Forum
Can you be more specific? Which update did you see exactly?
Can you be more specific? Which update did you see exactly?
Studio Albums of 1989, I just saw that figure were still at 10,819,000 yesterday, but today it rises by 1,000 sales, and almost nothing had happened to the album recently, why did it rise even in its pure sales?
Taylor Swift's CSPC has trends set in. It means that when we last estimated her sales, we also estimated the annual sales pace for her releases. For 1989, that's 25k, so basically every two weeks it will be increased by 1k. These trends were set before she purchased back her catalog; we review them when we also review the totals.
I was wondering if the upcoming passing of Madonna as the biggest female artists of all time warrent a pure update?
With the release of The Life of A Showgirl, Swift, easily the biggest pure seller of the streaming era (at least in the West) will have 4 major releases without by-country pure breakdowns, and several albums have accrued a relevant pure units since their last updates. The delta between the by-country sales breakdowns and her total pure is over 15 million, excluding the expected millions TLOASG will sell.
I understand pure albums accrue automatically and updates aren't prioritized over new studies, but I think many people would be interested in seeing the updated specific breakdowns.
Taylor Swift has sold 5 million albums in the US this year, 2025, according to Nielsen and is also the best-selling artist of this Nielsen era.
@komhstan13 yes, we do have most of the numbers updated in the backend, won't require too much work to clean/update the article and that is getting well overdue!
@Mikko do you refer to my last one?
Yeah, that 3/8 answer you gave was confusing, please explain what you mean a little more clearly.
Thank you, I'm extremely excited to see the update numbers and analysis.
I left a comment a few months ago on the CSPC introduction article but it's low trafficked so I don't think it was seen.
Have you considered using QQ heat values to reform your methodology for Chinese units? Below is my thinking:
For instance Netease had announced that on TTPD’s 19th day, it had surpassed 100m streams on the platform (all paid). NetEase has 46m subscribers to QQ’s 104m subscribers, but because QQ is less western-focused lets say at this point TTPD had 200m premium streams in China. TTPD had amassed 2.15Bn streams on Spotify at this time, extrapolating for the present we get that TTPD has 710m PAID streams in China totaling 475k units.
Using the current QQ video streams method (125万 * 5 ) calculates a measly 4k units. (Both calcs use the /1500 for paid streams).
Further, every song on QQ has a public ‘thermal value’ denoted by 热力值. While these thermal values are probably not directly streams, could it not be a more accurate predictor of Chinese streams, especially given the inherent advantage singles recieve from the QQ MV method from having a released MV (whereas nonsingles obviously don’t).
This can be further illustrated through the comparison of specific songs. The QQMV method suggests willow with 523万 is 227x bigger than champagne problem’s 2.3万. Using thermal values (which are indicative partially of actual audio stream data) you get willow’s 7,033万 being 1.77x bigger than willow’s 3,956万. In fact, QQ’s concurrent listeners stat even reveals that 129 people are listening to champagne problems presently, relative to the 116 listening to willow. That’s a part of the puzzle being completely misreported.
While heat values might represent more of a blend of popularity, I think they represent consumption more accurately, & accrue continiously overtime like streams, whereas QQ MV consumption has really stagnated.
Since this comment a few months ago, NetEase announced Taylor surpassed 10 BILLION audio streams on the platform, with its 16% market share that suggests she could have up to 65 billion (likely less, but the point stands).
This isn’t meant to be overtly critical, and I’m sure the intention is to err with caution as to not inflate artist totals, but was just curious your thought and if you had considered this.
TTPD/Lover eras need an update with the "Fortnight" and "Lover Live from Paris" vinyl sales
So let's assume that Swift sells the album 1989 25k per week, or 100k per month WW. You didn't explain further. Is this how the sales of all pure albums are calculated? There are so many
$5 budget albums and $300 box sets like Springsteen. And the comparison numbers for these.
@komhstan13 the current calculation would be *50 for albums not available for sell, so 40k units rather than 4k. Just to be precise as it doesn't change your point! About the overall argument, I checked not more than 2/3 weeks ago if the QQ API I can access to had that 'thermal' figure to improve our methods but sadly for now it seems unaccessible from an automated way ; for most artists it wouldn't make sense to manually retrieve all these figures, but I think that for Taylor, Justin, K-Pop acts, etc, it would be relevant for sure. I'll still look at a solution (maybe NetEase has better data through its API?) soon enough as Taylor Swift/BTS will require updates in the near future, I'll let you know!
So let's assume that Swift sells the album 1989 25k per week, or 100k per month WW. You didn't explain further. Is this how the sales of all pure albums are calculated? There are so many
$5 budget albums and $300 box sets like Springsteen. And the comparison numbers for these.
The trend is an annual figure. I did an article about the introduction of this method about two years ago. The idea is that when album sales are estimated, up to date, I also estimate a current ongoing trend for the release. For example, just now I completed Hozier's debut album data (I'm trying to complete 2010s best-selling albums), it has sold 2.55m pure to date. While doing this estimate, I also estimated that it did about 80k in 2022, 70k in 2023, 60k in 2024. So rather than printing 2.55m sales forever, I added a trend for upcoming years of 50k. Thus, from now on, the algorithm will add near 1k to the album every week. This way, the estimated figure stays loyal to the reality much longer, and if the album ends up booming/collapsing for some reason, we can update that trend in a click from the backend.
Excellent, I agree it wouldn't make sense for manual retrieval for most artists besides perhaps a few token hits
I look forward to seeing these updates in the future, thank you for your prompt resposiveness
