We have emphasised the benefits and savings associated with building a Translation Memory (TM) for a particular customer, and the way that the TM gets better and better the more documents are processed using it.
But a TM can also be built ‘retroactively’ – that is, using documents that have already been translated by someone else. When an LSP starts working with a customer who has already had documents translated, either in-house or by a different LSP or translator, these documents can simply be fed into TM software (‘aligned‘) alongside their translated versions, and a ready-made TM created.
CAT software usually features program-loading optimisation, which allows you to ‘align’ the two documents (the source and the translation). The texts should be as close to identical as possible in terms of formatting, and all the translator has to do is go through and ‘populate’ the new TM based on the correspondences between the two texts. If this process is repeated with various documents, a functional and detailed TM can be build for the new client, enabling them to take advantage of all the savings associated with TM from the beginning of their partnership with their new LSP.