📈 economics

Platform Competition with User-Generated Content

This paper presents a theoretical model of platform competition where user-generated content quality is endogenously determined by user composition, analyzing how platforms strategically set advertising intensity to influence user utility and shape market equilibria under both Nash and Stackelberg frameworks.

Original authors: Bohan Zhang

Published 2026-05-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Bohan Zhang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the internet as a giant, bustling town square where two rival marketplaces are trying to get people to hang out. These marketplaces are like social media apps (think YouTube or TikTok), and the "goods" they sell aren't physical items, but content created by the people themselves (User-Generated Content, or UGC).

Here is the story of how these two marketplaces compete, told through the lens of this paper.

1. The Two Ingredients of a Good Party

In this town, everyone wants to go to the party that offers the best mix of two things:

The Quality of the Music/Entertainment: Is the content interesting, smart, and high-quality?
The Size of the Crowd: Is there a huge party with lots of people, or is it a small, empty room?

However, there's a catch. The "music" isn't played by a DJ hired by the party host. The guests are the ones playing the music.

Some guests are "High-End Artists" who make beautiful, complex music but hate loud, annoying noises.
Other guests are "Casual Party-Goers" who just want a huge crowd and don't care if the music is a bit rough.

2. The Host's Dilemma: The "Ad" Noise

The marketplaces (the hosts) make money by selling "advertising space." In our story, advertising is like loud, flashing neon signs and annoying loudspeakers set up in the middle of the party.

The Good: More ads mean more money for the host.
The Bad: The loud ads annoy the "High-End Artists." If the ads get too loud, these artists pack their instruments and leave.
The Domino Effect: When the artists leave, the music gets worse. When the music gets worse, the "Quality-Sensitive" guests leave. Eventually, the party becomes a noisy, low-quality mess, and even the casual crowd gets bored and leaves.

3. The Big Discovery: The "Winner-Takes-All" Trap

The paper asks: What happens if two marketplaces compete to see who can get the most people?

The authors found a surprising and somewhat scary result: It is almost impossible for both to survive.

Here is why:

If Marketplace A tries to make money by turning up the "ad volume," the High-End Artists leave.
Marketplace B sees this and thinks, "Hey, if I keep my ads quiet, I can steal all the High-End Artists!"
Marketplace A panics and turns its ads down even further to steal them back.
This turns into a race to the bottom. One marketplace eventually realizes that to win, it must drive the other one out of business entirely.

The Result: One platform ends up dominating the whole market, while the other gets zero users and zero profit. The paper calls this a "tipping" effect. Even if the second platform tries to enter with a "better" product, the first platform can simply adjust its ad strategy to make the second platform's entry impossible.

4. The "First Mover" Advantage

The paper also looks at what happens if one platform gets to set its rules before the other one even opens its doors (like a leader and a follower).

The leader sets a specific, moderate level of "ad noise."
This level is just low enough to keep the High-End Artists happy, but high enough to make money.
The follower tries to enter, but realizes that no matter what they do, they can't beat the leader's setup. The leader effectively blocks them out.
The Lesson: Being the first to build a loyal, high-quality community creates a "fortress" that is very hard for competitors to break into.

5. Real-World Example Used in the Paper

The authors use a specific example to explain this: Academic Forums.

There is a famous, somewhat toxic forum called EJMR (Economics Job Market Rumors). It has a lot of "noise" (toxic content/ads), but it has a massive user base.
New, cleaner forums tried to start up to offer a "high-quality, professional" environment.
According to the paper's logic, these new forums failed not because they were bad, but because the "noise" on the old forum wasn't quite loud enough to drive everyone away. The old forum had already secured the "High-End Artists" (the active users), so the new, quiet forums couldn't get enough traction to survive.

Summary of the Takeaways

Quality is fragile: In a world where users create the content, the "quality" of the platform depends entirely on who is there. If you annoy the best contributors with too many ads, the whole platform collapses.
Competition is brutal: In this specific type of market, competition doesn't usually lead to two healthy companies coexisting. It often leads to one giant winner and one loser.
Strategy matters: Platform owners have to be very careful. If they push for too much short-term profit (too many ads), they might kill the very thing that makes their platform valuable: the high-quality users.

The paper concludes that while this model explains why some digital markets become monopolies, it also suggests that regulators should watch out for how platforms use advertising to push competitors out, rather than just looking at prices.

Technical Summary: Platform Competition with User-Generated Content

Problem Statement
This paper addresses a gap in the economic literature regarding platform competition where content quality is not an exogenous variable or a direct strategic choice of the firm, but rather an endogenous outcome determined by the composition of the user base. While traditional models emphasize network effects derived from user base size (quantity), they often overlook how the quality of user-generated content (UGC) interacts with advertising intensity. The central problem is to understand how platforms compete when their strategic variable (advertising intensity) simultaneously affects user utility and the endogenous quality of the content ecosystem, potentially leading to market tipping or coexistence.

Methodology
The authors develop a theoretical model involving two competing platforms ( $j=1, 2$ ) and a continuum of users with measure 1.

User Heterogeneity: Users differ in their relative valuation of network size versus content quality, denoted by a parameter $\beta_i \sim U[0, 1]$ . A higher $\beta_i$ indicates a stronger preference for network size relative to quality.
Endogenous Quality: Content quality is modeled as a function of user type. Specifically, the quality contribution of a user with parameter $\beta$ is $q(\beta) = q_m - \lambda\beta$ , implying that users with a higher relative valuation of network size (higher $\beta$ ) tend to contribute lower-quality content.
Utility Function: A user's utility from joining platform $j$ is given by $U_{ij} = Q_j + \beta_i n_j - a_j$ , where $Q_j$ is the average content quality on the platform, $n_j$ is the market share, and $a_j$ is the advertising intensity. Advertising reduces utility by lowering perceived quality.
Equilibrium Concept: The paper utilizes a generalized fulfilled-expectations equilibrium. An equilibrium allocation is defined where the realized user distribution matches the users' expectations, which in turn determine the average quality $Q_j$ .
Solution Approach: The authors first characterize the "realized allocation" for fixed advertising levels, proving the existence of a unique cutoff $\tilde{\beta}$ that separates users between platforms. They then define a mapping $\Gamma$ from expected market shares to realized market shares. Finally, they analyze the profit maximization problem for platforms under two scenarios: simultaneous moves (Nash equilibrium) and sequential moves (Stackelberg equilibrium). Stability refinements are applied to select among multiple equilibria.

Key Results

Existence of Multiple Equilibria: The model demonstrates that for certain parameter ranges (specifically when the quality sensitivity parameter $\lambda < 1/4$ ), multiple equilibria can exist where either platform dominates or where one platform holds a specific market share.
Stability and Refinement: Using a dynamic stability criterion (iterative adjustment of expectations), the authors show that interior equilibria with multiple solutions are unstable. Only the equilibrium with the smaller market share (for the interior case) or the boundary equilibria (where one platform captures the entire market) are stable.
Nash Equilibrium Outcome: In the simultaneous game, the unique stable Nash equilibrium results in a "winner-takes-all" scenario. The dominant platform sets an advertising level of $a^* = 1/8 - \lambda/2$ , while the competing platform is forced to set $a^* = 0$ and earns zero profit. The dominant platform effectively blockades entry.
Stackelberg Equilibrium: In the sequential game where one platform moves first, the leader can commit to the same advertising level ( $1/8 - \lambda/2$ ) that forces the follower out of the market. The result is identical to the Nash outcome, suggesting that the potential for the dominant strategy in a simultaneous game is sufficient to deter entry without the need for sequential commitment.
Strategic Trade-off: The analysis reveals a negative relationship between the sensitivity of content quality to user composition ( $\lambda$ ) and the optimal advertising intensity. As $\lambda$ increases (meaning user composition has a stronger impact on quality), the optimal advertising level for the dominant platform decreases to preserve the high-quality user segment.

Contributions
The paper makes three primary contributions to the literature:

Endogenous Quality Modeling: It extends platform competition literature by explicitly modeling the feedback loop between advertising, user base composition, and content quality, moving beyond the assumption that quality is fixed or directly controlled by the platform.
Equilibrium Characterization: It characterizes the properties of equilibria in this setting and introduces a refinement process based on stability to select among multiple plausible outcomes, resolving ambiguity in market predictions.
Market Concentration Insight: It demonstrates that platform competition with UGC naturally leads to market dominance by a single platform. The model explains why smaller platforms offering higher quality may fail to coexist if the dominant platform maintains a sufficiently low advertising intensity.

Significance and Claims
The authors claim that their model provides both theoretical advances and practical guidance. Theoretically, it integrates network effects, quality competition, and UGC mechanisms into a unified framework, explaining why competition can collapse in simultaneous play and how sequential commitment generates first-mover dominance. Practically, the results offer insights for platform operators and policymakers:

For Managers: Aggressive monetization can backfire if it drives away quality-sensitive users, triggering a negative feedback loop. Platforms must balance revenue with the preservation of their high-quality user base. First-mover advantages can be decisive and self-reinforcing.
For Policymakers: The model highlights risks of market concentration driven by strategic advertising choices rather than just pricing. It suggests that regulators should consider how platforms shape user composition and content quality, potentially supporting policies like interoperability or limits on excessive advertising to maintain a competitive environment.

The paper concludes by noting limitations, such as the assumption of symmetric costs and purely advertising-based revenue, and suggests future research directions involving dynamic user behavior, platform investments in moderation, and alternative monetization models like subscriptions.