Eurovision 2025: Semi-Final 2 Results Analysis

The second semi-final of the Eurovision Song Contest 2025 saw sixteen countries become ten in a battle to qualify for the final. Just like the first semi-final, the eventual result was very close, and also threw up a few surprises, so once again I think it’s time to dive into the numbers and find out what the scores on the doors really mean for these participating countries. Before we do, however, I would recommend you read my previous article, which was an analysis of the results of the first semi-final, if you haven’t already.

For the avoidance of confusion, I will once again clarify that the results of this semi-final were determined solely by televoting in each participating country, as well as France, Germany and the United Kingdom, voting as pre-determined automatic qualifiers, and an aggregated ‘Rest of the World’ vote. I’ve shared before on this blog that I don’t agree with this system, and would definitely favour the re-introduction of juries, but this is where we are at the moment. Anyway, let’s see what the big picture has to tell us.

Israel came out victorious in this semi-final, scoring 203 points (10.68 points per voting country), followed by Latvia, Finland, Greece, Austria, Lithuania, Luxembourg, Denmark, Malta and Armenia, who finished in tenth place, qualifying for the final with 51 points (2.68 points per voting country). The six countries failing to make it to the final were therefore Australia, Czechia, Ireland, Serbia, Georgia and Montenegro, who finished at the bottom of the pile with just 12 points (0.63 points per voting country).

This was Israel’s second consecutive semi-final victory, and their score is the highest semi-final score since the move back to a televote-only competition in 2023 (as well as being the only entry to score more than 200 points in a semi-final since then). This seems, once again, to be a result of the Israeli government’s relentless advertising campaign for the song, which no doubt provided a boost to pre-existing support for the country. I spoke about this in my overall Eurovision 2025 analysis, but clearly the thirteen sets of 12 points Israel received were not all entirely legitimate, and surely warrants some scrutiny of the televoting system. For the third year in a row, Israel received maximum marks from the Rest of the World vote, which has never failed to award the country douze points.

Israel received points from every voting country in this semi-final, and in fact only two countries ranked “New Day Will Rise” outside of the top three: Serbia, who awarded Israel 7 points, and somewhat surprisingly, Yuval Raphael only received 2 points from Armenia, despite having been placed in the same semi-final allocation pot. This result marks Israel’s third consecutive qualification to the final (for what it’s worth, I do think they would have qualified regardless of any external interference) and gives the country a solid qualification rate of 63% overall.

Second place went to Latvia, who scored 130 points, a marked decrease from Israel’s score. As with all three shows this year, the votes in this semi-final were much more widely spread than we usually see, and as such the scores are notably lower and also much closer, particularly at the top. I think this song had a similar journey to Iceland in the first semi-final in that it was significantly helped on its way by Eurovision fans nervous for its qualification chances. Whilst I didn’t personally vote for this song, I did wonder if it might struggle to reach the final and was very pleased when it got there.

Despite receiving points from every voting country, Latvia only received one set of 12 points, from neighbouring Lithuania, although did manage to make it in to the top three with another eight, including Nordic nations Denmark and Finland, diaspora-heavy UK and Ireland, and 10 points from the Rest of the World. This result is Latvia’s joint-highest semi-final finish ever, matching Aminata’s feat from 2015, and also gives the country its first qualification streak since 2016. However, this may not be good news for Latvia, who since 2009 have followed a pattern of six years out of the final followed by two qualifications. I don’t want to jinx anything, but I sincerely hope we see Latvia in the final again before 2032.

Latvia's Tautumeitas performing in the second semi-final of Eurovision 2025. (photo: Corrine Cumming/EBU)

Finland’s Erika Vikman took the bronze medal with 115 points, the last contestant to receive points from every voting country. I know that this came as a surprise to many in the Eurovision community, as Erika was definitely a pre-contest fan favourite, and had a very favourable running order position, so there was a fairly sizeable group of people expecting this to win the semi-final with relative ease. Again, I discussed Finland’s perceived under-performance in my overall Eurovision 2025 analysis, but I think the package was harmed by the muting of the live audience in Basel.

Despite the high position, Finland failed to receive any sets of 12 points, with their highest scores being 10 from Australia, Denmark and Malta, and 8 from Latvia and Serbia. The Rest of the World vote awarded Finland 6 points. This does of course extend Finland’s qualification streak, making them one of the most successful countries in Eurovision this decade, having qualified every year since 2021. Finland now has an overall qualification rate of 60%; though this only looks set to increase as the national final continues to go from strength to strength.

Just behind Finland in fourth place was Klavdia from Greece, who finished with 112 points. Having had this song down as a borderline qualifier, I was thrilled to see “Asteromáta” qualify and do as well as it did in the final. This result clearly shows that there is an appetite among a casual audience for Greek-language music and this very distinctive sound. No doubt the Greek diaspora also contributed to some of the higher marks Greece received, particularly the 12 points from Armenia and 7 from the Rest of the World.

In total, Greece received points from seventeen countries out of a possible nineteen, only failing to score from Latvia and Lithuania in what I might describe as a “reverse KAJ effect”, whereby this song connected much more with audiences in the South and East of the continent, scoring heavily from the Balkan countries present in this semi-final. Seven countries placed Greece in the top three, including Malta, Montenegro and Serbia, but also countries such as Germany and Luxembourg where diaspora is likely to have had more of an influence.

Klavdia from Greece finished fourth in the second semi-final of Eurovision 2025. (photo: Alma Bengtsson/EBU)

Next was the eventual winner of Eurovision 2025, Austria’s JJ with “Wasted Love”. Despite finishing fourth in the final televote, he could only manage one place further down two days earlier, coming in at fifth place with 104 points. This marks the third consecutive year in which the eventual winner has not won their semi-final (which makes sense given that none of these winners have won the televote in the final), and is also the lowest semi-final finishing position ever for a song that proved victorious in the end, breaking Nemo’s record from last year.

Like Greece, Austria scored points from seventeen countries, with the UK and the Rest of the World declining to reward this song. JJ didn’t receive any maximum marks, and only three countries ranked him in their top three, but he scored well from most countries, with “Wasted Love” appealing to voters across the continent and twelve countries awarding Austria at least 6 points. Austria have now qualified for the final for three consecutive years – one away from equalling their longest ever streak – and have an overall qualification rate of 53%.

Sixth place went to Lithuania, who finished only one point behind Austria in another example of the very tight results we saw this year. Katarsis scored 103 points and received votes from seventeen countries, including 12 points from friendly neighbours in Latvia and another five top three rankings. As is often the case for Lithuania, this song was helped heavily by the diaspora, and the country had something of an ideal draw, with Ireland and the UK also present in this semi-final. Lithuania received 5 points from the Rest of the World vote.

Lithuania is another country that must be considered a powerhouse of 2020s Eurovision. Like Finland, they have qualified every year since 2021 whilst sending a real variety of songs in the process. Whilst the all-televote competition will have helped them somewhat, clearly this is a country that knows what it’s doing. The country now has a qualification rate of 65% since 2004, although with “Tavo akys” qualifying this year, the fact remains that Lithuania has never failed to qualify for the final with a song containing lyrics in the Lithuanian language.

Luxembourg finished seventh in this semi-final, with Laura Thorn’s “La poupée monte le son” scoring 62 points. This represents a clear separation between the top six entries, all of whom scored points from at least seventeen countries, and the chasing pack. This was Luxembourg’s second appearance at Eurovision since their return last year, and given that they once again managed to qualify for the final, they have maintained a 100% qualification rate. I don’t think it’s quite on the same level as that of Ukraine, which has qualified from fifteen semi-finals, but worth celebrating all the same.

Luxembourg received points from fourteen out of nineteen countries, but its only top three placings came from France and Israel, both of whom awarded Laura Thorn 10 points. Israel voted heavily for Luxembourg last year, largely due to the background of the singer, but I’m not sure quite what it was that drove this push once again. Beyond those two countries, Luxembourg won widespread low-level support and scored in the 1-4 points range with eight countries. However, the country failed to score anything from the Rest of the World vote, which is possibly down to a smaller diaspora.

Luxembourg qualified for the final of Eurovision 2025. (photo: Sarah Louise Bennett/EBU)

Hot on the heels of Laura Thorn was Sissal from Denmark, who finished in eighth place with 61 points. This was Denmark’s first appearance in a Eurovision final since 2019, and, despite my earlier reservations about their chances, they qualified with relative ease. Despite finishing lower than Luxembourg on the scoreboard, Denmark actually scored points from more countries, with sixteen out of nineteen awarding “Hallucination” a place in the top ten. The country’s only top three position, however, was 8 points from Finland.

This entry is another of this year’s songs that was decidedly more successful in one part of the continent than the rest. Much like KAJ from Sweden, and the opposite of Klavdia from Greece, Denmark scored many more points from the Northern European countries in this semi-final, and failed to score at all from Georgia, Greece and Serbia. However, Denmark is also another country that managed to score consistently in most countries, with eleven countries awarding a score between 1 and 4 points, one of which was the Rest of the World vote, which gave Sissal 4 points.

Ninth place went to fellow diva Miriana Conte from Malta, who, much like Erika Vikman, could not live up to the expectations of fans. Whilst it must be recognised that “Serving” did bring Malta to the final for the first time since 2021, and that this is a huge achievement in itself, the televote scores for the song both in the semi-final and the final left much to be desired. I discussed a couple of possible reasons for this in my overall Eurovision 2025 analysis, but I am of the opinion that it was likely Eurovision fans providing a significant percentage of Malta’s votes in this semi-final.

Malta scored 53 points, and received votes from fifteen countries. The country’s highest score, and only top three placing, came from Australia (8 points), with Armenia awarding 7 and Greece 6. Other than that, Malta scored in the range of 1-4 points with eleven countries, including 3 from the Rest of the World. Fundamentally, Malta will always be at a disadvantage in the televote due to its size, and virtually non-existent diaspora. I am very glad that Miriana qualified this year, as the song and package deserved to be in the final, but I worry that the country might not be able to replicate this success whilst the voting system remains in its current form.

The last qualifier from this semi-final was Armenia, who finished just behind Malta on 51 points. I have said on numerous occasions that the real strength of this song was its uniqueness in its field, and much of its success can probably be attributed to there being nothing else like it in either this semi-final or the final. “Survivor” was another entry that I didn’t think was going to make it, but with this result, Armenia have now qualified for the final for four years in a row, and are just one year away from equalling their longest qualification streak (achieved from 2013-2017).

Despite finishing with almost the same score as Malta, Armenia only received points from nine countries out of a possible nineteen, with most of Western Europe, as well as the Rest of the World, ignoring this song altogether. However, Armenia was the only country other than Israel to receive more than one set of 12 points in this semi-final. Parg scored douze points from both Georgia and Israel themselves, reaping the benefits of a favourable semi-final draw that kept them away from rivals Azerbaijan. As well as this, Armenia received 8 points from France and Greece, both countries with a significant Armenian diaspora.

Go-Jo from Australia failed to qualify for the final of Eurovision 2025. (photo: Alma Bengtsson/EBU)

The unlucky country that just missed out on a place in the final was Australia, who finished eleventh in the semi-final for the second year in a row. Whilst I didn’t like the song, I was still surprised to see “Milkshake Man” bow out early. The song scored 41 points, ten behind Armenia, but actually scored points from twelve countries compared to the nine that voted for Parg. However, in what was possibly an adverse effect of opening the show, the highest score awarded to Australia by any country was 6 points from both Latvia and Lithuania. Go-Jo also scored 2 points from the Rest of the World, a vote less impacted by the running order on the night. Australia’s qualification rate now stands at 67%, a record low for the country, so hopefully they will come back stronger in 2026.

Twelfth place went to Adonxs from Czechia, another country that I, along with most of the Eurovision community, had expected to perform much better. In the end, “Kiss Kiss Goodbye” didn’t come particularly close to qualifying, only scoring 29 points. Eight of these came from the Rest of the World vote, in a similar result to Sudden Lights from Latvia in 2023. It has been suggested that this is the result of people in Slovakia voting for one of their own. However, if we remove the Rest of the World vote from our calculations, Czechia actually ends up fifteenth – i.e. second to last – a hugely disappointing result for the country, which has now failed to qualify for two years in a row.

Only one point behind Czechia, we have a three-way tie on 28 points. Winning the tie-break by virtue of receiving points from more countries is Ireland, who failed to replicate last year’s success with Emmy’s “Laika Party”. The song received points from nine countries, including 6 and 7 from fellow anglophone nations Malta and the UK respectively, and were the lowest-ranked country overall to score from the Rest of the World vote, receiving 1 point. Unfortunately, this seems to be a bit of a return to form for Ireland, who have followed a pattern since 2014 of four non-qualifications followed by an appearance in the final. If this continues, we can look forward to Ireland next qualifying in 2029, but here’s hoping the country can pull it together before then.

In fourteenth place, and also on 28 points, was Serbia. Despite my personal feelings about the song, which I will still defend with my life, I must accept that the prevailing opinion was the prevailing opinion for a reason, and this is unfortunately Serbia’s worst ever result at Eurovision. Princ received points from just five countries, including 10 points from Austria and maximum marks from neighbours Montenegro. This is likely a result of many known factors, including mutual language intelligibility and a music industry largely shared between the two countries. Despite this result, Serbia’s qualification rate still stands at a very respectable 75%; and the country’s previous non-qualifications have often resulted in a bit of a change in direction from the broadcaster, so what Serbia may do in 2026 remains a bit of an unknown for now.

Georgia finished fifteenth in this semi-final, again scoring 28 points from five countries, though being ranked below Serbia as the country did not score any sets of 12 points. It did, however, receive 10 from Armenia and 7 from Israel, again benefitting heavily from the semi-final allocation draw. It is perhaps, therefore, reasonable to assume that Georgia may have scored notably fewer points had this draw worked out differently. After Nutsa’s qualification last year broke a six year non-qualification streak for Georgia, the country seems to be back in the wilderness, and its qualification rate now stands at 47%.

Last, and unfortunately least, is Montenegro. Returning to Eurovision for the first time since 2022, many Eurovision fans had high hopes for the country, particularly when Nina Žižić was chosen as the representative for the second time, however “Dobrodošli” completely failed to go anywhere in Basel. I don’t doubt that performing in the so-called ‘death slot’ did nothing to help their chances, but frankly I think this song was doomed wherever in the running order it ended up. This is clear from the fact that, had Serbia not been voting in this semi-final, Montenegro would have ended up with nul points. It’s only thanks to the aforementioned factors that the country finished with a not dreadful score of 12 points. I really do hope we see Montenegro back at Eurovision in 2026, and we have reason to be optimistic at the moment, but I don’t know how many more poor results the country will tolerate.

Montenegro finished last in the second semi-final of Eurovision 2025. (photo: Sarah Louise Bennett/EBU)

As promised, I’ve also been analysing running order trends in Eurovision semi-finals and have a few findings to share. Since 2013, the running order for all Eurovision shows has been determined by producers, with the only constraints being a random draw to determine which half of the show each country will perform in, and another draw to determine the exact performance number of the host country. This has now changed slightly for the final, but for semi-finals, it remains the same. Since then, we’ve had twenty-four semi-finals so can begin to draw some conclusions.

Firstly, a note about how I tackle the issue of different numbers of participating countries. Since 2013, we’ve seen as few as fifteen countries competing in a semi-final but also as many as nineteen. Clearly, it doesn’t seem entirely fair to compare position 15 in a fifteen-song competition (i.e. the closing act) with position 15 in a nineteen-song competition (midway through the second half), so I instead compare songs based on how close they were to either the start or end of the show.

For example, in a semi-final with seventeen songs, I would assign the first eight entries positions 1-8, the middle entry position 10 and the last eight entries positions 12-19. Position 19 is always used for the song closing the show, and position 10 is used if and only if the show has an odd number of songs. I understand that this is a little confusing, but I believe that it provides a better foundation for analysis. To demonstrate how this works, here are the running order numbers assigned to each of this year’s participating entries across both semi-finals.

For complete accuracy, I have included the relevant automatic qualifiers in each semi-final based on where they performed. Marko Bošnjak from Croatia cannot be deemed to have performed second-last in the first semi-final when viewers still had Switzerland’s entry to sit through afterwards. All running order trends ultimately boil down to the variable ability of an audience to remember a song depending on when in the show it was performed, and whilst the automatic qualifiers didn’t compete in the semi-finals, they performed as if they were competing and this must be acknowledged.

Statistically, the ‘best’ running order position in semi-finals since 2013 has been right in the middle, so-called ‘position 10’. As I said earlier, this is only used when there is an odd number of performances in a show (and as such was not used in the first semi-final of 2025), however, every entry that has performed in this slot has qualified for the final, with an average finishing position of 6.10. 

This year, Miriana Conte from Malta performed 9th out of 16 competing entries in her semi-final, however, with adjustments being made for the presence of automatic qualifiers, she ended up with this magic position and continued its run of luck by qualifying for the final. Other notable acts to qualify from this position include Koza Mostra from Greece (2nd in SF2, 2013), Softengine from Finland (3rd in SF2, 2014) and Nadav Guedj from Israel (3rd in SF2, 2015).

Position 10 may have the highest qualification rate, but a different position has the highest average finishing position. This is in fact position 19 (A.K.A. the show closers), with an average finishing position of 5.21 and a qualification rate of 83%. It is worth noting, however, that this position has been used a lot more than position 10 (every show, after all, must have a closer). Theo Evan’s failure to qualify for the final represented the first non-qualification for this position since 2021.

In fact, since 2013, only one show closer has finished lower than eleventh in a semi-final, and that is Latvia’s Triana Park in 2017, who came dead last in the first semi-final despite closing the show. However, this position has seen eleven top three finishers, including Eleni Foureira from Cyprus (2nd in SF1, 2018), Destiny from Malta (1st in SF1, 2021) and Käärijä from Finland (1st in SF1, 2023), and any act chosen to close a semi-final in the years to come should feel relatively confident in their chances of qualification.

The ‘worst’ running order position is the ‘death slot’, position 2. Since 2013, only 29% of acts performing in this position have qualified for the final, thanks in part to a complete qualification drought between 2018 and 2022, and this slot has an average finishing position of 12.21. It does seem that unless you are providing a bombastic opening to the show, the earlier slots are not a good place to be; no act performing in position 2 has finished higher than sixth in a semi-final, and the position has seen four last-place finishers.

The graphic above shows an overview of statistics for each running order position since 2013; and whilst the average finishing position and qualification rates for most slots are hovering somewhere between 45% and 65% (with a little variance which is to be expected), there is a definite skew towards entries performed closer to the end of the show. As well as position 2, positions 3, 11 and 14 all have qualification rates lower than 50%, although both 3 and 14 could change that with a good couple of years.

None of the ‘best’ four running order positions in terms of qualification rate are in the first half of the show, however, the average number of qualifiers from the first and second halves aren’t as different as you might think. For obvious reasons, position 10 isn’t counted in either the first or second half, which means that the two averages don’t add up to the ten qualifiers each semi-final has, as, in a semi-final in which the ‘middle song’ qualifies, we only have nine qualifiers when considering the first and second halves. 

However, since 2013, we see that we have an average of 4.4 songs qualifying from the first half, and 5.2 from the second. Whilst there is still an advantage to being in the second half of the show, and this is still considered favourable, the gap isn’t huge, and in the last three years we’ve actually seen more qualifiers from the first half than the second (if 2022 is included as well, the numbers are equal). This year was particularly skewed, and the first semi-final became the only semi-final since 2013 to have more than six qualifiers from the first half, as seven of the first eight competing songs made it through to the final.

Perhaps this indicates that running order ‘bias’ is diminishing over time, or perhaps this is simply the result of some unusually skewed allocation draws. It’s difficult to say, as this trend has really only just started to emerge, but it will be very interesting to see what happens over the next few years. The above graph illustrates the qualification rates for each position, and the dotted line shows the overall trend – i.e. entries are more likely to qualify for the final when performed later in the show.

That’s just about everything I wanted to share in this article. Once again, I have collated all the relevant qualification streaks and rates for your interest. Next up on this blog will be my analysis of the jury votes in the grand final, with many interesting patterns and trends to look at, so do keep an eye out for that, and I hope you have enjoyed reading this article.