ETH Zurich has published a study in the Journal of the Royal Society Open Science analyzing 70 years of Eurovision. They examined nearly 1,800 songs, combining musical data, lyrics, AI models, and voting results. The work identifies three phases in the contest's evolution, from initial diversity to current homogenization. The conclusion is clear: the recipes for success no longer work as they used to.
Collective learning that leveled the playing field 🎵
The study reveals that in the formation phase (1958-1974), performances were diverse, with national languages and no defined strategies. During consolidation (until 2003), countries began to imitate each other, adopting catchy melodies, English lyrics, and danceability. These elements became constants, but according to the researchers, collective learning among nations has homogenized the competition. Current rules seek to level the field, but constant evolution ensures the contest does not stagnate.
So the old trick no longer fools anyone 🤷
It turns out that after decades of copying the same formula, countries have reached a point where they all sound the same. Science confirms what many of us suspected: throwing in an English chorus and a dancer in sequins no longer guarantees a single point. Now we just need AI to predict when a song in Finnish will win again. Meanwhile, we'll keep watching them try to be original... all in the same way.