It's working for sub-directories actually, not sub domains.
Each sub directory is called the country code - /uk/, /fr/ etc.
Each country site is in target language and includes things like local office addresses etc.
No noted impact on traffic yet, but we're only on day 1.
Some very noticeable issues with dupe content filters. We have a large dupe content issue which is being worked on, but the wheels turn slowly. Dupe filters on 'pages from' exhibit different behaviour. Either that or it's an issue with what is currently indexed as 'pages from'.
Will try to update as I have more news.