Frontiers: Content Quality Provision and Welfare Implications of Digital Platform Commission Fees

1. Introduction

Over the past few years, a major conflict has emerged between platforms such as Apple’s App Store and Google Play Store and the developers of leading digital services. Prominent firms such as Spotify and Epic Games argue that the high commission on in-app revenues stifles innovation (Allyn 2020, Hoppe 2020, Geradin 2021).¹ This dispute has attracted growing scrutiny from regulators in the United States, European Union, and Asia, who question whether prevailing fee structures promote or hinder technological progress (Warren 2021, Jimenez 2024). The issue has also entered the public spotlight with media coverage such as John Oliver’s HBO segment claiming that high commissions suppress developer innovation (Kelly 2022).

At first glance, it seems intuitive that lower commission fees should spur innovation: reduced commissions raise developers’ marginal returns, encouraging investment in higher quality content (a direct effect). However, platforms also benefit from high-quality content as it boosts engagement and usage, thereby increasing their own revenues. To enhance content quality, platforms can expand the demand for content, thus increasing the marginal returns of developers to investing in quality. Understanding how platforms shape this demand is therefore crucial.

A defining feature of many platforms is that they are hardware-gated ecosystems: consumers must purchase a proprietary device to access content, for instance, smartphones for mobile apps, video game consoles (Sony PlayStation, Microsoft Xbox), e-readers (Amazon Kindle, Barnes & Noble Nook), or smart-home systems (Amazon Alexa, Google Home).² Because the demand for content depends on the installed base of compatible devices, platforms often price devices aggressively to expand both hardware adoption and content consumption. A reduction in commission rates weakens this incentive by lowering the platforms’ revenue from content, leading to higher device prices and a smaller installed base. This contraction reduces the developer’s effective market size and, in turn, the incentive to invest in quality (an indirect effect). The direct and indirect effects, thus, operate in opposite directions, making the overall impact of commission changes on content quality fundamentally ambiguous.

This mechanism reflects real-world practice. Revenues from digital content distribution now make up a substantial share of the earnings of many platforms. For example, Apple’s services segment—including the App Store—generated nearly $96 billion in 2024, contributing significantly to the company’s gross profits.³ Such figures create strong incentives for Apple to factor service-side revenues into device pricing. Cross-subsidization across markets is also well-established, for instance, Google’s Fiber initiative was not aimed at direct profits but at accelerating adoption in complementary markets such as smart homes and video streaming (Levin and Downes 2018). These cases highlight the plausibility of platforms strategically subsidizing hardware to expand demand for complementary service-side revenue streams.

These observations lead to our main research questions. We seek to investigate how the commission fees charged by hardware-gated platforms may impact digital content quality innovation as well as the profitability of developers. Additionally, we aim to assess how the commission rate influences consumer welfare. By exploring these aspects, we provide a comprehensive understanding of the economic implications of the commission rates charged by platforms for different stakeholders in the ecosystem.

We develop a stylized model of a market with two competing platforms and a digital content developer who invests in content quality and distributes it on both platforms. Each platform sells a proprietary device that consumers must purchase to access the content and also collects commissions on the developer’s revenue. Differences in design and user experience cause consumers to engage more on one platform, creating an asymmetry in which that platform yields higher content revenue for the developer.

A reduction in the commission rate lowers the importance of commission revenues in platform profits and leads platforms to raise device prices. Because the platform with higher consumer engagement depends more on commissions, it has a stronger incentive to raise prices, and this reduces its market share. This shift dampens developer revenues and weakens quality incentives. When the indirect loss in incentives outweighs the direct benefits of lower commissions, the developer invests less in quality. The developer may still earn more overall because of the lower commission burden, but consumers face higher device prices and possibly lower content quality. Thus, a commission cut can reduce quality provision or consumer welfare, improving the developer’s payoff, showing that developer incentives may diverge from consumer interests.

Our work relates to research on platform strategies that shape innovation by third-party suppliers. Previous studies show how platform actions can influence developer incentives; for example, the launch of Google’s first-party apps reduces the innovation of existing third-party apps on Google Play (Foerderer et al. 2018, Wen and Zhu 2019); up-front fixed access fees make failed innovation more costly, encouraging sellers to exert more effort (Lin et al. 2011); and rating systems affect incentives for high-quality apps on the Apple App Store (Leyden 2021). Angelini et al. (2025) further demonstrate how demand-enhancing platform actions can asymmetrically affect sellers. Based on this stream, our article shows how the commission of a platform interacts with the pricing of devices in a hardware-gated ecosystem to influence developers’ innovation incentives. Unlike much of the prior literature, which examines a single platform with many sellers, our setting features two competing platforms and a single developer, highlighting how cross-platform competition and device cross-subsidization shape quality investment incentives.

Our paper also contributes to the literature on pricing in platforms (Rochet and Tirole 2003, Parker and Van Alstyne 2005). Although much of this work emphasizes direct subsidies to one side of the market (typically consumers) to stimulate growth, we study an indirect subsidy: device pricing. Because consumers must purchase the platform’s device to access content, device subsidies directly affect demand for content and, therefore, alter developers’ incentives to invest in quality. Finally, our work connects to the literature on vertical relationships, which examines how upstream–downstream arrangements shape innovation. Wang and Shin (2015) show how contract form and downstream competition influence upstream innovation, whereas Guo (2020) analyze how upstream suppliers may exploit downstream firms through contractual terms. In contrast, we study hardware-gated ecosystems in which the interaction of commission rates and device pricing may serve as the upstream lever shaping a developer’s innovation incentives.

In the next section, we present our model, and we present our findings in Section 3. We discuss our concluding thoughts in Section 4. All proofs are presented in the appendix, and all model extensions are contained in the Online Appendix.

2. Model

We consider a setting with two competing digital platforms, labeled a and b. Each platform owns and sells a proprietary hardware device—denoted A and B—that provides users access to its respective digital ecosystem. A representative third-party developer D creates digital content (e.g., an app, game, or service) that can be accessed on either platform, and users must purchase one of the hardware devices to do so. We focus on a setting in which the developer possesses market power, with monopoly as the extreme case, in order to isolate how platform commissions and device pricing shape the developer’s incentives to invest in quality.⁴

2.1. Developer and Platforms

The developer, D, develops content of quality q by investing $k{q}^2$, where $k>0$.⁵ We assume that a consumer spends q on content purchases if the consumer uses the content version from platform a and $\alpha q$ if the consumer uses the content version from platform b, where q is the quality of the content and $\alpha >0$. Without loss of generality, the setup captures the fact that higher content quality results in higher content purchases on both platforms. Note that $\alpha =1$ implies that a consumer on both platforms generates an equal amount of revenue per unit quality for the developer, whereas $\alpha \ne 1$ implies asymmetric returns to quality for the developer from the two platforms.

In practice, developers often earn different average revenues per user across platforms. For instance, Howarth (2025) reports that, in the mobile app ecosystem (iOS versus Android), users not only spend different amounts of time engaging with apps but also differ systematically in income levels. These differences in engagement and user demographics together help explain why app developers may experience heterogeneous per-user earnings across platforms. The difference in engagement could be due to a better user experience (a supply-side factor) on one of the platforms. In the mobile app market, the open nature of Google’s ecosystem—in which numerous manufacturers deploy customized versions of Android—results in significant fragmentation in both hardware and operating system (OS) versions. In contrast, iOS operates in a more controlled environment with a limited number of device models and much greater consistency in OS adoption. As a result, iOS apps need to be optimized for a narrower range of hardware and software configurations, enabling a more streamlined and consistent user experience (Brovkin 2017, DDI Development 2018). As also reported by Gyorödi et al. (2017), iOS outperforms Android in two of the three core performance tests, specifically asynchronous JavaScript object notation parsing and user interface rendering speed, both of which are critical for the smooth functioning of modern apps.⁶ Consequently, we model a setting in which one of the platforms provides a better user experience. To do this and without loss of generality, we assume that $\alpha >1$, which implies that the content earnings per user from platform b are higher compared with platform a. In one of the model extensions, we also consider the alternative scenario in which the two platforms are asymmetric in the utility they provide, and the reason for the differential per-user revenue across platforms is the demand-side factor of heterogeneity in consumers’ willingness to pay for quality. See Section 3.1 for more details.

We assume that both platforms charge a commission rate of r on the developer’s revenue from content purchases.⁷ This assumption is motivated by the following observations. First, competing platforms across industries consistently charge similar commission fees, typically around 30% (Perez 2020, Borck et al. 2021), despite having vastly different business models. For example, both Apple’s App Store and the Google Play Store charge 30% on in-app purchases and subscription revenue.⁸ Likewise, Xbox and PlayStation maintain 30% commission rates despite offering different subscription and cloud gaming models (LeBlanc 2022, Wolfe 2025). Second, this 30% rate has remained remarkably stable over time; even as the app ecosystem has expanded from hundreds to millions of apps, smartphone adoption in the United States has grown from 35% in 2011 to more than 90% in 2024 (Pew Research Center 2024), and monetization has shifted from paid downloads to subscriptions and in-app purchases.⁹ The substantial variation across market structures, platform monetization models, business strategies, and time periods makes it unlikely that the commission rates arise from a competitive, Bertrand-style equilibrium. Relatedly, Apple has also said that it followed the prevalent industry standard to choose its commission rates (Miller 2021).

2.2. Consumers

We capture the device market through a Hotelling line of unit length, allowing for a differentiated market because of a variation in consumers’ brand preferences for different devices. We assume A is located at the left end of the line (at zero), and B is located at the right end (at one).¹⁰ The prices of the devices are set at $p_A$ and $p_B$, respectively. A unit mass of consumers are uniformly distributed over the line and estimate their utility from device A to be $v_a$ and from device B to be $v_b$, respectively. We assume here that the utility the consumers get from devices is independent of the content quality.¹¹ A consumer located at distance x from device A gets a net surplus of $v_a-x-p_A$ from buying A and $v_b-(1-x)-p_B$ from buying B. We assume that $v_a$ and $v_b$ are large enough so that the market is covered.

A consumer who is indifferent between devices A and B is denoted by $x_{AB}$. Based on the location of the marginal consumer, we can derive the demand function for each device. Let $m_i$ denote the demand for device $i\in \{A,B\}$. We have $m_A=x_{AB}=(1-p_A+p_B+v_a-v_b)/2$ and $m_B=1-x_{AB}=(1+p_A-p_B-v_a+v_b)/2$. Then, the profit functions of a, b, and D are given by the following expressions: ${\pi }_a=(p_A+rq)m_A$, ${\pi }_b=(p_B+r\alpha q)m_B$, and ${\pi }_D=(1-r)q(m_A+\alpha m_B)-k{q}^2$, respectively.

3. Analysis

In this section, we present the results of our analysis under a subset of parameter conditions described in the model, and this allows us analytical tractability, preserving insights. In particular, we assume that $v_a=v_b=v>0$.¹² The timeline of the game is as follows. Based on the commission rate r, the content developer D first decides the level of quality (innovation) because the core development of the content is a more long-term decision than setting device prices.¹³ Then, the two platforms set their respective device prices.

Using the timeline, we use backward induction to solve the game. First, we obtain optimal prices of the devices as functions of the content quality by solving the first order conditions for the profit functions of A and B. These prices are

Accordingly, we represent the market shares of the devices and the profits of the developer as a function of quality by $m_i^q$ for $i\in \{A,B\}$ and ${\pi }_D^q$, respectively. From the price expressions, one can see that, as $\alpha >1$, $\frac{d{p}_B(q)}{dr}<\frac{d{p}_A(q)}{dr}<0$, which, in turn, implies that the market share of B increases with r as that of A decreases ($\frac{d{m}_B^q}{dr}>0>\frac{d{m}_A^q}{dr}$).

Using the above price expressions, we examine the impact of content quality on prices. We present this relationship in the following lemma.

Lemma 1.

Both device prices decrease with content quality.

$$\frac{d{p}_B(q)}{dq}<\frac{d{p}_A(q)}{dq}<0.$$

The first observation from the above lemma is that the prices of both devices decrease with the content’s quality. This price reduction happens because a higher content quality increases consumers’ spending on the content. Consequently, the revenues per consumer through commission fees from the content purchases for the two platforms also increase. Therefore, both of these platforms have an incentive to reduce the price of their devices to gain market share to capitalize on the increased revenue per consumer from commissions.

The second observation from Lemma 1 is that the rate of reduction in the price of B is higher than that of A. The reason is that a unit increase in the content’s quality improves the revenue per consumer for platform b more than that for platform a (because $\alpha >1$). Hence, platform b has a stronger incentive than a to reduce the price of its device to gain market share.

An implication of Lemma 1 is that $\frac{d{m}_B^q}{dq}>0>\frac{d{m}_A^q}{dq}$.

Lemma 2.

Device prices decrease with content quality at a higher rate when the commission rate increases.

$$\frac{d^2{p}_B(q)}{\mathit{\text{dqdr}}}<\frac{d^2{p}_A(q)}{\mathit{\text{dqdr}}}<0.$$

Lemma 2 says that the rate of price reduction of the devices with content quality increases with the commission rate, r. Further, this rate of reduction is higher for B. The reason for these observations is that, as r increases, the marginal value of increasing the market size increases for both A and B because of higher earnings from commission revenue. Further, because B earns higher revenues than A per consumer (because $\alpha >1$), this incentive of B is greater than that of A.

Continuing our analysis, we substitute the first stage prices $p_A(q)$ and $p_B(q)$ into the profit function of developer D and obtain the equilibrium content quality by solving the first order condition with respect to quality.¹⁴ We then substitute the equilibrium quality expression into the first stage prices to get the equilibrium prices. We present the expressions of equilibrium content quality and phone prices below:¹⁵

the corresponding phone market shares in equilibrium are represented by $m_i^{*}$ for $i\in \{A,B\}$, and the profits of the developer are denoted by ${\pi }_D^{*}$. From the price expressions, one can see that $\frac{d{p}_B^{*}}{dr}<\frac{d{p}_A^{*}}{dr}<0$ for $r<\frac{1}{2}$, which, in turn, implies $\frac{d{m}_B^{*}}{dr}>0>\frac{d{m}_A^{*}}{dr}$ for $r<\frac{1}{2}$. Having obtained the equilibrium content quality, we now analyze how it is affected by the commission rate charged by the platforms.

Proposition 1.

Content quality ($q^{*}$) decreases as r decreases when $r<\overline{r}$.

Proposition 1 provides the insight that a decrease in the commission rate may create incentives for the content developer to decrease quality.¹⁶ The intuition behind this result can be explained by two opposing forces created by the reduction in commission rate on the content developer. To explain these two forces, we write the marginal value of quality,

where we use the fact that $m_A^q=1-m_B^q$ because the market is fully covered. Equating the previous expression to zero, we get the expression for equilibrium quality as a function of the market size of B as where $M_B=\frac{d{m}_B^q}{dq}$. From Lemma 1, we note that $(\alpha M_B+1-M_B)>0$ because an increase in content quality results in the rate of price drop of B being higher, resulting in an increase of market share of B (i.e., $M_B>0$). We conclude that a decrease in r increases the numerator and creates a force to decrease the denominator in the expression for quality. Both these effects create an incentive to increase the content’s equilibrium quality and capture the direct effect of a decrease in commission rate, r. However, there is an indirect effect of r as well. Because the rate in price reduction of B is higher with r (Lemma 2), we have $\frac{d{M}_B}{dr}>0$. Therefore, as r decreases, the expression $(\alpha M_B+1-M_B)$ also decreases and creates a force to increase the denominator in the expression of quality. Thus, the indirect effect of a reduction in the commission rate, r, is to decrease the content quality. When r is relatively low, the term $(1-r)$ in the denominator of the quality expression is relatively larger, thus accentuating the negative indirect effect of a decrease in r. In sum, a decrease in commission rate incentivizes platform b to increase its price at a higher rate, reducing its market share. As the more profitable platform loses market share, the developer has a weaker incentive to invest in quality.

Next, we turn to the effect of the commission rate on the profits of the content developer.

Lemma 3.

The content developer’s profit (${\pi }_D^{*}$) decreases as r decreases when $r<\widehat{r}$.

Intuitively, a decrease in the commission rate r should increase the content developer’s profits because it retains a larger share of the content revenues.¹⁷ On the contrary, Lemma 3 says that the content developer’s profits decrease with a reduction in r when r is relatively small. The reason becomes clear from applying the envelope theorem to the content developer’s profit function ${\pi }_D^q$ with respect to r. We find that the direct effect of a decrease of r on the developers’ profits is indeed positive because

However, the indirect effect of a decrease in r on the developer’s profit, because of a decrease in the market share of B, is negative because as $\frac{d{m}_B^{*}}{dr}>0$ for $r<\frac{1}{2}$. When the indirect effect is strong enough to overcome the direct effect, we have a situation in which the content developer’s profits decrease with a reduction in r.

Combining the results from Proposition 1 and Lemma 3, we have the following.

Proposition 2.

When $r\in (\widehat{r},\overline{r})$, a decrease in r decreases the content quality but increases the content developer’s profits, and when $r<\mathit{\text{Min}}\{\widehat{r},\overline{r}\}$, a decrease in r decreases both the content quality and the content developer’s profits.

The above finding reveals a scenario in which the incentives of the content developer are not aligned with an improvement in quality provision.¹⁸ We note that, as the commission rate reduces, the content quality also reduces (for $r<\overline{r}$). However, the reduction of the commission rate on the content developer’s profits can be both positive and negative. This result is because of two opposing effects of a decrease in commission rate on the content developer’s profits. On the one hand, a reduction in commission rate improves the profits of the content developer as it keeps a larger fraction of the content revenue. On the other hand, however, as the commission rate decreases, both platforms’ revenue from the content decreases, and it creates an incentive to increase their device prices. As B’s price increase is greater (when $r<\overline{r}$), the relative decrease in B’s market share hurts the content developer’s marginal value of quality (because $\alpha >1$). Therefore, the content developer creates a lower quality that increases its profits because of a reduction in convex quality creation costs. We find that the positive effect of a decrease in the commission rate on the developer’s profits dominates when $\widehat{r}<r<\overline{r}$. Hence, it is possible that content developers complain about the commission rate being too high even when a reduction in the commission rate leads to a lower content quality. When $r<\mathit{\text{Min}}\{\widehat{r},\overline{r}\}$, the negative effect of a decrease in the commission rate dominates, and a reduction in r reduces the content developer’s profits. This implies that, in this parameter range, lowering the commission rate may not only hurt content quality but also the content developer’s profits.

Last, we look at the impact of decreasing the commission rate on consumer welfare.

Proposition 3.

As r decreases, the content quality increases, but consumer surplus reduces when $r\in (\overline{r},1/2)$.

In our model, a reduction in the commission rate affects consumer surplus through its influence on device prices.¹⁹ When content quality increases with a reduction in r (when $r>\overline{r}$), the platforms’ content revenue can decrease because it retains a smaller fraction of content revenues. This feature may, therefore, lead the platforms to increase their device prices (when $r<1/2$), thus hurting consumer welfare. The implication of this result is that consumer welfare does not move in lockstep with the content quality and may even decrease when the content quality increases.

Beyond consumer surplus, we examine the broader implications for social welfare by incorporating the platforms’ and the developer’s profits. Our analysis reveals that, in certain parameter ranges (e.g., at $k=0.25$, $\alpha =2$, and $r=0.1$), a reduction in the commission rate leads to a decline in both consumer surplus and overall social welfare. This occurs when the lower commission weakens the incentives of platform b more to lower its device price, shrinking its market, and reducing content quality. We find that, whereas the developer’s profits may rise because of improved margins, the sum of the profits of the two platforms declines. The decline in consumer surplus and platforms’ profits may dominate the increase in the developer’s profits, resulting in lower social welfare. These results underscore that lower commission rates, despite potentially benefiting developers individually, can be welfare-reducing when the indirect effects on quality are considered.

3.1. Robustness Checks

In Online Section 1.1, we analyze a model in which platforms are vertically differentiated and consumers differ in their willingness to pay for device and content quality. In this setting, a reduction in commission rates leads platforms to raise device prices, and this can contract overall market coverage and indirectly reduce the developer’s incentive to invest in quality. However, if the lower bound of willingness to pay is sufficiently high so that full coverage is maintained even after device prices increase, this extensive margin channel disappears, and the main findings can no longer hold. The unifying theme across both models is that the developer’s quality incentives weaken when the effective market shrinks through intensive margin reallocations in the main model, in which users shift toward the lower revenue platform, or through extensive margin contractions in the vertical differentiation model, in which higher device prices reduce market coverage.

Beyond the vertical differentiation case, we also analyze several other extensions to assess the robustness of our main insights (see details in the Online Appendix). These include allowing the developer to offer asymmetric quality across platforms, linking device utility directly to content quality, relaxing the assumption of identical commission rates, and comparing open versus closed platform systems. Across these variations, the qualitative mechanism remains consistent: reductions in commission rates can raise device prices and indirectly dampen developers’ incentives to invest in quality by contracting their effective market size. Taken together, these analyses demonstrate that our central finding—that lower commissions need not always increase innovation or consumer welfare—holds under a broad range of modeling assumptions and continues to capture the key dynamics of hardware-gated platform ecosystems.

4. Discussion

In this section, we discuss the role of various parameters in obtaining our results. We note that the parameter $\alpha$ influences the thresholds presented in Propositions 1 and 2 ($\overline{r}$ and $\widehat{r}$). We capture the interplay between r and $\alpha$ through a visual summary of our main results in Figure 1 (for $k=0.1$). We observe that, in the leftmost region, a decrease in r leads to higher quality content, and the content developer’s profits also increase. However, for a given commission rate, as $\alpha$ increases, the relationship between content quality (or developer profits) and the commission rate changes. Specifically, (i) both the content quality and the developer’s profits decrease with a reduction in r in the rightmost region, and (ii) the content quality decreases but the content developer’s profits increase with a reduction in r in the middle region. The consumer surplus always decreases with a reduction in r throughout the presented region.²⁰

Regulators around the world are actively debating whether high commission rates harm innovation and consumer welfare. For example, the European Union recently required Apple to allow consumers to bypass the App Store’s payment system and avoid commission fees (Hancock 2024). Our findings caution against assuming that lower commissions necessarily foster innovation or benefit consumers. As Proposition 3 shows, even when reduced commissions lead to higher content quality, consumer welfare can decline if device prices rise in response. This occurs when commission rates are relatively low because the platforms then rely more heavily on device revenues. By contrast, when commission rates are high, a reduction can intensify device price competition and improve welfare. These results suggest that the welfare effects of commission cuts depend on the prevailing level of commissions and that regulators should carefully consider how commissions interact with device markets when assessing interventions.

A promising avenue for future work is to analyze settings in which commission rates are determined endogenously by competing platforms, which would provide a richer view of how setting commission itself interacts with device pricing and innovation incentives. Another direction is to consider competition among multiple developers or study market entry, and this could further shape how commissions and device pricing affect quality incentives and welfare.