Match rate is the percentage of source content in a project that can be retrieved from a translation memory at a given similarity threshold, such as exact, fuzzy, or no match, after analysis against existing translation units. It is used in translation pricing and planning to estimate how much new translation is required versus how much existing approved content can be reused or edited.
Match rate appears in translation memory analysis reports generated by CAT tools before a project is quoted or assigned. Project managers use it to apply word-count discounts, set fuzzy-match thresholds, and compare output from different engines or TMs when deciding between machine translation post-editing and human translation.
A common mistake is treating a high fuzzy-match rate as if it means most content is ready to use, when in reality each fuzzy segment may require extensive editing and QA. Conversely, ignoring match rate entirely can cause clients to pay full per-word rates for repeated or previously translated content that a TM could have leveraged.
For creative and consumer content the failure mode is flatness, not inaccuracy — and flatness passes an accuracy check. Terms here also carry technical constraints (reading speed, character budgets) that are part of the brief.
Angel Translation locks terms like this into a project terminology base before expert review begins, so the same source term resolves to the same target term across every document in a submission — and stays consistent in the next revision. See the 8-step AI + expert workflow.
Send us the files. Acknowledged in 4 hours, quoted in 24.
Request a Quote →Content Globalization solutions →