How it works
Every transaction on a public blockchain is visible. Analysts index those transactions, group addresses by entity (exchanges, miners, large holders, retail wallets), and aggregate the flows. Standard metrics include exchange inflow and outflow, miner reserve balance, supply held by long-term holders, and the realised value of recent moves. Services like Glassnode, Nansen, and Arkham publish these metrics with daily or hourly updates.
Example
A sustained outflow of BTC from exchanges into self-custody wallets historically signals accumulation, often preceding price rises by weeks. A spike in stablecoin issuance combined with stablecoin inflow to exchanges often precedes buying pressure on BTC and ETH. Conversely, large miner sells into exchanges can pressure price short-term. These are probabilistic signals, not certainties.
Why it matters
Price-only analysis is reactive; on-chain analysis is partially predictive because flows often lead price action. It is the closest crypto has to fundamentals data. The catch is interpretation: address heuristics are imperfect, exchange labels lag reality, and many metrics generated useful signal in one regime and noise in another. Treat on-chain signals as one of several inputs, not as the deciding factor.
