Scavenger AI systematically analyzed 68 years of the Eurovision Song Contest: 1,734 entries, 54 countries, 51,354 voting lines, and 1,722 song lyrics were analyzed. The data show which lyrics resonate best, who is always right in their predictions, and whether men or women win more often.
The fan community is always right
Between 2016 and 2024, the MyESC community (an app in which Eurovision enthusiasts from all over Europe aggregate their predictions) placed all eight winners in 7th place or better. The OGAE fan club vote, by contrast, regularly overestimated the wrong contenders: France 2023 with 302 points ended up in 16th place, Moldova 2017 came third with zero points. In total, the fan club placed only five of the last eight winners in their top 7. This also shows that more data sets lead to a reliable result; after all, the MyESC scoreboard has more than 10,000 users, while OGAE has just under 900 members.
Things get interesting from starting position 11
Whoever performs first loses. In eight consecutive Grand Finals (2016 - 2024), not a single country that drew one of the first five starting positions won. With an average total score of 119 points, this part of the show is by far the weakest segment. What comes next is more surprising: it is not just the end of the show that dominates, but two specific windows. Slots 11 to 15 and 21 to 26 together produced seven of the eight winners — with almost identical overall averages of 213 and 216 points. Between them lies a noticeable dip: slots 16 to 20 drop back to 175 points and have no wins at all.
Language as Strategy
When the requirement to sing in one's own national language was lifted in 1999, the share of English-language songs rose sharply from 10 percent to just under two thirds of all songs. In recent years, a countertrend can be observed - and the trigger for it was Portugal. In 2017, Portuguese singer Salvador Sobral scored 758 points with the song „Amar Pelos Dois" - giving Portugal its first victory in 53 years and the highest score ever achieved at the ESC to this day. The winning song was raw, melancholic and above all one thing: in Portuguese. Since then, the number of songs sung in the national language has been steadily increasing: from 16 percent the year before (2016) to 37.5 percent in 2018 to almost half of all songs in 2024 being non-English-language songs. And this is also reflected in the winners: 4 out of 8 winning songs between 2016 and 2024 were songs in the respective national language.
No One Wins with “La La La”
If you want to win, you sing about eternal love - and under no circumstances “La La La”. At least if you go by the statistics: the word that has appeared most often in winning songs since 1956 is “forever” (14.5 percent), while “La La La” appears in none of them - but in 6.9 percent of all last-place songs. On average, those are also more negative in their lyrics, while just under two thirds (63.3 percent) of all winning songs have a positive undertone. Incidentally, the most sung word on the ESC stage is “Love”, which appears in 27.4 percent of songs and thus in almost every third song. Paradoxically, Europe still prefers to sing about war rather than love. For while “war” appears in 129 songs (7.5 percent of all songs), only 22 songs (1.3 percent) mention “peace”.
Placements: solo artists before solo female artists before groups
Men win more often, but only slightly. Between 2016 and 2024, four of eight victories went to male solo artists, three to women, and one to a mixed group. On average, men place in 12th.6, women in 13th.8, which is a difference of just over one place. Both groups achieved exactly nine top-5 placements in this period.
The difference between solo artists and groups is more pronounced: solo artists qualify for the final 66% of the time, groups only 60%. While six of eight victories went to individuals, mixed groups perform the weakest: lowest score, worst average placement, only one win. The field itself is perfectly balanced by gender, with 139 male and 139 female entries. The gender difference exists in the dataset, but it is too small to serve as a strategy. However, the fact that solo artists do better than mixed groups can be a deciding factor.
The data show what has repeated over decades: a solo female artist or a solo male artist, starting position between 11 and 15 or 21 and 26, a song in the native language, lyrics about love. Statistically speaking, that would be the strongest starting position for a win at ESC 2026. But at ESC, numbers are not the only thing that counts - in the end, it is above all the three minutes on stage, the music, and the moment that decide.
You can easily carry out more ESC analyses yourself: https://app.scavenger-ai.com/