Volume data: reading and judging the numbers
A volume figure is a summary of transactions that already happened. It carries no information about who made them, why, or whether the same balance moved back and forth. This desk is about closing that gap.
Volume is an aggregate, and aggregates hide things
Any volume number is a sum over a window. Summing destroys structure. Ten thousand dollars of volume could be four hundred separate wallets buying twenty-five dollars each, or one wallet cycling five hundred dollars twenty times. Both produce the same figure. Neither is visible from the figure alone.
This is not a criticism of the metric. Aggregation is what makes a number readable at a glance. The problem is treating the aggregate as if it were evidence about composition, which is exactly what most readers do, and exactly what a token page invites by showing volume in large type next to a price change.
The three checks this desk keeps coming back to
First, composition: what is the buy-versus-sell split inside the window, and how many distinct signers produced it. Second, persistence: does the activity survive the window boundary, or does it disappear the moment a leaderboard cut-off passes. Third, correspondence: did holder count move in the same direction as volume, and if not, what mechanism explains the divergence.
None of these checks requires special tooling. All three require going one level below the summary, usually into a block explorer, and all three are described concretely in the notes below. The desk also covers the tooling class that produces some of the activity you will see, because pretending that activity is always organic makes every other reading wrong.