Translation memories are databases that retain previously translated sentences, paragraphs, and other forms of text. Each item (segment) in a translation memory includes texts in both the source language and its translated target language. A pair consisting of such is called a translation unit. Generally referred to as TMs, translation memories are used in our computer-assisted translation (CAT) tools and other professional translation software and automatically suggest identical or similar pre-existing segments to our linguists as they translate a new document. Using TMs eliminates the need to translate previously processed sentences or paragraphs again. TMs support all of our translation and localization processes and is an indispensable piece of technology that significantly improves the quality, speed, consistency, and efficiency of all translation processes.