Fees vs revenue vs profit: the DeFi accounting guide

Crypto analytics is full of mislabeled charts. If you can't separate fees, revenue, and profit, you can't compare protocols, and you definitely can't value tokens.

Fees: The Gross Top-Line

Fees represent what users pay to use the protocol. When you swap tokens, you pay a swap fee. When you borrow, you pay interest. When you trade perpetuals, you pay trading fees and funding rates. These are gross payments flowing into the system.

Swap fees on a DEX are typically 0.01% to 1% of transaction value. A $1M swap at 0.3% generates $3,000 in fees. That number is observable onchain. Every transaction records the fee amount. Aggregating across all swaps gives total fees collected.

Interest payments on lending protocols are accrued over time. Borrowers pay interest continuously based on their outstanding loan balance and the current rate. A borrower with a $100K loan at 5% APY pays $5,000 annually. Summing across all active loans gives total interest fees.

Spreads and funding rates in derivatives markets create implicit fees. A perpetual futures exchange might capture the difference between the index price and perpetual price through funding rate mechanisms. Market makers pay takers or vice versa. The protocol can extract value from this flow.

Not all fees are created equal. Voluntary fees from users choosing to trade are different from mandatory fees extracted through protocol mechanics. Liquidation penalties are fees, but they're not a sign of healthy usage. MEV extracted from users is economically a fee, but it's not always attributed to the protocol.

Revenue: What's Retained

Revenue is the portion of fees that the protocol keeps after all mechanically required distributions. The difference between fees and revenue is pass-throughs.

On most DEXs, liquidity providers receive the majority of swap fees. Uniswap v2 and v3 historically sent 100% of fees to LPs. The protocol retained zero revenue. Uniswap v3 introduced an optional protocol fee that governance could enable. Until activated, protocol revenue remained zero despite billions in fees collected.

Lending protocols retain the spread between borrow rates and supply rates. If borrowers pay 5% and suppliers receive 3%, the protocol captures 2% as revenue. That 2% might fund the treasury, get distributed to tokenholders, or pay for protocol operations. The 3% paid to suppliers is not revenue. It's the cost of attracting capital.

Stakers represent another pass-through. Protocols that require staking for security or governance often pay stakers from fee revenue. If a protocol collects $10M in fees and distributes $7M to stakers, the protocol revenue is $3M. The $7M is a required payment to maintain network security.

Treasury versus tokenholder revenue creates another layer. Some protocols send revenue to a DAO-controlled treasury. That revenue benefits the protocol's long-term sustainability but doesn't immediately accrue to token prices. Other protocols distribute revenue directly to tokenholders through buybacks, burns, or dividends. The economic reality differs.

The challenge is attribution. Smart contracts execute distributions automatically. Identifying which wallets represent the treasury, which represent staker rewards, and which represent LP payments requires interpretation. Most dashboards default to reporting fees because the calculation is cleaner.

Profit: What's Left After Incentives

Profit requires subtracting expenses from revenue. The largest and most commonly ignored expense in crypto is token emissions.

When a protocol distributes tokens to users as incentives, that's an expense. The tokens have economic value. Users receive them and can sell them. The protocol is paying users in dilutive currency. The accounting should reflect this cost.

Measure emissions at fair value at the time of distribution. If a protocol distributes 1 million tokens worth $5 each, that's a $5 million expense. Compare that to revenue. If revenue is $3 million, the protocol is unprofitable by $2 million. Existing tokenholders bear this cost through dilution.

Subsidy dependence is the norm, not the exception. Many protocols operate at a loss for extended periods. This can be rational during growth phases. Customer acquisition costs in traditional businesses are upfront expenses that pay back over time if retention is strong. The question is whether crypto protocols are buying durable users or renting mercenaries.

Unit economics reveal sustainability. If the protocol spends $100 in tokens to acquire a user who generates $10 in lifetime fees, that's a bad trade. If the user generates $500 in fees, it's a good investment. Most protocols don't disclose or track cohort economics. This makes profitability analysis harder.

Quick sanity checks help. Compare annual emissions value to annual revenue. If emissions are 3x revenue, the protocol is deeply unprofitable. If emissions are 0.3x revenue, the protocol might be approaching sustainability. If emissions are zero and revenue is positive, the protocol is genuinely profitable.

Model Differences Across Protocol Types

The relationship between fees, revenue, and profit varies by protocol architecture.

DEXs collect swap fees. Most DEXs distribute 100% to liquidity providers initially. Protocol revenue is zero. Later versions might introduce protocol fees, typically 10-20% of total fees. Revenue becomes a small fraction of fees. Profitability depends on whether the protocol issues tokens to incentivize liquidity.

Lending protocols collect interest from borrowers and pay interest to suppliers. Revenue is the spread. Profitability depends on whether the spread covers bad debt, operational costs, and token emissions. Many lending protocols subsidize lower rates through token distributions, making them unprofitable despite positive revenue.

Perpetual futures protocols collect trading fees, capture funding rate spreads, and earn liquidation fees. Revenue after paying market makers and insurance fund allocations can be substantial. Profitability requires emissions to be lower than this retained revenue. Some perps protocols achieve this. Many don't.

Liquid staking protocols collect staking rewards from validators and distribute most of it to depositors. The protocol retains a small percentage, typically 5-10%. Revenue is this retention. Profitability is high because operating costs are low and emissions are often zero. This is one of the few protocol types with clear paths to sustainable profit.

L2 sequencers collect transaction fees from users and pay data availability costs to L1. The difference is gross revenue. Net revenue depends on operational infrastructure costs. Profitability depends on whether the L2 issues tokens to subsidize low fees. Some L2s operate profitably. Others burn capital to maintain cheap transactions.

A Quick Comparison Checklist

When comparing protocols, ensure like-for-like comparisons. Different business models require different frameworks.

Within the same category, compare retained revenue as a percentage of fees. DEX A retains 15% of fees as revenue. DEX B retains 10%. DEX A has a better take-rate, all else equal. Check if this is structural or configurable through governance.

Across categories, compare capital efficiency and unit economics. Lending protocol revenue per dollar of TVL versus perps protocol revenue per dollar of TVL reveals which model monetizes capital more effectively. But recognize the business models are different. Lending has credit risk. Perps have liquidation risk.

Always normalize for emissions. Protocol A has $50M revenue and $10M emissions. Net: $40M. Protocol B has $50M revenue and $80M emissions. Net: -$30M. Raw revenue is identical. Economic reality is opposite. Protocol A is profitable. Protocol B is subsidy-dependent.

Track trends over time. Is revenue growing faster than fees? That means take-rate is improving. Is revenue growing slower than fees? That means the protocol is passing more value to users or intermediaries. Is profit improving while emissions decline? That's the path to sustainability. Is profit declining while emissions increase? That's reflexive growth.

Avoid misdefinitions. If a dashboard labels fees as revenue, adjust your mental model. If it doesn't account for emissions, add that expense manually. If it combines treasury revenue with tokenholder distributions, separate them. Precision in definitions prevents capital misallocation.

For standardized definitions across protocols, refer to DefiLlama's data definitions, which explain their methodology for distinguishing fees, revenue, and other metrics consistently.

Related Concepts

Understanding protocol accounting requires these related frameworks:

• Protocol revenue: Deep dive into what counts as retained revenue
• Emissions vs revenue: How to account for token incentives as expenses
• DeFi income statement: Apply GAAP-like principles to protocol analysis
• Take rate: The percentage of fees a protocol actually keeps
• Cost of funds in DeFi: Understanding pass-throughs to capital providers
• Onchain profit: The bottom line after all expenses