Blockchain analytics tools are essential for compliance teams, regulators, and investigators to uncover illicit activity and hold bad actors accountable. However, their effectiveness depends entirely on the quality of underlying data. Incorrect data can waste resources, undermine investigations, and lead to wrongful actions like customer terminations. To select the right provider, users must ask critical questions about address grouping, entity labeling, and methodology testing. Address grouping methods vary between probabilistic and deterministic approaches, and providers must handle edge cases like CoinJoin transactions. Entity labels require reliable evidence, distinguishing between wallet controllers and users. Providers should have their methods tested in court (e.g., Daubert standard) and through independent accuracy studies. They should also clarify the role of machine learning and provide transparency in cluster construction. Chainalysis offers a formal ontology called 'Defining the Cluster' to standardize these processes.

Key facts

• Address grouping methods must handle edge cases like CoinJoin transactions.
• Entity labels must be supported by reliable evidence, distinguishing controllers from users.
• Methodology should be tested in court and via independent accuracy studies.
• Machine learning outputs must be clearly labeled and not treated as confirmed facts.
• Providers must explain how any given cluster was built and what evidence supports it.