On demand streaming solved the problem of finding a song whose name you already know. Radio has never solved the opposite problem, which is far more common: a song plays, it is good, and by the time you can reach a phone the presenter has moved on and the title is gone. This is the one functional advantage on demand services have always held over broadcast, and it is solvable.
Track Radar removes the need to react in time. While a station plays, each track it announces is recorded along with the station it came from and the time it aired. Nothing has to be tapped at the moment the song is on. Hours or days later the list is still there, in order, and the song that could not be named at the time is sitting at the timestamp where it played. The mechanism underneath is the stream's own metadata, described in more detail in our guide to station directories, which means accuracy depends on how carefully the station writes its now playing field. Stations that publish a clean artist and title pair produce clean logs. Stations that publish only a slogan produce a log entry that is honest about knowing nothing.
Saving a track moves it from the passive log into a library that persists across devices when signed in. From there the practical step is usually leaving the app entirely, and that is deliberate. Each saved track carries links out to Spotify, YouTube, and SoundCloud with the artist and title already filled in, because the natural next action after finding a song on the radio is adding it to a playlist somewhere else. The free tier keeps three saved tracks at a time, which is enough to find out whether the habit is useful. Premium removes the limit.
The statistics dashboard is a slower reward and works on accumulated history rather than any single session. It reports listening streaks, the hours of day when listening actually happens rather than when it is assumed to happen, and the spread of countries a listener's stations broadcast from. That last figure tends to surprise people, because the perception of listening widely and the reality of returning to four stations are usually different. The dashboard also derives a listening persona from the mix of genres and hours, which is a lighter feature than the rest and is meant to be read that way.
A routine that gets value out of these three pieces without much effort looks roughly like this. Leave a station running while working and do not think about the log at all. Once a week, open the radar, scan the timestamps around the hours when something stood out, and save the handful worth keeping. Move those out to whichever service holds your playlists. The statistics take care of themselves and are worth reading monthly rather than daily, since streak and footprint figures only become interesting once several weeks of history exist.
One limitation is worth stating plainly, because it changes what the log can contain. Guest listening history is held locally in the browser and never reaches an account, which means it does not follow you to another device and is lost if site data is cleared. Signing in is what makes the library and the statistics durable.
References:
Xiph.Org Foundation. Icecast 2 Documentation: Server Statistics and Metadata.
Berry, R. (2006). Will the iPod kill the radio star? Profiling podcasting as radio. Convergence, 12(2), 143-162.
Keith, M. C. (2010). The Radio Station: Broadcast, Satellite and Internet (8th ed.). Focal Press.