Me and my partner are going to split test running some english banners and native language banners on mobile. Trying this because we read somewhere that ads often do better in english even in foreign countries. Is there any merit to that? Anyone have experience doing this?
if the country primarily speaks english like singapore, sure
if the country primarily speaks another language then I'm pretty sure localized banners will perform better. Koreans and Japanese don't like English lol
Yeah depends on the GEO.
If it's just a quick test it might give you some info but if you plan doing volume translating is a MUST.
I run in 30+ countries right now.
What I do with a new country is google "languages of [country]." It'll show you a breakdown of the most spoken languages.
For instance I just did a search for Germany and 95% of the country speaks German. Common sense to just jump into Germany.
Some countries are trickier such as Hong Kong. They speak both Chinese and English. In this case I would split test the entire funnels. One english funnel (ads AND lander in english) vs one Chinese funnel (ads + landing in Chinese).
Yes it costs extra money and time, but getting the right language can be the difference between a dud and a whale.
Damn 30+ countries?! *drools over*
I'm struggling with 3 
I did this initially on some global app installs, only splitting up my traffic with Spanish landers for LATAM and English for the rest of the world.
As I started to gather data I could see some countries just weren't converting. So I made landers in their native language and redirected those visitors over to the new pages. Most geos went into the green after the lander translation, so I'd definitely say it's a good idea to put the extra effort into translating pages, even from the get-go.