I doubt 38K from physical sales is accurate for China considering she grew in popularity there since 1989 in 2014, but I understand your reasoning. It sucks that they aren't more transparent with their data.

Here's a list of the best-selling digital albums in China based on RTD sales, including all artists, aggregate from all music services that disclose sales data. Link: http://y.saoju.net/szzj/

Taylor has two of the top 3 best-selling albums with 1989 generating 28.7M yuan (1.511 million digital copies) and reputation generating 28.2M yuan (1.41 million digital copies). 1989 costs 19 yuan while reputation costs 20 yuan.

Pretty impressive numbers considering she's the only person to be in the top 10 twice and the only western artist in the top 10 as well. Her success in China is, although expected, quite amazing. Is there any way to include the digital sales or at least mention them somewhere? This is becoming an important part of China's music market. I don't remember a western artist ever performing as well in China to be honest.

The biggest publisher of physical albums in China just revealed the top 10 best-selling artists of 2018 and Taylor is #1 (source: https://mp.weixin.qq.com/s/TsQxg6RcSwgEJtVhY8JwzA). Using the worldwide average drop to determine China's sales is most likely very innacurate.

Hi Yin!

These shows nothing about numbers. Physical sales are basically non-existent in China. Reputation was also far and away the top physical seller in the US, it doesn't mean it hasn't sold much less than 1989.

To reiterate what these users have said: your formula for on-demand audio streaming relies almost entirely on Spotify. Taylor was absent from the platform during the discussed time period, but she indeed did have 1989 available on Apple Music. That means the album was open for consumption in on-demand streaming on that platform from mid-2015 to 2017. Your formula basically ignores any units that the album acquired from those streams, no matter how small or irrelevant you feel that number may be.

From what I've observed, this has been brought to your attention multiple times and instead of accepting this very valid criticism you consistently enter defensive territory retorting with accusations of biased fandom. From an objective lens, no matter how you spin it, your formula literally claims that 1989 got ZERO streams from Apple Music in the nearly 2 years it was exclusively available there. How does this not ring as troublesome to you? It is simply impossible and illogical to make such a claim, and quite blasphemous to not account for it in the data you are presenting as accurate.

As someone who has always boasted the value in giving the most accurate representation of data possible in regards to music consumption, accounting for different markets, different mediums of consumption, and applying numerical modifications to your formulas to account for said discrepancies, it is very interesting to me that this is how you respond. Knowing that your formula for streaming places almost sole weight on Spotify, one would think you would take into consideration the very unique case Taylor and 1989 had with their Spotify absence - but you are blatantly choosing to ignore it in your math and saying that those units literally never existed. It is false.

I strongly urge you to revisit your take and approach to this. Not everyone that criticizes your data comes from a place of biased fandom, but rather from valid observations such as this case. You need to account for unique situations like this, not completely ignore them - otherwise you have inaccuracies, and said inaccuracies lead to the criticism that you are getting now.

Hi Chris!

No, the formula definitely does not claim 1989 got zero streams from Apple music in any given period. That's the whole meaning of an average. By definition, an average will use a higher figure at times, a lower one in some others, but the total will be fine. If we start making assumptions everytime (and there is multiple such 'times' for every artist) we think the average is over-performed, but let it go the rest of the time, then we aren't reaching a more accurate representation of data, but instead corrupting it and inflating totals.

For this concrete example, adding a figure for Apple music would completely ignore the fact that 1989 got very high streams in Spotify when added back, a boost that hasn't happen in Apple Music. Since Spotify was a lot bigger than Apple back then and an awful lot bigger than Apple back in 2015, this boom upon availabiity is highly relevant, just as relevant as streams it could have got on Apple in late 2015/2016. In the same way, the second biggest streaming service in the past after Spotify was Deezer. Deezer has got millions of users in France, a country where the popularity of Taylor is nowhere near as high as in countries using Spotify. Then, should I downgrade the average ratio for her to take it into account?

That's why saying the average must be manually corrupted for a one-time deflation isn't a valid observation at all, it's just that, manual corruption. It's the big picture that must be looked at, not a specific, isolated case that fits one side of the spectrum and ignores the other one, ultimately breaking the consistency of the data.

Hello based on current trajectories, do you see 1989 hitting 20m in the next 5 years? also do you think it can surpass GGGB/IASF/TD especially with an album release boost?

Hi Taylorstan!

It's hard to tell since tracking CSPC ongoing sales of all these albums takes a lot of time, and that won't give a definite answer since a lot can happen in 5 years for all these artists, just look at how Gaga reversed the trend lately.

1989 hitting 20 million in 5 years is possible, it seems a bit optimistic but definitely not unrealistic!

According to Billboard, Taylor has more streams on Apple Music than Rihanna, yet if we follow your formula, one would have to believe Rihanna has way more Apple Music streams. https://www.billboard.com/articles/columns/pop/8501699/apple-top-streaming-female-artists-international-womens-day

Hi Samsas!

In these lists, they always go by lead artists only. If you take off features of Rihanna, she isn't much over Taylor in updated audio streaming units.

Then as I said multiple times, non-Spotify isn't Apple Music only. They got big in very recent years, but many others are concerned by the extrapolation, others that were bigger in the past, including many cheap Asian platform where Taylor wasn't available too just like on Spotify. The simple fact that Ariana tops them both Taylor and Rihanna shows very well that streams remained weak until fairly recently. Even SZA is higher than Katy Perry, this says it all! It's also platforms like Deezer where Taylor is way weaker than on Spotify, and Deezer used to be bigger than Apple Music not that long ago.

I just don't get it. How do these shipments work?? It's reported that 4.5M was shipped for REP. You said there were mass returns. But how could you know how many returns there was??? Cause albums are always shipped and I never knew that you could calculate how many exactly were shipped back. It's the first time I've ever heard of this. Artist always ship why aren't theirs being shipped back. So you are telling me only Taylor Swift this has happened too??? Yes I know it's been reported some places that the physicals weren't selling as much as 1989. But still, I find it unfair. You don't bat an eye when other artist shipments are reported and you included them without any hesitation. Or maybe I just don't understand how this works.

Hi Xavier!

When there is a large gap between retail sales and shipments it's only logical that some copies are shipped back. It happens very frequently, to tons of albums. Robbie Williams' Rudebox is a popular example, same for Forever by the Spice Girls, the last Rolling Stones album, their A Bigger Bang in the US, etc.

I would say that most of CSPC studies include at leat 1 number that is actually lower than the amount certified / shipped at first. It happens quite often with the album that follows a string of highly successful records without managing to do as well itself, like Reputation.

Hi there, I couldn't help but notice a double standard and I just wanted an explanation.

On one hand, in your analysis, you count extremely cheap ringtone downloads from South Korea (that were even free at one point and were less than 0,10$ most of the time) as normal song sales for artists like Beyoncé. In fact, you said notably that Halo is at 13.1M sales thanks to 2.9M ringtone downloads from South Korea. However, on the other hand, you currently refuse to include China's digital album downloads stats into the analysis of artists like Taylor. I've seen you used arguments in the past saying that it's because 1) it's too cheap (to that I would ask how come ringtones from South Korea are counted then), 2) that it's impossible to say which % comes from actual sales and not premium users (that is false, specific albums can never be downloaded for free. 1989 and reputation are among them) and 3) that, as you said on this page, "as long as the IPFI doesn’t treat this as sales it would be pure inflation to act as if they are legitimate purchases", however in that case wouldn't the 13.1M number for Halo also be inflation considering the song didn't even make the IFPI top 10 list in 2009 (sold less than 5.5M sold during its first year)?

Hi Minerva!

It's very wrong to apply old comments to a situation that is now getting sorted but that was previously a huge mess. To exclude the first numbers coming out of QQ was definitely the correct move, tons of fanatics who act as if they understood everything from day 1 were really believing that 1989 sold 14 million units on that platform only, to point out the most absurd example. Right now, most issues are cleared, the only remaining step is to value available metrics correctly, which will likely be possible with the upcoming IFPI report. Also, I always stated Chinese data would be included, retrospectively, as soon as we are able to address them the good way. It was never a matter of "excluding" this market. I was pretty much the first, way back in 2013, to claim that China was coming to be a Top 5 global market by 2020. I haven't said that to them ignore it.

As for South Korea, it's a very different situation. Firstly, they were nowhere near as cheap as many claimed as shown by IFPI reports. Then, they represented a huge market. Many act as if there is "a few cheap downloads" in SK and millions of album sales happening in China. The truth is that even with recent booms for the latter country, the last IFPI report still shown that SK is a much larger market than China. It grossed $473 million in 2017 against $270 million for China, and even focusing in digital gross SK wins the fight. By 2014, when downloads were still high, SK was outperforming China by 2.5 to 1. As you can see, we ain't talking about a few cheap downloads, but about the 6th largest market in the World, ahead of Canada and Australia and only topped by the US, Japan, the UK, Germany, and France. That the very biggest foreign hits get 3 million digital singles sales there, 450k EAS, is very fine and representative of the strength of this market. That's also what the very biggest foreign albums were selling during the 90s. So no double standard, figures are consistent with each other. No need to say that SK figures were never ignored among IFPI reports, that Halo moved 13 million without making the Top 10 in 2009 only resumes it has been selling well during more than 10 years.

I'll answer in 2 times.

1. " To exclude the first numbers coming out of QQ was definitely the correct move, tons of fanatics who act as if they understood everything from day 1 were really believing that 1989 sold 14 million units on that platform only. Right now, most issues are cleared"

That's simply because some people were ignorant and probably mixed yuans and actual sales, I remember in the past you said you saw articles mixing them. Peterstyles13 from ATRL was right in the end. He said this months ago in response to one of your comment claiming it was impossible to know the sales: ‘’QQ Music uses yuans to certify albums, but you can easily track how many albums have been sold from the certifications, which are constantly updated in real time. I literally never saw someone claiming yuans = units. By that logic, reputation sold 13.2M albums in China STICTLY based on QQ Music, which we never claimed it did. Clearly, Chartmasters just does not take the time (most likely out of bias and lack of interest) to understand how things really work. Reputation was certified 2x Diamond a few months ago on QQ because it passed the threshold of 10M yuans spent on the album, which equals 500K copies bought and downloaded considering the album costs 20 yuans. Now, it's almost 3x Diamond (34% away). QQ Music is extremely transparent with their data. Not only do they give the revenue through their certifications, but they also give you the total amount of songs downloaded from the album (you simply divide the total by the amount of tracks to know how many albums were downloaded). Over 9.88M single sheets were downloaded as of right now from reputation, which means around 658K people decided willingly to PURCHASE the album from that one application. Based solely on QQ Music and NetEase Music, reputation was digitally bought around 1.2M times. People claiming that Katy Perry or Rihanna sold "xx million albums in China" are wrong, because they haven't. They most probably are mixing single sheets (song downloads) or yuans with album sales. In Taylor's case, she literally SOLD 1.2M copies from 2 music services.”

2. “As for South Korea, it’s a very different situation. Firstly, they were nowhere near as cheap as many claimed as shown by IFPI reports. Then, they represented a huge market. Many act as if there is “a few cheap downloads” in SK and millions of album sales happening in China. The truth is that even with recent booms for the latter country, the last IFPI report still shown that SK is a much larger market than China. It grossed $473 million in 2017 against $270 million for China, and even focusing in digital gross SK wins the fight. By 2014, when downloads were still high, SK was outperforming China by 2.5 to 1. As you can see, we ain’t talking about a few cheap downloads, but about the 6th largest market in the World, ahead of Canada and Australia and only topped by the US, Japan, the UK, Germany, and France. That the very biggest foreign hits get 3 million digital singles sales there, 450k EAS, is very fine and representative of the strength of this market.”

You showed that South Korea’s market is sizable, but you were unable to demonstrate how it proves songs like Halo weren’t cheap. Is there evidence? Because I vividly remember single sales in Korea being even less than a cent for quite a while. I talked with various Koreans and they all confirmed that you could get Halo for practically less than a nickel. I also checked articles about their music industry and again, it claims the same thing. An article from BBC from 2011 says “the CD industry is stagnant, and digital music sites are seen as vastly underpriced, with some charging just a few cents a song.” ( https://www.bbc.com/news/world-asia-pacific-13760064) Even nowadays, despite them hiking the price of digital songs multiple times since 2013, the price for digital songs is only at around 35 cents ( https://www.digitalmusicnews.com/2019/01/02/south-korea-mobile-music-services-more-royalties/). You use the size of South Korea’s market as evidence that Halo was not cheap, however during the years where Halo was doing good, downloads only represented 10% of the digital revenue and the digital revenue only represented around half of the revenue of SK’s music industry. Around 75% of the digital revenue came from subscriptions. So if, as you say, songs like Halo were never as cheap as people claimed (no idea why my friends would lie about it being that cheap), according to you, what was the price of Halo from 2009 to 2013 in South Korea?