What is Wallet Clustering?

Wallet clustering groups multiple wallet addresses that available evidence suggests belong to the same user or entity. It can reduce double counting and connect fragmented journeys, but it is not proof of real-world identity. Responsible clustering records the evidence, confidence, time, privacy constraints, and reasons a relationship may be uncertain.

How wallet clustering works

Strong signals include explicit account links, verified ownership, or wallets connected within an authenticated product context. Weaker behavioral or transaction patterns may support probabilistic relationships but should carry lower confidence. Embedded wallets, smart accounts, exchange deposit addresses, multisigs, and shared devices can complicate the one-wallet-one-user assumption.

Wallet clustering example

A user signs in to one product account and connects an EOA, an embedded wallet, and a smart account. The three addresses can be represented as one analytical cluster while each address retains its own transactions, balances, source, and confidence record. Reports can then switch between address-level and cluster-level counts.

Common mistakes

Do not treat every transfer between addresses as common ownership or present a probabilistic cluster as a verified person. Avoid irreversible merges, hidden confidence thresholds, and analysis that discards the original addresses. Clustering can introduce false positives, false negatives, bias, and privacy risk.

Using Wallet Clustering in practice

Use clustering when the analytical question requires user-level journeys, deduplicated activity, or related-wallet context. Keep address-level reporting available, document the method, and allow uncertain links to be reviewed or reversed. Apply the same clustering version when comparing periods or cohorts.

Related reading

Explore wallet profiles, Identity Resolution, Wallet Profiles, Unique Active Wallets, and the wallet and web analytics guide.

FAQs

Is wallet clustering proof that addresses belong to one person?

No. A cluster is an analytical relationship supported by stated evidence and confidence, not automatic proof of real-world identity.

How do smart accounts affect wallet clustering?

One user may interact through owners, signers, session keys, or account contracts, so the cluster model must preserve those different roles.

Can wallet clusters change over time?

Yes. New evidence can add, remove, or revise links, so reports should record the clustering method and evaluation time.