How to Read On-Chain Data

On-chain analysis involves examining transactions and balances recorded on the blockchain to gain insight into network usage and asset movements. In Turkish, it is called “zincir üstü analiz.” Public records are visible to everyone; however, to whom an address belongs or why a transfer was made is not always known. This distinction between data and interpretation is the most significant limitation of the analysis.

What Is Seen on the Chain, and What Is Not?

Information such as the asset sent, amount, addresses, fees, and block time can be seen in a transaction. Depending on the network’s model, contract calls, token transfers, and transaction outputs can also be examined. In contrast, the agreement between parties, an order within an exchange, or the personal purpose of the transfer is not directly written on the chain. The same movement could be preparation for a sale, a change in custody, or a transfer between one’s own accounts.

Trading on centralized exchanges mostly occurs within the internal records of the exchange. Not every match turns into a blockchain transaction. Therefore, it is not appropriate to compare the number of on-chain transactions directly with exchange volume. An exchange can batch withdrawals from many customers into a single chain transaction; a single transaction count can include many economic movements.

Address Count Is Not User Count

One person can use multiple addresses. Conversely, an exchange address can represent the assets of thousands of customers. The number of active addresses measures addresses that transacted during a specific period; it does not directly count unique individuals. An increase in new addresses may indicate expanding usage, but automated accounts, address-switching habits, or app design can also affect the figure.

Data providers may attempt to aggregate related addresses into clusters belonging to the same entity or institution. This is called entity adjustment. These methods use estimation and heuristic rules; they are not definitive public identity records. If clusters change over time, past metrics may also be updated. The fact that a chart is “adjusted” does not mean all uncertainty has been eliminated.

Does an Inflow to an Exchange Necessarily Mean a Sale?

Seeing a transfer from an address to one believed to belong to an exchange can provide evidence that an asset was sent to the exchange. However, it does not alone prove that a sell order was placed or that a sale occurred. The user might be depositing collateral, using a custody service, or transferring to another account. Furthermore, movement between an exchange’s own hot and cold wallets can also be seen.

For example, 1,000 BTC going to a known exchange address might be a significant movement. To say “1,000 BTC were sold,” additional evidence is required. A transfer between two addresses of the same exchange should not be counted as a new inflow from the outside. This is why the source of the label, the address cluster, and the transaction structure are important. The fact that a large figure is noteworthy does not mean its purpose is clear.

How Is Net Flow Calculated?

Net flow is calculated by subtracting outflows from inflows to exchanges for a given period. Hypothetically, if there are 5,000 BTC in inflows and 4,200 BTC in outflows on the same day, the net inflow is 800 BTC. This does not mean that all 5,000 BTC were sold on the market or that 800 BTC represents new investor demand. The calculation only shows the difference between the amounts entering and leaving the defined set of addresses.

The net figure can also mask the magnitude of gross movements. 100 BTC in and 100 BTC out produces the same zero net result as 100,000 BTC in and 100,000 BTC out. Despite this, the level of activity is different. Therefore, seeing gross inflows and outflows, the time frame, and which exchanges are covered alongside net flow is more explanatory.

Coins That Have Not Moved for a Long Time

The time elapsed since the last movement of coins is used in some analyses to examine holding behavior. However, if a coin has not moved for a long time, the owner is not necessarily a long-term investor. Keys may have been lost, institutional custody may be in use, or the individual simply may not have changed addresses. Movement does not necessarily mean a sale either.

Age calculations can be performed via unspent transaction outputs (UTXO) in Bitcoin. A change output returning to your own address can also be created in a new transaction. How the provider classifies these affects the metric. When reading the phrase “old coins have moved,” it is necessary to distinguish how much volume was measured, at what age threshold, and with which adjustment method.

What Do Realized Value and Realized Profit Tell Us?

Some metrics value coins at the price they last moved on-chain instead of today’s price. This approach differs from market capitalization, which multiplies the entire supply by the current market price. Using the last movement price can provide a cost estimate; however, it does not definitively give the coin owner’s actual purchase cost. Purchases occurring within an exchange or transfers between one’s own addresses can disrupt this relationship.

For example, a person might buy BTC on an exchange at $40,000 and withdraw it to a personal wallet when the price is $45,000. Even if the last movement price seen on-chain is $45,000, the actual purchase price is different. Moving it to another of one’s own addresses later also does not create a real realized profit from a sale. The presence of the word “realized” in the metric name does not mean it carries the same certainty as an accounting or tax record.

Comparing Transaction Counts of Different Networks

A Bitcoin transfer, an Ethereum contract call, and a system transaction on another network may not represent the same workload. Some networks may include votes or validation messages in their transaction counts. A smart contract transaction can create numerous sub-movements. A portion of Layer 2 activity may also be aggregated under a single data submission on the mainnet.

For this reason, you cannot claim to have found the most used network simply by ranking daily transaction counts. Additional information is needed, such as the definition of a user transaction, whether failed transactions are included, fees, and economic value. Monitoring the change within a time series using the same methodology can be more meaningful than comparing raw numbers with different definitions.

Workflow for Evaluating a Chart

First, read the definition of the metric: is the unit in coins or dollars, a daily total or a moving average? Next, find the data scope and labeling method. Finally, investigate whether a large change stems from actual usage, price movement, a new address label, or a data adjustment. Even if several independent indicators point in the same direction, this does not make the future price certain.

In your own notes, separate the observed event from the interpretation. “Net inflows to labeled exchange addresses increased” is a data statement; “investors will definitely sell” is an unproven inference. On-chain analysis is a powerful tool for making sense of public records. Its strength comes not from describing unseen intentions on the chain as if they were certain, but from clearly showing what is being measured.

Sources

• Glassnode: address clusters and adjusted metrics
• Glassnode: holding time classification
• Bitcoin developer documentation: transactions and outputs