Address Clustering

💡 The Plain-English Definition

Address clustering is a technique used by blockchain analysts to link multiple Bitcoin addresses to the same owner — building a map of which addresses likely belong to the same wallet, even when no identity information is directly available.

🤔 But Why Though?

Bitcoin transactions are fully public. Every input, every output, every amount, every address — all of it sits on the blockchain (Bitcoin’s permanent public record) forever, visible to anyone. What’s not public is the identity behind each address. Bitcoin is pseudonymous: addresses don’t have names attached to them.

Chain analysis firms figured out that while individual addresses are anonymous, the patterns created when you use multiple addresses together can reveal that they belong to the same person. The most powerful of these patterns is the common input ownership heuristic — if a transaction has multiple inputs (sources of Bitcoin being spent), those inputs were almost certainly controlled by the same person, because creating a transaction requires signing each input with its private key. So if inputs from Address A and Address B appear together in one transaction, they probably share an owner.

From this starting assumption, analysts build clusters — groups of addresses that likely belong to the same wallet. Once even one address in a cluster is linked to a real identity (through a KYC exchange deposit, a forum post, a public donation address), the entire cluster potentially becomes identifiable.

Major chain analysis companies — Chainalysis, Elliptic, CipherTrace — have built sophisticated clustering tools that law enforcement, exchanges, and financial institutions use regularly.

🌍 The Real-World Analogy

Imagine every phone call you make is logged publicly — the number you called, when, and for how long. Nobody knows it’s you making the calls, but the pattern is visible. Now imagine an analyst notices that calls from number A and number B always happen from the same cell tower, often within minutes of each other, and sometimes call the same numbers. The analyst concludes A and B probably belong to the same person. The moment one number is linked to an identity — say, it called a business that requires ID — the other number becomes identifiable too. Bitcoin address clustering works the same way: behavioural patterns, not names.