The limits of every valuation ratio on this site
One page collecting the failure modes shared by all of them, so no ratio here gets read as a statement about what an asset is worth.
Every ratio in this track shares a set of weaknesses that no amount of care in construction removes. They are stated together here rather than repeated as footnotes, because a reader holding all of them at once will read any individual metric more accurately than one who meets them piecemeal.
None of what follows is an argument for ignoring the numbers. It is an argument for knowing precisely what they describe and where the description stops.
Nothing in the denominator is audited
Ledger data is verifiable in the sense that anyone can re-derive it from a node. That is not the same as being audited. What counts as a fee, which contracts belong to a protocol, which addresses belong to one entity, and how a multi-chain deployment is aggregated are all decisions made by whoever built the pipeline. Different providers make them differently. This site's choices are set out on the methodology pages, and the point of publishing them is that they are choices rather than discoveries.
Off-chain data is weaker still. Exchange volumes, holdings inside wrappers and custodial balances arrive from parties with an interest in the figure, and wash trading has been documented repeatedly on venues with weak surveillance. Any ratio using such a denominator inherits that weakness, and no amount of decimal places conceals it.
Both sides revise
Indexers backfill, correct and reclassify. A revision to an address-labeling set changes a historical series that a reader may already have used in a note. A chain reorganization can invalidate recent blocks. Supply figures change when a provider reassesses what counts as circulating. A ratio recorded on one date may not reproduce on another, and a chart drawn today may differ from the same chart drawn last month without anyone having made an error.
The numerator is most of the answer
Market capitalization moves faster and further than any denominator. Fee and revenue series are comparatively slow, because usage changes on a different timescale from sentiment. Over short windows, therefore, most of the variation in a valuation ratio is variation in price. A ratio that has halved has usually done so because the price halved, and describing that as a change in valuation is close to circular.
The problem compounds where the denominator is itself priced in the asset. Fees paid in the native token, and total value locked denominated in correlated assets, mean both sides move together. The residual signal is what remains after that correlation, and it is smaller than the headline movement suggests. Very little published commentary separates the two.
There is no discount rate and no terminal value
A multiple is a compression of a discounted cash flow model. It is defensible because, behind it, there is a stream, a discount rate reflecting the cost of capital, and a terminal assumption. Remove those and a multiple is a number with units, not a valuation.
For digital assets, all three are missing. There is no contractual stream. There is no accepted discount rate, because there is no established risk model and the risks include events with no probability estimate attached: a bridge exploit, a governance attack, an adverse ruling under the Howey test, or a change in the regulatory perimeter such as the one that arrived with MiCA entering application in 2024. And there is no terminal value, because a protocol carries no going-concern presumption. The 2022 failures of Terra, Celsius, Three Arrows and FTX established that entities and mechanisms in this sector can reach zero quickly and without an orderly process.
No ratio has a reference level
Equity multiples are interpretable because a long record of outcomes gives them a distribution. The ratios here do not have one. The sector is young, its composition has changed several times, and the population that would populate a historical sample includes many assets that no longer exist, which makes any backward-looking band a product of survivorship bias. Fitting a threshold to that record and applying it forward compounds the problem with look-ahead bias, since the fitting used information that was not available at the time the rule would have had to be applied.
This is why no page on this site labels a ratio as high or low in a normative sense. A number can be described relative to its own history and relative to a stated peer set. It cannot be described as correct, and a site that describes it that way has substituted an assumption for a measurement.
Cross-sectional comparison is the weakest use
The most common presentation, a table ranking assets by a single ratio, is the least defensible. A base-layer network, a decentralized exchange and a lending protocol earn fees through unrelated mechanisms, over different unit economics, with different supply structures and different competitive dynamics. Ranking them by market cap to fees produces an ordering with no interpretation attached to its positions. The next article covers what a defensible comparison set requires.
What remains true
Within these limits, the ratios do real work. They describe magnitudes: whether a network's valuation sits above a large fee stream or almost none. They describe direction: whether fee growth is moving with or against valuation. They describe structure: how much supply is unissued, how concentrated activity is, and how much of the fee take reaches the protocol rather than passing through it.
Those are facts, and facts narrow the space of things that can be true. Turning them into a view about worth requires assumptions that belong to a reader, not to a data page. The risk and incidents pages hold the failure record that any such assumption has to survive.