What’s the best way to implement a language switcher — Part 1 (Desktop Web)
A language switcher is either one of the first things you look for, or the very last when using digital products. Due to their tricky frequency of use, they are either extremely hidden, and automated by some form of logic, or very easily accessible occupying, usually important screen real estate.
Since Arabs have had to deal with English websites more than Arabic, it became a natural habit to look for the switcher in certain locations based on what most websites, and apps implemented overtime based on what they thought made sense.
More importantly, it also became a standard expectation to have a language switcher, and this directly ties into the previous post “Not all Arab users are the same”. This is where automated solutions on websites can fail and be way more irritating than helpful, when Arab users, or even worse, non-Arab expats are pre-assumed to want an Arabic version. Specially if the language switcher is hidden deep away. Google used to serve people the Arabic version of their search results, and a language switcher that was no-where to be found. This was eventually fixed by adding an obvious and clear sidebar of the language switcher in the results page, although this has been an inconsistent experience of mine as sometimes it would just be the standard search results page, with the switcher either at the homepage, or hidden deep in the settings of the search results page. Likely was part of an A/B test.
So let’s look at some common examples of language switcher implementations:
Top nav bar
It’s clear that most Arab-locale websites highlight the language switcher in an easy to access location, most commonly on the top left or top right. This has become a standard expectation. While the big international players opt for a footer implementation of the language switcher. From my experience, Arab users are used to look for the switcher at the nav area first, if it’s not there, they know to look at the footer area.
Some websites implement automated methods to localization detection, but these methods are tricky and should be handled with care, as they could be served to the wrong audience incorrectly as stated in the article “Not all arabs are the same”.
The best implementation is likely to be a combination of multiple forms of detection at different levels of priorities, and geographical location should not be the ultimate deciding factor.