VIDEO - Data experiment: Is this what the perfect ESC song sounds like?

The Eurovision Song Contest has been around for almost 70 years. In that time, the ESC has produced some real hits. And a few unexpected winners. But what does it take for a song to win?

In this video, we look at data and studies and try to produce the statistically perfect Eurovision song. And we want to know what a real Eurovision pro thinks of our song. We want to find out: Is there an ultimate Eurovision formula for success?

Let's start with the question: How fast does a song have to be to have a good chance at the Eurovision Song Contest? To find out, we'll analyze all the Eurovision Song Contest songs that have made it to the final since 2009. We chose this period because the judging system was adjusted in 2009. Since then, the jury's opinion and the audience's opinion each count half. If you sort these 388 songs by tempo and ranking, you can see that tempo has hardly any influence on how a song performs. There are fast songs that perform well, like the Finnish entry from 2023, which ultimately finished in second place. But there are also slow songs that are successful, like this entry from Austria, which reached third place. So, tempo alone doesn't seem to be that crucial.

Instead, let's try sorting the songs by style. We looked at all the entries that made the top 5 since 2009 and divided the songs into three categories. First: Mainstream Pop. That is, danceable pop music in 4/4 time, with an electronic sound and a catchy chorus.

A good example of this is "Heroes" – the Swedish winner from 2015. Entries in this category are often combined with elements of traditional music – creating a kind of ethno-pop. For example, here with the Ukrainian entry from 2021.

Then there are the ballads. They begin quietly and often end powerfully. Many of these songs have orchestral elements. For example, "Rise Like a Phoenix" from 2014.

And then there are the entries that are a bit out of the ordinary. Songs that are rather unconventional. Like "The Code" by Nemo, with its mix of pop, rap, opera, and drum 'n' bass. Especially in recent years, these unconventional songs have made it into the top 5 more often.

But overall, regular pop entries dominate, accounting for about half of the songs in the top 5.

Another integral part of the Eurovision Song Contest are fun entries. Entries where entertainment is the main focus – rather than musical quality. But if you look at the data since 2009, you'll see that only one such fun entry has made it into the top 5 since then. That was the singing grandmas from Russia in 2012.

What does this mean for our own song? The tempo isn't that important. But with a ballad or an unconventional song, we're pretty out of luck. And a fun song is out of the question, too. A danceable song with electronic sounds would be best.

Danceable and electronic, to me that sounds like a tempo somewhere between 120 and 130 beats per minute. We've chosen 130 BPM for our example because we want to incorporate the electronic subgenre garage, or UK garage to be more precise, into our pop-style song. The Eurovision Song Contest often features influences from current electronic music. For example, techno or psytrance, or in our case, UK garage.

So it's quite simple, but still danceable and it combines very well with pop elements and is perhaps a small factor that could set us apart from other songs.

With this beat, we already have a basis for our song. But what do Eurovision producers really look for? One person who knows this is Benji Alasu. He is one of the songwriters and producers of "The Code." This is the song with which Nemo won Eurovision in 2024.

Benji Alasu: "A producer once told me: 'Yeah, Eurovision songs, it's always about impact.' So, impacts are super important, and that just stuck with me. And by impacts, he simply meant the transitions, the way it hits the mark and then fades away again, and things like that. That really stuck in the back of my mind. And I tried to take that with me."

Okay, we'll try to incorporate that into our song as well. But first, let's look at what the data says about keys and harmonies.

If you look at all 388 final songs, you'll see that the majority are composed in a major key. And major is more likely to be associated with qualities like "bright" and "cheerful." But if you just look at the songs that ultimately made it into the top 5, things look completely different. Minor keys dominate. And minor keys sound rather serious, perhaps even a little sad.

One key stands out particularly often: D minor. About one in six songs in the top 5 was written in D minor. This includes the winning songs "Fairytale" and "Rise Like a Phoenix."

And the data reveals something else surprising: Until the mid-2010s, about one in three entries in the final featured a so-called modulation. That is, a sudden key change – usually as a major, emotional climax. But since 2016, such modulations have become significantly less common. The dramatic key change that many associate with the Eurovision Song Contest has thus become quite out of fashion. The last Eurovision Song Contest victory with modulation was more than ten years ago.

Therefore, the case for our song is clear: Statistically speaking, a song in D minor has the best chances – and please, without a key change.

The key actually gives us a framework into which we can produce the song. And since we're using D minor here, based on the data, we set it up accordingly in the music software and then choose chord progressions that fit.

What you often find in Eurovision songs are these very orchestral elements, which are then combined with pop and electronic music. And in this case, we chose a cello, which sets a very good tone for the mood. So it's very simple, but effective, and immediately creates a mood.

Then, of course, it's about adding more elements to the cello. These are pads, like this one, which simply creates a kind of atmosphere. In the verse, we also use a guitar. In the pre-chorus, a few more instruments are added, which are essentially meant to make the atmosphere bigger.

Another genre that's trending right now is so-called stutter house. We incorporated it with a choir and a synth.

With “The Code,” too, care was taken to incorporate trendy elements.

Benji Alasu: "We actually wanted to make a song that would catch the audience and be a bit TikTok-like and crazy. And we knew that if we could combine a lot of things that harmonize well together, it could work. And we knew that Nemo is so talented that we have almost no limitations."

What's missing for our song now is the vocals. And for that, we first need lyrics. For a long time, Eurovision entries had to be sung in the national language of each country. But since 1999, countries have been free to choose. And one language has clearly prevailed: English.

To find out what the lyrics are about, we analyzed all 388 songs using various large language models. We tested different models and prompts – and here's what we found:

Firstly, the lyrics often deal with strong emotions such as love, longing or heartbreak.

Second, many songs are about fighting back or being strong. They're often about self-determination, courage, or breaking chains.

Third, simple imagery is used to express emotional processes. For example, fire, storm, or light.

Fourth, many choruses have simple, repetitive speech patterns—often complemented by sounds like "uh-oh-oh." This makes them memorable.

For simplicity, we're building the lyrics with ChatGPT. We've now packaged the data insights into a prompt, and ChatGPT has crafted the lyrics from it. The verse begins, "They told me I should smile less wide, hide the fire I hold inside."

So you can see that this text is about not letting yourself be held back by others and about being yourself and I would say that the text is okay, but it is really not a masterpiece.

Then there's the question of who should sing the song. Here, too, the data is quite clear: Women perform best at the Eurovision Song Contest. Around 46 percent of the top five entries were sung solo by a woman. This means for our song: We now need a female singer for our lyrics.

We then created a melody for this text and passed it on to a singer. The singer then returned this file to us. Combined with the instrumental, the whole thing sounds like this.

Benji Alasu: "The first reaction is nice. I think the vibe is already nailed. The vibe, in the sense that you can hear it and you can already see the stage setting, is more ESC, I think. What's still missing to make the song more ESC-like is if I went up a half tone for the singer, maybe, and then I looked at the melody. I can't sing it back to you right now. I'd work on that."

What I've built here, single-handedly, in this short time, in this experiment obviously can't really compete with ESC production. But it still beautifully demonstrates that all the small nuances, those artistic decisions in the details, which are nevertheless very, very important, are simply missing from the data.

Benji Alasu: "Why does this melody do something to me and the others don't? That's always the question. Why does this chord sequence work better than the other? These are all factors, but they can't be captured data-based. I think the start is a bit data-based, in the sense of, okay, we want to make a ballad or we want to make something more powerful. But after that comes the creative part, after that comes how you feel and what the song wants, I always say."

In the end, production is the work of professionals, of artists who have to master their craft more than just very well in order to win a competition like the ESC.