Intermediate level
DeFi dashboards report a handful of standard metrics. TVL is the value deposited, so it moves with prices and can count the same capital more than once. Fees are what users pay, revenue is the part the protocol keeps, and earnings subtract the tokens it pays out as incentives. Volume can be washed, and addresses are not people. Compare protocols on the same definitions, the same time window and the same denominator.
Five numbers and what they measure
Every DeFi dashboard leads with the same handful of figures. Each answers one narrow question, and each has a trap that catches people who read it as an answer to a different one.
| Metric | What it measures | The trap |
|---|---|---|
| TVL | The value of assets deposited in a protocol’s contracts | Moves with token prices, and the same capital can be counted in several places |
| Volume | The value of trades that went through | Cheap to fake on low fee venues, and easy to double count |
| Fees | Everything users paid to use the protocol | Not what the protocol keeps |
| Revenue | The part of fees the protocol keeps | Measured before the tokens it pays out to attract users |
| Active addresses | Unique addresses that transacted | Addresses are not people, and bots count too |
None of these is the market value of a protocol’s token. That is a separate number, and the DeFi market cap meter explains why the two can move in opposite directions for months.
TVL moves with prices
DefiLlama defines TVL for a protocol as the value of all coins held in its smart contracts, and compares it to assets under management. The key word is value. Suppose a protocol holds 10,000 ETH at $3,000, a TVL of $30 million. ETH falls 20%, nobody withdraws anything, and TVL drops to $24 million. The chart looks like an exodus that never happened.
So separate prices from flows. Read TVL in the native asset where the dashboard allows it, or use a flow metric: DefiLlama’s USD inflows multiply each day’s change in token balances by price, which shows zero in the example above. Also remember what TVL is: money that belongs to depositors and can leave. Deposits attracted by token incentives tend to leave when the incentives stop, which is why the guide to where yield comes from matters here.
Providers also disagree about what belongs in the number. DefiLlama’s published rules leave these out of the default figure:
| Left out by default | Why |
|---|---|
| Borrowed assets | Only coins actually sitting in the contracts count, which stops deposit, borrow and redeposit loops from inflating TVL |
| The protocol’s own token staked in it, and pools paired with it | Shown separately as staking and pool2, because minting your own token costs nothing |
| Vesting and non circulating tokens | A token with a small float could otherwise show a huge locked value |
| Liquid staking, in chain totals | Tracked as protocols but not added to a chain’s TVL by default |
| Bridges, in chain totals | Counted as protocols but not attributed to any chain |
| Native staking | Staked ETH or SOL securing the chain is never counted |
A protocol’s own website often uses the most flattering version. When two sources disagree, the definition is usually the reason.
How one dollar gets counted twice
DeFi is composable, so a receipt token from one protocol can be deposited in the next. Follow two users:
Alice stakes 100 ETH worth $300,000
A liquid staking protocol now shows $300,000 more TVL, and Alice holds a liquid staking token.
She deposits that token in a lending market
The lending market shows $300,000 more in deposits. It is the same ETH.
Bob supplies $150,000 of USDC to the same market, and Alice borrows it
A naive dashboard adds Bob’s $150,000 to the lending market’s deposits.
Alice adds the borrowed USDC to a DEX pool
The DEX shows $150,000 more TVL. It is Bob’s money.
Summing every deposit gives $900,000. The real capital is $450,000: Alice’s ETH and Bob’s USDC. Under DefiLlama’s default rules, which keep liquid staking out of chain totals and count only coins actually sitting in the lending contract, the chain total comes to the $450,000. Every extra layer, such as restaking the same token again, adds another claim on the same collateral, which is also why a failure at the bottom of the stack travels upward (see liquid staking tokens in practice).
Fees, revenue and earnings
DefiLlama models protocol finances as an income statement, and its data definitions are worth reading once in full. The chain of terms runs like this:
Work it through for a lending market whose borrowers pay $20 million a year in interest, with a 15% reserve factor and $5 million a year of token rewards (illustrative figures):
| Line | Amount | Where it goes |
|---|---|---|
| Fees | $20.0 million | All interest paid by borrowers |
| Supply side revenue | $17.0 million | Paid through to depositors |
| Revenue | $3.0 million | Kept by the protocol: treasury, team or token holders |
| Token incentives | $5.0 million | New tokens paid to users to attract activity |
| Earnings | −$2.0 million | Revenue after the cost of incentives |
Three readings follow. A protocol can have large fees and no revenue, as a DEX that passes every swap fee to its LPs does. It can have healthy revenue and negative earnings, because it is paying for its activity with its own token. And revenue only reaches token holders if a mechanism routes it there: DefiLlama reports token holder revenue (buybacks funded by revenue, fee burns, distributions to stakers) as its own line, and a governance token with no such mechanism shows zero however much revenue the protocol makes.
Volume, and how it gets washed
DefiLlama counts one side of each swap as DEX volume. It reports aggregator volume separately because those trades are already in the underlying DEXs’ figures, so adding the two double counts. Watch for that in any chart that sums categories.
Wash trading is trading with yourself to inflate the number, usually to farm volume based rewards or points, or to push a token up a trending list. It is cheapest where fees are lowest, and it can cost almost nothing when the washer is also the pool’s main LP, because the fees flow back to them. Signs to look for:
- Turnover out of proportion to liquidity. A pool with $2 million of liquidity doing $80 million a day turns over 40 times daily. Real demand rarely looks like that for long.
- Few addresses, identical sizes. The same handful of wallets trading round amounts back and forth.
- Timing. Volume that starts with a points campaign and stops when it ends.
DefiLlama says it removes volume it identifies as washed, including in low fee pools that fail its liquidity rules and in pairs distorted by incentive campaigns. Protocol websites and token listing sites often do not.
Active addresses are not users
One person can run hundreds of addresses, which is exactly what airdrop farmers do. One address can serve thousands of people, as an exchange or a smart contract does. Bots transact around the clock. Definitions differ too: DefiLlama counts only addresses that interact with a protocol directly, so users routed through an aggregator do not appear in a DEX’s count at all.
Better questions than "how many" are "how many came back" and "how concentrated is it". Addresses that return month after month, fees paid per active address, and the share of volume from the top ten addresses tell you far more than a daily headline. Answering them usually means querying the chain, which the guide to onchain SQL teaches.
Annualized figures and valuation ratios
Dashboards often annualize a recent period, multiplying 30 days of revenue by 365 ÷ 30, about 12.17. That is only as good as the month it starts from. A single liquidation cascade or an incentive burst can inflate a month and everything built on it. Prefer 90 day windows and look at the monthly bars, not only the total.
Ratios such as price to fees or price to revenue compare a token’s valuation with the business under it. The denominator you choose can reverse the ranking. Two illustrative protocols:
| Protocol X | Protocol Y | |
|---|---|---|
| 30 day revenue | $1.5 million | $0.6 million |
| Annualized revenue | $18.25 million | $7.3 million |
| Market cap | $300 million | $120 million |
| Fully diluted value | $900 million | $150 million |
| Market cap ÷ revenue | About 16.4 | About 16.4 |
| FDV ÷ revenue | About 49.3 | About 20.5 |
On market cap the two look identical. On fully diluted value, X is priced at more than twice Y, because two thirds of X’s supply is still to come. Neither ratio is right on its own: state which one you used, and pair it with the unlock schedule from researching a token. None of this is a signal to buy anything. It is context for a judgment.
Where to find the numbers
DefiLlama
TVL, fees, revenue, volume, unlocks and stablecoins across most chains, from open source adapters anyone can inspect. The best first stop, and its definitions pages explain every number.
Financial terminals
Token Terminal lays protocols out as financial statements period by period. Artemis pulls crypto and equity data into one research terminal. Useful for standardized comparisons; check each definition.
Dune
Community dashboards built on SQL queries over decoded chain data. You can usually open the query behind a chart, which makes it the place to answer your own questions.
Block explorers
The ground truth. When a dashboard number looks odd, read the contract balances, the events and the largest addresses yourself.
The site’s page on DefiLlama covers the interface in more detail.
Comparing two protocols fairly
- Same source, same definitions. Take both protocols from one provider, or confirm the definitions match.
- Same window. Compare 30 or 90 day totals, never one protocol’s best day against another’s average.
- Prices out. Use native units or inflows for TVL, so a price move is not mistaken for growth.
- Incentives out. Compare earnings, not only revenue, so paid activity is not mistaken for demand.
- Follow the money. Check token holder revenue if you are judging a token rather than a protocol.
- Check concentration. See how much volume or TVL comes from the largest few addresses.
- Verify one number yourself. Pick a figure and confirm it on an explorer. If it does not reconcile, find out why before trusting the rest.
Common questions
What is the difference between TVL and market cap?
TVL is the value of assets users have deposited in a protocol. Market cap is the value of the protocol’s own token. A protocol can hold billions in deposits while its token is worth little, or the reverse. TVL measures use, and market cap measures what the market expects.
Why do DefiLlama and a protocol's own site show different TVL?
Usually because of definitions. DefiLlama leaves out borrowed assets, the protocol’s own staked token, vesting tokens and some double counted deposits by default. A protocol’s own figure often includes some of these. Timing and price sources differ as well.
Is higher TVL always better?
Not by itself. TVL rises when prices rise and when incentives attract deposits that will leave later, and deposits are owed back to depositors. Fees earned per dollar of TVL, and whether deposits stay after incentives end, say more about the health of a protocol.
What is the difference between protocol revenue and token holder revenue?
Both are parts of revenue, the share of fees the protocol keeps. Protocol revenue goes to the treasury or team. Token holder revenue reaches the token through buybacks funded by revenue, fee burns or distributions to stakers. A token with no such mechanism has none, however profitable the protocol is.
How can I tell if DEX volume is wash traded?
Compare volume with liquidity, look at how many addresses produce it and whether trade sizes repeat, and check whether it began with a rewards campaign. Volume many times the pool’s liquidity every day, from a few wallets, during a points program, is the usual pattern.
Where to go next
Next: Living on more than one chain
Guide 15 of 16 in the Intermediate level. Your place is kept on this device, so you can pick up exactly here next time.