I run a site with mixed-language content, and language switching turned out to be harder than I expected.
The main problem wasn’t translation, it was UX:
dropdowns in the header kept causing layout shifts,
search bars moved around,
and the whole page felt jumpy, especially on mobile.
I ended up changing the approach:
It feels calmer now, but I’m still unsure about the tradeoffs.
For those who’ve dealt with multilingual feeds:
do you prefer instant filtering,
or reload-by-language with shareable URLs?
What broke for you in the long run?
Building on a point raised above about overlays vs dropdowns — stability really matters here. Layout shifts tend to quietly erode trust, especially in feed-based experiences.
On the filter vs reload question, one thing that often gets overlooked is shareability. When content is shared, the language context needs to travel with the link; otherwise people land and bounce before they even engage.
A hybrid approach can work well in practice: instant, session-level filtering for browsing speed, paired with clean, language-aware URLs for sharing and deep links.
Curious — do you have any data yet on how many users actually switch languages vs sticking to one? That usually helps determine whether language controls should be prominent or stay out of the way.
Thanks, that’s a really good point about shareability.
I don’t have solid data yet, traffic is still small, but so far most users seem to stick to one language per session. That’s partly why I’m trying to keep the language UI quiet for now.
The overlay approach sounds right - you're trading one interaction (dropdown) for another (overlay), but gaining stability. Layout shifts are trust killers.
On the filter vs reload question: shareable URLs matter more than most people realize. When someone shares a link to content, it needs to work in their language context. Instant filtering is nice for browsing, but if someone lands on a shared link and sees content in a language they can't read, they'll bounce.
A hybrid might work: filter the current session instantly, but generate a clean URL that reflects the language preference. Best of both - fast UI locally, shareable URLs externally.
What percentage of your users actually switch languages, vs. sticking to one? That data might tell you whether the feature even deserves prominent UI real estate.
Agreed, layout shifts were the main reason I moved away from dropdowns.
I like the hybrid idea a lot: fast session-level filtering, but language-aware URLs for sharing. That’s likely the next direction I’ll experiment with.