No published 24-hour volume figure for a crypto exchange has been independently audited. It is a number the exchange reports itself, then filtered by whichever site is displaying it – and the
No published 24-hour volume figure for a crypto exchange has been independently audited. It is a number the exchange reports itself, then filtered by whichever site is displaying it – and the filtering rules differ enough between CoinMarketCap, CoinGecko, Coin Metrics and Nomics that the same exchange can show meaningfully different volume depending on which one you check.
Trading volume is easy to manufacture. Coin Metrics data scientist Jon Geenty told Decrypt that exchanges are “notorious for boosting volume numbers” to climb ranking sites, according to Decrypt’s report published 28 July 2020. CoinMarketCap’s own methodology documentation makes a similar point about specific trading pairs, describing pairs with no fees or heavy trading incentives as “susceptible to wash trading, resulting in artificially inflated reported volumes” and excluding them from its calculation. CoinGecko’s Trust Score Methodology page makes the same underlying point without a fixed formula: it says exchanges in crypto markets can inflate what they report, which is why the Trust Score looks past the headline figure rather than ranking on it directly.
CoinMarketCap takes an exchange’s own reported 24-hour figure and converts it to US dollars using reference prices, according to its support documentation, which carries no publication date. Two of its rules matter for judging whether a listed number is inflated. First, it excludes trading pairs with no fees or significant trading incentives, on the reasoning that subsidized or fee-free trading is where wash trading concentrates. Second, rather than summing every pair an exchange lists, CoinMarketCap averages only the top 25 pairs, specifically so an exchange cannot pad its ranking with a long tail of inactive pairs, per the same documentation. Neither rule tells a reader what share of any specific exchange’s current volume was excluded – the documentation describes a method, not a per-exchange result.
CoinGecko treats reported volume as one input rather than the primary signal. This is set out across two separate documents. Its Trust Score Methodology page describes the score as considering liquidity – meaning volume consistency and order book depth – alongside cybersecurity history, regulatory status, past incidents and whether an exchange publishes proof-of-reserves. Exchanges whose reported figures deviate sharply from reference benchmarks are penalized for that deviation, per the same page. A separate document, CoinGecko’s volume methodology page, covers a narrower mechanical rule: it excludes trading pairs blacklisted for inconsistent data or left more than three hours without an update, and calculates volume using a volume-weighted average price across a set of reputable exchanges. That page also states that in May 2026 CoinGecko rebuilt the Trust Score to replace a web-traffic comparison with direct evaluation of volume and order book depth – worth knowing if you are comparing a Trust Score from before that change with one from after it.
Two outside attempts to grade exchange volume illustrate how wide the gap between reported and “real” volume can look on paper, though both results are now years old.
In 2020, Coin Metrics built what it called a Trusted Volume Framework, certifying only 13 exchanges – among them Binance, Bitfinex, Bitstamp, Coinbase, Gemini and Kraken – as of 1 July 2020, according to Decrypt’s report published 28 July 2020. The framework ran each exchange through three tests: whether its price feed correlated at 80% or higher with a set of established exchanges, whether its reported volume looked proportionate to its web traffic, and a qualitative score covering features such as KYC requirements and regulatory status, per Decrypt. On that basis, Coin Metrics put the market’s real 24-hour volume at roughly $13.25 billion against nearly $113 billion in volume reported by exchanges, as of 1 July 2020 – Decrypt described the real figure as “close to one tenth” of the reported one. One exchange, OKEx, failed all three tests; BiBox, ZB.com, LBank and HitBTC each failed two, according to Decrypt’s account of the same framework.
A year earlier, analytics firm Nomics launched a Transparent Volume service grading exchanges from A to D on how much raw trading data they published, per FxStreet’s report published 28 August 2019. Coinbase Pro, Kraken and Binance were cited as exchanges that could earn an A or A+ rating by publishing a complete trading history, while a D rating meant an exchange provided only ticker-level data – daily volumes and prices with no underlying detail – according to FxStreet. The same article cited a Bitwise Asset Management report finding that 95% of trading volume on unregulated exchanges was fake. That figure comes from FxStreet’s account of the Bitwise report, not from this desk’s own reading of it.
The International Organization of Securities Commissions published 18 policy recommendations for crypto and digital asset markets on 16 November 2023, covering conflicts of interest, market abuse, custody of client assets and disclosure, according to IOSCO’s final report. The recommendations are aimed at regulators building rules for crypto-asset service providers – they address how exchanges should govern themselves and disclose trading arrangements, not how a data aggregator should calculate a 24-hour volume figure. IOSCO’s report does not set out a volume-verification standard of its own.
The mistake is treating the “24h Volume” figure on an aggregator site as a settled, comparable fact, the way a reader might treat a company’s audited revenue line. It isn’t that. It is a number the exchange supplied, passed through one aggregator’s filtering rules, which may differ from another aggregator’s rules applied to the same exchange on the same day. A number that looks large on one site can look ordinary on another simply because of which pairs got excluded and which weighting method was used.
This page explains how the major volume-filtering methodologies work in principle. It does not tell you whether any specific exchange’s volume figure, right now, is real – that would require applying each framework to current data, and none of the evidence gathered here does that.
The Coin Metrics and Nomics material is old. The Coin Metrics result is dated 1 July 2020, as reported by Decrypt on 28 July 2020, and the Nomics grading and Bitwise finding come from an FxStreet article published 28 August 2019. Nothing in the evidence for this page shows whether either framework is still maintained, or whether the exchanges named above still carry the same ratings today.
The CoinGecko and CoinMarketCap descriptions here come from partial summaries of their published methodology pages, not the full documents, and both carry no publication date – so thresholds and edge cases beyond what is quoted above, such as the exact scoring weights behind CoinGecko’s Trust Score, are not covered.
The 95% figure attributed to Bitwise comes secondhand, through FxStreet’s account of the Bitwise report, and has not been checked here against the underlying Bitwise report.
None of the four frameworks described above is an independent financial audit. Each is a methodology designed and run by a data provider, not verified by an outside auditor the way a reserves attestation would be – so a high rating from any one of them reflects that provider’s own criteria, not a regulator’s or auditor’s sign-off.
Every fact above is attributed to one of these reports. Where they disagree, the article says so.
The post Which exchange volume figures can be trusted, and why appeared first on TheCoinrise.com.
