discover on it.

None of this means friction is irrational for every platform in every position. The dominant network with genuine winner-take-all economics can sometimes benefit from raising switching costs. But the paper's setting is the ordinary case: platforms competing for the same finite pool of creators and attention, where the marginal creator is going to multihome no matter what the fence looks like. In that world, and it is most of the world, the fence forfeits the spillover without stopping the behavior.

Creators Were Never Going to Stay Home Anyway

The premise underneath the defensive posture, that a creator could be persuaded to live on one platform, was already dead. The evidence that multihoming is the creator economy's default state predates the anchor paper. In a large-scale study of the creator economy published in Social Media + Society in 2023, Bernhard Rieder, Erik Borra, Òscar Coromina, and Ariadna Matamoros-Fernández examined a near-complete sample of 153,000 "elite" YouTube channels, each with at least 100,000 subscribers, and mined the URLs embedded in 137 million video descriptions. They found that the "big three" off-YouTube destinations, Facebook, Instagram, and Twitter, dominate the linking across every content category, functioning as what the authors call unavoidable promotion tools for building an audience.

That phrase should reset the strategic frame. Multihoming is not a threat a platform can choose to face or avoid; it is the ambient condition of professional content creation. Elite creators do not treat any single platform as home. They treat the whole constellation as their distribution, and they route attention among the nodes deliberately. A platform that designs as though it can wall its creators in is not defending a realistic position; it is refusing to compete for the spillovers that the Rao paper shows are on the table.

Once you accept that creators multihome regardless, the incumbent's decision variable narrows to something concrete: given that your best creators will post on a rival, do you architect your platform to receive the audience they bring back, or to repel it? The watermark answers "repel." The paper says that answer costs you.

The Fund Was Always a Spillover Machine

Here is the tell that platforms already half-understand the spillover, even while they fight it: they pay to manufacture the same effect from the other direction. In May 2021, YouTube launched a $100 million Shorts Fund to reward top short-form creators over 2021 and 2022. TikTok had preceded it, announcing a $200 million U.S. Creator Fund in July 2020 that it said it expected to grow past $1 billion in the U.S. within three years. When TikTok retired that fund at the end of 2023, it did not abandon the model; it replaced it with the Creativity Program, which offered qualifying creators substantially higher payouts for longer original videos.

Strip away the branding and a creator fund is a subsidy to attract multihoming creators and the audiences that trail them. YouTube was not paying Shorts creators out of charity; it was paying to pull TikTok-native talent, and TikTok-native audiences, onto Shorts. That is inbound spillover, purchased at nine figures. The incumbent that watermarks its outbound clips while writing checks to import a rival's creators is running both strategies at once, paying dearly for the attention it flows in and taxing the attention it could flow out.

The cleaner move is to treat facilitation as the whole strategy. YouTube has, in fact, drifted toward it: it has offered creators free access to third-party tools that automatically repackage their TikTok and Instagram Reels videos into Shorts, greasing the exact cross-posting its watermarks were built to discourage. That is the paper's logic operationalized. Lower the cost of multihoming into your platform, make the imported audience easy to follow and monetize, and the spillover that a creator's rival presence generates accrues to you instead of leaking past.

What To Build Instead of a Fence

The forward-looking reading of Rao and colleagues is not "abolish exclusivity everywhere." It is that the default posture toward multihoming should flip from suppression to capture. The near-term product agenda writes itself. Make cross-posting into your platform frictionless, not penalized, so a creator's rival presence routes viewers to a followable home base. Instrument the spillover so you can see which creators are importing audiences and reward them for it, rather than ranking them down for a competitor's watermark. Accommodate, in your content-mix and recommendation systems, the reality that multihoming creators will shift their style toward whatever platform they just joined, because the paper shows they do, and fighting that adaptation only degrades the content your own audience sees.

The deeper shift is conceptual. Platforms have modeled creators as a stock to be locked in and attention as a fixed pie to be divided. The multihoming evidence says both models are wrong. Attention is not conserved across platforms; cross-platform exposure creates it, and a meaningful share of what a creator's second platform generates flows back to the first. The 23.5% and 25.4% figures are the size of the prize an incumbent forfeits every time it treats a creator's ambition as betrayal. The platforms still building higher fences are optimizing against their own growth, and the receipts, this time, are in the panel data.