The earlier search will be on a shorter input which will find more results; imagine waiting for everything that contains "e" before you can get the one result for "europa" that you wanted. Isn't that pattern going to happen most of the time - waiting for a large amount of useless generated and transferred and disposed-of, before the search you want will run?

That's why I set it to run after the third letter and only return up to 7 suggestions. In your example, it wouldn't run until "eur", which should narrow it down quite a bit. Given my rather small blog article corpus and aggressive full text search indexing, it's very fast.