|
By Volkan Güvenç, Founder — Alafranga Language Solutions
We have been working with some clients for over a decade. Toyota for seven years. HiRez Studios for four. Solplanet is now in its fourth year. When I look at how we handled their first project versus where we are now, the difference is not just speed. It is the depth of shared knowledge. We know their terminology. We know which reviewers they trust. We know which document types they consider compliance-critical and which ones they are comfortable running through a faster workflow. That knowledge did not come from a system. It came from doing the work, repeatedly, and paying attention. |
▮What actually improves
After every project, something changes. Sometimes it is a glossary entry — a term the client corrected that we had translated one way for three years. Sometimes it is a workflow decision: a file type that kept causing DTP problems, now handled differently from the start. Sometimes it is a personnel decision: this translator works better for this client's register, that one is stronger on the technical side.
None of this is dramatic. It is incremental. But over years, it compounds.
The translation memory for a long-term client is not just a repository of past translations. It is a record of every decision we made together about how their content should read. Every approved segment is a small agreement between our team and theirs.
The glossary is the same. It is not a dictionary. It is a history of corrections, preferences, and domain decisions accumulated over dozens of projects.
▮Where errors actually come from
In my experience, most translation errors in long-term programmes do not come from poor translators. They come from gaps in context.
A new file type arrives. The translator does not know how this document is used operationally, so a term that works in a manual is applied incorrectly in a safety procedure. Or a new translator joins the project without a proper briefing on the client's style guide. Or a glossary has not been updated after a product revision.
These are not linguistic failures. They are information failures. And the only way to address them systematically is to treat every project as a source of information, not just a deliverable.