What is a Translation Memory and Why You Need One

A translation memory (TM) is a database that stores previously translated text segments (typically sentences or paragraphs) along with their source language equivalents. When a translator encounters text that matches or resembles a stored segment, the TM automatically suggests the previously approved translation, ensuring consistency and improving efficiency.

TM technology works through exact matching (100% matches) and fuzzy matching (similar but not identical segments). Modern CAT tools display match percentages and allow translators to accept, edit, or reject suggestions. TM leverages are a common pricing model in the translation industry, where clients pay less for repetitions and high-match segments because the translator's work is reduced. A 100% match might be billed at 30-50% of the full rate, while a 75% fuzzy match might be billed at 60-70%.

Building and maintaining a high-quality TM is a strategic investment for professional translators. A TM with over 100,000 segments in your specialization, clean of errors and consistently formatted, becomes a significant business asset. It enables faster delivery times, higher consistency across projects, and the ability to take on larger projects with confidence. TM data is portable through the standard TMX format, meaning you can move your TM between CAT tools without losing your accumulated translation history.