Deterministic Matching uses exact correspondence between instances of an identifier or set of identifiers to produce a match. Deterministic matching assumes that unique identifiers such as email addresses or IP addresses can be used to definitively link identity fragments across channels or interactions. When making a deterministic match, you are using some known identifier from first-part data, which can improve match reliability. But deterministic matching can potentially err in several ways: missing data can mean a real match is missed; a shared identifier can mean multiple identities are treated as one; mis-entered data or multiple identifiers owned by a person (e.g. multiple email addresses or telephone numbers) can lead to incorrect splits or failed merges. Using a combination of deterministic, probabilistic, and heuristic approaches can reduce errors at the expense of additional work in defining or performing matches.