Your website has a design. Your studio has a look. But the thing most listeners actually see is one line of text in whatever player they happen to use — and it comes from your stream metadata.
The template
Encoder profiles carry a metadata format template, defaulting to {artist} - {title}. That is what appears in the listener's player.
The default is right for music. It matches what people expect and it is what music recognition and logging services parse.
Station identity
Profiles also carry station name, genre, website URL and description. These appear in directory listings and in some players' station information.
The website URL is the one most often left blank and most worth filling. It is a direct path from someone listening to someone visiting — and directory listings are one of the few places new listeners find stations at all.
Where it comes from
Metadata is generated from the track currently playing through the engine, which means it follows whatever is actually on air rather than a separate list that can drift out of sync.
Because the encoder taps post-mix audio and metadata comes from the same playback state, what listeners hear and what they read stay aligned.
Tags are the input
All of which depends on your files being tagged. A library full of track07.mp3 with no artist tag produces a metadata line that says nothing.
Tags are read with TagLib#, falling back to the filename where they are missing. That fallback keeps things working but it is not what you want listeners reading.
Right-click → Information… shows exactly what is on a file, including the fields that are empty. Fixing tags is unglamorous work with a directly visible payoff.
What about streams and imaging
When a relay stream is playing, the metadata reflects the playlist entry — which is why giving stream URLs a sensible display name when you add them matters. "Network News" reads considerably better than a URL.