That really depends on how many (or who) you get to help on a specific translation. Also, when someone request a specific language, you have the option to direct them to crowdin to help out.
If time is of the essence, I guess paying is the best bet. But then we won’t see this software localized in our language anyway, so this effort would only make sense for those languages you don’t have time/budget to hire translators for in the first place.
By using something like Crowdin, you will slowly cater for more and more languages. Of course, there might be some languages that go stale once in a while (where we lack translators or proofreaders), but this is probably expected, and if someone comes along and wants to help out in these later, they can always be added at a later time.
On my part, having GlassWire localized to Norwegian will give me two things; 1. Bragging rights because I partook in the translation. 2. The option to recommend this software to anyone who is insecure about english and would avoid it if it were untranslated.
So for Norwegian, time is not of the essence anyway. Besides, community efforts always creates ambassadors that will help out and people generally become more friendly towards the idea that they need to partake to make it happen - if that makes sense to you.
Localization will be a slowpoke in the beginning. But once a language is up to speed with source, it is easier to spot-translate source-strings as they are created by you. (And you can anyway include new translations once they are done, in your regular updates).
In the end, it is up to you how you want to deal with odd-locales, but this solution will at least let us take part in making it happen for those locales that would anyway never be included in the software.
