TVL research: separate deposits from asset prices
TVL research should separate changes in deposited asset quantities from changes in the prices used to value them. A higher dollar total is not, by itself, evidence of new deposits, new users, or higher protocol revenue. Start with the metric’s definition and then reconstruct the movement.
DefiLlama’s data definitions describe its treatment of TVL and exclusions. That is a provider methodology rather than a universal accounting rule. Use the definition belonging to the actual series you inspect, and record its scope before comparing it with another provider.
Create a quantity and price worksheet
For each included asset, retain its contract identity, opening quantity, closing quantity, and corresponding prices. Record the contract set included in the protocol boundary. If the provider changes coverage, treat that as a methodology event rather than silently calling the resulting increase a deposit.
Use a fixed-price comparison to isolate one component. Revalue closing quantities at opening prices, then compare that subtotal with opening TVL. This estimates the quantity contribution under a stated decomposition. It does not automatically measure net external deposits because accrued balances, rewards, internal movements, or classifications may also change quantities.
| Worksheet column | Meaning |
|---|---|
| Q0 and Q1 | Asset units at opening and closing observations |
| P0 and P1 | Consistent prices at those observations |
| Q0 × P0 | Opening value |
| Q1 × P0 | Closing quantity at opening prices |
| Q1 × P1 | Closing value |
| Scope adjustment | Contracts or assets added or removed |
Document the decomposition convention. With several changing variables, alternative ordering can allocate interaction effects differently. Consistency matters more than presenting one allocation as uniquely correct.
Hypothetical unchanged deposits
Assume a fictional protocol contains 100 units of asset A at both endpoints. A’s reference price rises from 10 to 12 reporting units. TVL increases from 1,000 to 1,200, while the asset quantity is unchanged. Reporting a 20% increase in deposits would be incorrect. The defensible observation is a 20% increase in measured value under the selected price series.
Now suppose the closing quantity is 110 units. At opening prices, the closing balance is worth 1,100; at closing prices it is worth 1,320. Under this worksheet, 100 of the increase is allocated to quantity and 220 to repricing of the closing quantity. Those ten added units still need transaction-level analysis before being called external inflows.
All numbers are hypothetical. They illustrate a measurement bridge, not the actual performance of any protocol or an expectation about future asset prices.
Inspect duplicated claims
Look for receipt tokens, borrowed assets, staked representations, and positions held through other protocols. A deposited asset and a claim on that deposit may appear at several layers. Whether a provider excludes a layer must be read from its methodology rather than assumed.
Create a small dependency diagram for the largest positions. Identify which balances represent distinct underlying assets and which represent claims on already counted assets. Report a gross view and a defined adjusted view when useful, labeling both clearly. Do not publish an adjusted figure without showing the exclusions.
For time comparisons, keep the same observation convention and address coverage. For peer comparisons, align categories and borrowing treatment. An apparent gap between protocols may come from different inclusion policies rather than different economic usage.
Limits and common mistakes
TVL does not directly measure solvency, revenue, user satisfaction, security, or available spot trading depth. It may change because of valuation conventions, stale prices, inaccessible assets, or methodology revisions. Quantity data can also be incomplete when contract coverage is undocumented. In such cases, limit the conclusion to the observed series and identify unverified components.
Common errors include treating TVL growth as user growth, calling all quantity changes net deposits, adding recursive positions without inspecting dependencies, and mixing daily closing values with intraday observations. A useful note ends with the decomposition, the coverage definition, and the unresolved contribution rather than an unqualified growth claim.
FAQ
Is rising TVL evidence of new money entering a protocol?
Not by itself. Prices, quantities, and scope can all change. External flow requires a separately defined and verified transaction analysis.
Can TVL compare two protocols directly?
Only after aligning inclusion rules, asset valuation, observation times, and relevant product categories. Otherwise the comparison may be mostly a methodology difference.