Public blockchains expose extensive activity data, yet a September BIS working paper finds that commonly used measures of crypto, stablecoin and DeFi activity can differ substantially depending on how researchers treat that data. Transparency of transactions does not remove the need to define the metric.

What the evidence shows

Blockchain records are transparent, but the economic meaning of each record is not self-evident. A single actor can control many addresses; a bridge can move the same asset through several representations; exchanges may move funds internally. The BIS working paper makes these choices central to interpreting activity measures.

A good dataset documents its chain coverage, contract selection, spam filtering, entity clustering and treatment of stablecoin minting and redemption. A growth chart should use the same method in both periods. When definitions change, an apparent surge may reflect a methodology revision rather than more economic use.

Why it matters

A bridge transfer, exchange movement and user payment can all produce on-chain transactions with different economic meanings. The same collateral may pass through multiple protocols. Comparing raw counts without accounting for this structure can overstate adoption or confuse changes in routing with new demand.

Deeper context and limits

Consider a stablecoin moving from an issuer to an exchange, then to a bridge and into a DeFi pool. Several on-chain transfers can occur without several independent end-user payments. Counting all of them as equal activity would exaggerate the economic event. Conversely, off-chain exchange activity may not appear in public-chain counts at all.

The choice of unit matters: transaction count, transfer value, active addresses, distinct entities and net issuance each answer a different question. Wash activity and incentives can distort all of them. A robust analysis compares measures, describes filtering and checks whether results survive reasonable alternative assumptions. 'On-chain' means observable records, not automatically unbiased interpretation.

Bull, bear and neutral cases

Confirmation and invalidation

Reproduce the metric with the published method and compare independent sources. A thesis is invalid if it relies solely on an unexplained transaction count.

Ask whether a report includes internal transfers, bots, bridges and repeated collateral. Compare definitions over time and publish the method next to any chart. This feature is an explanation of measurement limits, not an estimate of current market size.

Reader checklist

• Ask what a 'user' and a 'transaction' mean.
• Check for bridge and exchange routing.
• Demand comparable methodology across dates.

Published 26 September 2026. Dated report, not a live price feed or personal investment advice. Source documents may be revised after publication.