A Platform's Favor to Influencers Is a Hidden Tax on Advertisers

Based on the research ofZhu, Chen and Raghunathan, "Helping Oneself by Helping a Rival: Social-Media Platforms' Policies on Enabling Features for Influencers," Information Systems Research, 2026

That is the uncomfortable finding at the center of a new game-theoretic paper by Yan Zhu, Jianqing Chen, and Srinivasan Raghunathan, published in *Information Systems Research*. The question they formalize is one every platform executive already wrestles with informally: when an influencer's content on your platform works almost like a tiny, unregulated ad unit, do you help it work better, or do you cripple it? Instagram, TikTok, and YouTube have each answered differently at different times, and the paper's mechanism explains why those answers have nothing to do with being nice to creators and everything to do with advertiser pricing power.

The Swipe-Up Threshold Was Never About Technology

For four years, Instagram restricted the ability to attach a clickable external link to a Story, the "swipe-up", to accounts with roughly 10,000 followers or a verification badge. In August 2021 the company replaced swipe-up with the link sticker, and by that October it had opened the sticker to every account, regardless of size. Instagram's own explanation, relayed to Adweek, was that the change came from "feedback we heard from the community about how impactful it would be for accounts of all sizes to benefit from link sharing to engage with their followers." That is a user-experience justification. It is also, read through the Zhu-Chen-Raghunathan model, a textbook enabling decision: the platform had spent years gating a feature that made smaller creators' content more commercially effective, then chose to switch it on broadly once doing so no longer threatened its own position with advertisers.

The contrast case sits a few thousand miles away. In 2021, Xiaohongshu, the Shanghai-based lifestyle app, moved in the opposite direction, stripping out written links to products sold on rival marketplaces like Taobao and Tmall from its own feed. Caixin reported that insiders described the move as a "strategic adjustment, not a technical one," driven by Xiaohongshu's conclusion that outbound product links were generating commercial value the platform itself never captured. Different feature, different company, same logic in reverse: a platform will disable a monetizable hyperlink when it siphons value away from the platform's own commercial relationships, and enable one when it doesn't.

Both examples make the same point the paper makes formally: the enable/disable decision is not a creator-relations decision. It is a decision about who captures the value that a more effective influencer creates.

Why the Platform Doesn't Need to Charge the Influencer

The paper's central surprise is how little the platform has to ask for in return. In the baseline model, the platform cannot charge the influencer anything for using the enhanced feature, and it still comes out ahead, as long as advertisers do not value the platform's own advertising service dramatically more than they value the influencer's. Enabling the feature can cost the platform some market share, because a more effective influencer pulls some advertising spend away from the platform's own inventory. But the platform can claw that loss back, and then some, simply by raising the price it charges the advertisers who stay. The influencer gets a genuinely better product for free; the platform gets to charge more for what remains of its own service. Nobody at the platform has to negotiate a revenue share with any individual creator.

That is not how every platform has chosen to operationalize the idea once it becomes profitable enough to bother metering. TikTok Shop launched in late 2023 with no seller commission at all, by its own account, in explicit "growth mode." Within months, reporting from The Information, relayed by Tech.co, described a plan to lift the US referral fee from about 2 percent plus a flat per-transaction charge to as high as 8 percent on most items by the middle of 2024, alongside a pullback in merchant subsidies. TikTok built the audience first and only started billing once sellers were committed to the channel. The Zhu-Chen-Raghunathan model has a second regime for exactly this situation: when the platform *can* charge a fee for the feature, it should always enable the feature at some fee, because doing so strictly helps the platform, the extra fee revenue can arrive with no change to advertising revenue, a boost to it, or even at some advertising revenue's expense, depending on how the platform tunes the fee to steer how much service the influencer chooses to offer. Enabling for free and enabling for a fee are not two different policies in the paper's framework. They are the same lever, set at two different settings, and the fee-free setting is the one platforms reach for first because it requires no negotiation and produces no visible price increase for anyone to resent.

The Advertiser Pays for the Peace Treaty

This is where the paper's sharpest claim lives. A platform and an influencer compete for the same advertising budgets, that much is obvious, since both are, in the end, selling attention to the same marketers. So when a platform hands the influencer a feature that makes the influencer's service more effective, basic competitive intuition says the platform should lose. Instead, the paper shows that the platform's own price increase to advertisers can more than offset the lost market share, producing a genuine win for both the platform and the influencer. The authors call this what it looks like from the advertiser's chair: a "collusive" outcome, arrived at without any actual collusion, no backroom deal, no explicit coordination, just two formally competing parties whose joint payoff rises once one of them unilaterally decides to make the other stronger.

YouTube's own numbers hint at how this plays out at scale once a platform leans into it rather than away from it. In its April 2024 announcement of Shopping Collections and an in-app Affiliate Hub, YouTube disclosed that people watched more than 30 billion hours of shopping-related video on the platform in 2023, with a 25 percent increase in watch time for videos that help people shop. Every one of those hours makes YouTube's own advertising inventory around that content more valuable to the brands trying to reach shoppers, the same audience the affiliate-tagged influencer content is also competing for. The platform does not need a revenue share from the Affiliate Hub to benefit from its existence; a richer shopping ecosystem around creator video raises the price of being anywhere near it.

The platform doesn't need a cut of the influencer's commission. It needs a less price-sensitive advertiser standing next to a now-more-effective one.

What the Next Round of Feature Rollouts Signals

Instagram's own trajectory since 2023 reads like a live test of the model's two regimes. After years of leaning into Shop features, Meta killed Instagram's live-shopping tagging that March, with Meta telling TechCrunch it would "focus on products and features that provide the most value to our users" instead, a retreat from a feature whose returns, in a market where live commerce was never more than roughly 5 percent of US e-commerce, didn't clear the bar. Three years later, in April 2026, Instagram reversed course again, letting creators tag up to 30 affiliate products directly in Reels, with Meta's Nicola Mendelsohn declaring at Shoptalk that "the era of link in bio is finally over." The gating condition this time is not a follower count; it is that the tagged brand must sit inside Meta's own commerce catalog, a structural choke point that lets Meta decide, product by product, how much of this newly enabled value flows through inventory it can also price around.

None of these moves require Meta, TikTok, or YouTube to ever hand an influencer a line item labeled "platform fee for this feature." The paper's contribution is showing why that line item was never the point. A platform enabling a feature for an influencer looks, on its face, like generosity toward a competitor. It functions, in the aggregate, as a repricing event aimed at whoever is actually footing the bill, and that party has never been the influencer.