How much service fee should Web3 platforms charge?
TechFlow Selected TechFlow Selected
How much service fee should Web3 platforms charge?
Designing a合理的 fee structure is not at odds with decentralization, but rather a core element in building a functional decentralized market.
Navigating Web3 tides with focused insightsAuthors: Gérard Cachon, Tolga Dizdarer, Gerry Tsoukalas
Translated by: Luffy, Foresight News
Web3 aims to reduce reliance on intermediaries, thereby lowering service fees and giving users greater control over their data and assets. For example, Gensyn (a decentralized AI computing platform) offers AI compute services at a fraction of the cost of Amazon Web Services (AWS); Drife (a decentralized mobility platform) promises to free drivers from Uber's commission fees as high as 30%.
However, while the idea of reducing costs for users is appealing, establishing fair fee and pricing structures requires platforms to balance multiple interests. The most successful decentralized markets do not completely abandon fees; instead, they combine "decentralized pricing" with well-considered, value-adding fee structures to achieve supply-demand equilibrium.
Based on our research, this article explains the role of pricing control and fee structures in platform economics and governance; why "zero-fee" models are ultimately doomed to fail regardless of their designers' good intentions; and how blockchain platforms should design pricing strategies. We propose a new transaction-volume-based "affine pricing" model that resolves the tension between private information and market coordination.
Why Pricing and Fees Matter
The success or failure of digital platforms hinges on their ability to manage two core levers: pricing control and fee structure (i.e., how much the platform charges buyers and sellers for using its services). These are not merely revenue tools but also market design instruments that shape user behavior and determine market outcomes.
Pricing control determines "who sets the transaction price." For instance, Uber uses centralized algorithms to set fares, optimizing for supply-demand balance and pricing stability; in contrast, Airbnb grants hosts autonomous pricing power, offering only algorithmic suggestions as mild guidance. Each model addresses different priorities: centralized pricing ensures coordination efficiency in large-scale markets; decentralized pricing allows service providers to incorporate private information (such as costs, service quality, and differentiation advantages) into their pricing decisions. Neither approach is universally superior—their effectiveness depends on the specific application context.
Fee structures influence more than just platform revenue—they determine which participants enter the market and how it operates. Apple’s App Store charges up to a 30% commission, funding infrastructure and filtering high-quality app supply, though it may frustrate developers—but typically does not directly affect users. In contrast, Ticketmaster’s high fees drive artists and fans to alternative channels when available. On the low-fee end, Facebook Marketplace’s free listing service has fostered fraud; numerous near-zero-fee NFT platforms have suffered degraded user experiences due to floods of low-quality NFTs.
The pattern is clear: excessively high fees drive away suppliers; excessively low fees undermine service/product quality.
Many blockchain projects adopt zero-commission models based on the logic that by forgoing value extraction, better outcomes emerge for suppliers and users. But this view overlooks the critical role well-designed fees play in enabling effective market operation: fees are not merely extraction tools—they can serve as coordination mechanisms.
The Trade-off Between Information and Coordination
The central challenge in platform design lies in balancing "leveraging private information from service providers" against "coordinating the market for efficiency." Our research shows that the interaction between pricing control and fee structure determines whether this trade-off is mitigated or exacerbated.
When platforms set prices directly, they can more easily coordinate supply-side efforts and align competitive behaviors among providers. However, because platforms lack access to each supplier’s private costs (e.g., operating costs, marginal costs), such pricing often results in mismatches: prices too high for some users, too low for some suppliers. Platforms typically charge commissions based on transaction value, and inefficient pricing ultimately leads to profit leakage.
If service providers set their own prices, those prices could theoretically reflect true costs and capabilities: lower-cost providers can gain competitive advantage through lower prices, achieving better supply-demand matching and higher market efficiency. But uncoordinated pricing can backfire in two ways.
When products or services are highly homogeneous, it can trigger destructive price wars. High-cost providers are forced out of the market, reducing supply—just as demand may be rising—ultimately weakening the platform’s ability to meet demand. While lower average prices might benefit consumers, they directly threaten the platform’s commission-based revenue model.
When products or services must be combined to deliver maximum value, providers tend to set prices too high. Although many suppliers may join the platform, individually high prices collectively raise average market prices, driving users away.
This is not mere theory: in 2020, Uber tested its “Project Luigi” in California, allowing drivers to set their own prices. Results showed drivers consistently set fares too high, causing users to shift to other platforms. The program was terminated after about a year.
Key conclusion: these outcomes are not accidental—they represent equilibrium results under standard commission contracts. Even optimized commission contracts may lead to persistent market failures. Thus, the central question is not "how much commission should the platform charge," but "how to design fee structures so the market works effectively for all participants."
Solving the Problem
Our research identifies a targeted fee structure that elegantly resolves market coordination issues while preserving the benefits of "personalized pricing." This affine fee model uses a two-part tariff, where service providers pay the platform:
-
A fixed base fee per transaction;
-
A variable fee that increases with transaction volume (surcharge) or decreases with transaction volume (discount).
This model differentially affects providers based on their cost structure and market position.
In such markets, provider costs vary significantly: some have naturally lower costs due to advanced technology, access to renewable energy, or efficient cooling systems; others have higher costs but offer premium services like high reliability.
Under traditional commission models, if competition becomes excessive, low-cost GPU providers will set aggressively low prices, capturing disproportionate market share and triggering the market distortions described earlier: some suppliers exit, limiting transaction volume, while average market prices fall.
For this scenario, the optimal solution is a "transaction-volume surcharge": the more customers a provider serves, the higher the fee per transaction.
This mechanism imposes a "natural constraint" on aggressive low-cost providers, preventing them from capturing excessive market share via unsustainable pricing, thus maintaining market balance.
When competition is moderate or insufficient, the optimal strategy shifts to a "transaction-volume discount": the more customers a provider serves, the lower the fee per transaction. This incentivizes providers to expand volume through price reductions, effectively boosting competitiveness without pushing prices below sustainable levels.
For example, on a decentralized social platform, creators with higher user engagement could be charged lower fees, encouraging them to set more competitive prices for paid content while attracting more users.
The elegance of the affine fee mechanism lies in its independence from detailed knowledge of individual provider costs. The fee structure creates positive incentives that guide providers to self-regulate based on their private cost information. Low-cost providers still gain advantage through lower prices than high-cost competitors, but the fee structure prevents them from monopolizing the market in ways that harm overall ecosystem health.
We validated through mathematical simulations that a properly calibrated "transaction-volume-based fee structure" enables platforms to achieve over 99% of theoretical optimal market efficiency. Within our theoretical framework, it significantly outperforms both "centralized pricing" and "zero-fee" models. The resulting market exhibits the following characteristics:
-
Low-cost providers retain competitive advantages but do not dominate market share;
-
High-cost providers can sustainably participate by focusing on niche markets offering differentiated services;
-
The overall market reaches a more balanced equilibrium with reasonable price dispersion;
-
The platform achieves sustainable revenue while enhancing market functionality.
Moreover, analysis shows that the optimal fee structure depends on "observable market characteristics," not on individual providers’ "private cost information." When designing contracts, platforms can use observable signals such as "price" and "transaction volume" as proxies for "latent costs," allowing providers to retain pricing autonomy based on private information while resolving inherent coordination failures in fully decentralized systems.
Future Pathways for Blockchain Projects
Many blockchain projects, by adopting traditional commission models or zero-fee models, have compromised both financial sustainability and market efficiency.
Our research confirms that well-designed fee structures are not contrary to decentralization—they are essential components for building functional decentralized markets.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
a16z
