Web search is challenging partly due to the fact that search queries and Web documents use different language styles and vocabularies. In order to bridge the lexical gap between queries and documents, we are leveraging well-known models developed for bilingual translation to solve our monolingual problem. We assume queries to be our source language, while documents aligned thanks to user click-feedback to be our target language. From the resulting aligned corpus, several variants of Translation Models have been generated and experimented as ranking features in the live search engine with certain success. We will also talk about how a close collaboration between Microsoft Research and Bing enabled a research project to have direct product impact, along with the challenges we had to overcome in order to ship on a live system.