Price tells you what the market thinks. On-chain metrics tell you what's actually happening underneath it. Here's a starter toolkit for reading network health directly from the ledger.
Dario FennDeFi & Markets Lead · DeFi protocols, yield, market structure and on-chain data
On-chain metrics describe what a network is actually doing, but each carries a known distortion. Active addresses counts addresses rather than people; exchange flows are muddied by custodial rebalancing; MVRV indicates whether holders are collectively in profit rather than predicting direction. Read them together and against price, because a conclusion only one metric supports is usually that metric's distortion.
Price is the least interesting number a blockchain produces, because it's downstream of everything else and it's the number every participant is already staring at. The genuinely useful information — whether a network is actually being used, whether holders are accumulating or distributing, whether a token's circulating supply is about to expand — sits in the transaction data itself, freely available to anyone willing to look. Learning how to read on-chain metrics is less about finding secret data and more about knowing which public numbers actually mean something and which are vanity statistics dressed up to look analytical.
Active addresses — the count of unique wallets that transacted in a given period — is the most commonly cited on-chain metric and also the easiest to misread. A rising active address count is often taken as unambiguous proof of growing adoption, but it's trivially inflated by airdrop farming, wash trading, or a single actor cycling funds through hundreds of wallets to appear as organic activity. The more useful version of this metric isn't the raw count but its trend relative to transaction value and relative to new versus returning addresses. A network gaining genuinely new users looks different in the data to one where the same core group of addresses is simply transacting more often with each other — and conflating the two is one of the most common analytical mistakes newcomers make.
Net exchange flow — the difference between coins moving onto centralised exchanges and coins moving off them — is one of the more reliable behavioural signals available, because moving an asset onto an exchange is usually a precursor to selling it, and moving it off is usually a signal of intent to hold. Sustained net outflows, especially during periods of flat or falling price, tend to indicate accumulation by holders who aren't in a hurry to sell. Sustained net inflows ahead of a price rally are worth treating with some suspicion, since they can indicate holders positioning to sell into strength rather than confidence pushing the market higher. The caveat worth remembering is that exchange flows capture custodial movement, not ownership change — a large inflow could just as easily be an exchange consolidating its own wallets as it could be a whale preparing to sell, so it's a signal to combine with others, not to read in isolation.
Market cap — price multiplied by circulating supply — treats every coin as though it were acquired at today's price, which is obviously untrue and makes market cap a poor gauge of aggregate holder cost basis. Realised capitalisation fixes this by valuing each coin at the price it last moved on-chain, effectively summing up what the network's holders actually paid, in aggregate, for the coins they're sitting on. The ratio of market cap to realised cap — MVRV — then tells you roughly how far the market is trading above or below the network's aggregate cost basis. A very high MVRV has historically coincided with market tops, where the average holder is sitting on outsized unrealised gains and the incentive to take profit is strong; a very low or negative MVRV has historically coincided with capitulation bottoms, where the average holder is underwater and further selling pressure tends to be exhausted. It isn't a timing tool on its own, but it's one of the more grounded ways to answer "is this market historically stretched or historically cheap" without relying purely on price charts.
Most retail attention goes to demand-side signals — volume, addresses, social sentiment — while the supply side gets comparatively little scrutiny despite being just as decisive. Token unlock schedules are the clearest example: a protocol with a large tranche of investor or team tokens vesting on a known date is facing a mechanical increase in sell-side pressure that has nothing to do with sentiment or fundamentals, and that date is public information sitting in the tokenomics documentation, not a secret. Circulating supply relative to fully diluted valuation is worth checking for the same reason — a token trading at a modest market cap but a huge fully diluted valuation is one where today's price doesn't reflect the supply still to come, and that gap tends to close through downward price pressure as unlocks land rather than through the market simply absorbing it painlessly.
Holder distribution — how concentrated a token's supply is across wallets — is a blunt but genuinely useful check. A token where the top ten non-exchange wallets hold a large share of circulating supply is one where a small number of decisions can move the market disproportionately, regardless of how healthy the rest of the metrics look. It's worth distinguishing genuine concentration from false alarms, though: some of the largest addresses for any given token are usually exchange hot wallets, bridge contracts, or staking contracts rather than a single whale, so the raw "top holder" list needs a bit of labelling before it tells you anything. Where the data is available, tracking how concentration has changed over time — is supply spreading out to more holders, or consolidating into fewer — tends to be more informative than a single snapshot.
For layer-1s and layer-2s specifically, a few less-cited metrics are worth adding to the toolkit. Fees paid to the network, rather than transaction count alone, capture how much economic activity users are actually willing to pay for, which filters out a lot of the low-value spam that inflates raw transaction counts. Total value locked is useful for gauging DeFi activity on a chain but is easy to double-count across protocols that build on top of each other, so treat cross-chain TVL comparisons with some caution rather than as a clean apples-to-apples ranking. Developer activity — commits, active repositories, unique contributing addresses to a chain's core codebase — is a slower-moving signal but a genuinely predictive one, since sustained building tends to precede sustained usage rather than the other way round.
For any asset with a liquid futures and options market, on-chain spot metrics tell only half the story, and the other half sits in derivatives data that's public but rarely checked with the same rigour. Funding rates on perpetual futures — the periodic payment traders make to each other to keep the perpetual price anchored to spot — reveal positioning that price alone hides. Persistently high positive funding means longs are paying shorts to stay in their positions, which is a sign of crowded, leveraged bullishness that tends to precede sharp downside liquidation cascades rather than sustained rallies. Open interest climbing alongside price is a healthier signal than open interest climbing while price stalls, since the latter often means leverage is building without genuine conviction behind it. None of this replaces on-chain analysis, but combining the two — what holders are doing with their actual coins, and what traders are doing with leveraged bets on the price of those coins — closes a gap that either data set leaves wide open on its own.
None of these metrics is a standalone trading signal, and treating any one of them as such is how people end up overconfident in a read that a second metric would have contradicted. The actual discipline is building a small dashboard — active addresses alongside exchange flows, MVRV alongside supply unlock schedules, holder concentration alongside developer activity — and looking for the metrics to agree with each other before drawing a conclusion. When exchange outflows, low MVRV, and rising developer activity all point the same direction, that's a meaningfully stronger signal than any one of them in isolation. When they conflict, that's useful information too — it usually means the story the price action is telling isn't the whole story, and that's precisely the gap on-chain data exists to fill.
Smart contracts can't see the outside world on their own. Oracles are the systems that feed them prices, events, and data — and getting that design wrong has cost DeFi protocols hundreds of millions.
Perpetual futures now trade several multiples of spot crypto volume most days. Understanding how perps and funding rates actually work is the difference between using leverage deliberately and being liquidated by it.
A token's price chart tells you nothing about the supply mechanics working against it. Here's a framework for reading tokenomics properly before you decide what a token is actually worth holding.