Payment Match Score

Overview

What does Payment Match Score mean?

Payment Match Score is a numerical or categorical result produced by evaluating a transaction against payment matching evidence and rules.

How it works in practice

A payment match score estimates how strongly an observed transfer belongs to an invoice when exact deterministic evidence is incomplete. The engine combines features such as address match, reference, allowed asset, amount difference, time window and known customer attribution. Some signals are hard constraints: an unsupported token contract should not be compensated by a high time-match score. Scoring thresholds should be calibrated on real exceptions, and every automated assignment should retain the feature values and rule version that justified it.

Limits and common misunderstandings

A high score is not a statistical guarantee of ownership unless the model is validated and calibrated. Similar invoice amounts, reused addresses and correlated signals can create falsely confident matches. Cases near the decision threshold deserve manual review.

Additional implementation considerations

Thresholds should be chosen from actual labeled cases and reviewed when merchant behavior changes. A score of 0.92 means nothing unless its calibration and components are defined. Keep strict rejection rules for incorrect asset identity and chain, and use scoring only for genuine uncertainty in attribution. Reviewers should see why a candidate ranked first, what alternate candidates existed and which evidence would remove ambiguity.

Related glossary terms

Payment Match Confidence; Payment Match Candidate; Payment Match Rule