Most corpus-based studies of translation use published texts as the basis for their corpus. This overlooks interventions by other agents involved in translation such as editors, who may have significant influence on the translated text. In order to study editors’ influence on the translation product, this paper presents a comparative analysis of manuscript and published translations, which allows a differentiation of actual translated language and edited translated language. Based on a tripartite parallel corpus of English business articles and their translations into German, I analyse translators’ and editors’ influence on grammatical metaphoricity of the text, specifically on the use of nominalisations. One finding is that a significant amount of nominalisation is re-verbalised by editors. The results show that translated language may often be the result of significant editorial intervention. Thus, by just considering source text and published translation, our picture of what translators actually do may be significantly distorted.