Axios: Apple Has Acquired Asaii, a Music Analytics Platform 'Able to Find the Next Justin Bieber' [Update: No]

MacRumors

macrumors bot
Original poster
Apr 12, 2001
7,447
8,515



Apple has acquired San Francisco-based music analytics startup Asaii, according to unnamed sources cited by Axios. The deal, which has not been confirmed by Apple, was reportedly worth less than $100 million.


Asaii built tools that allowed music labels to discover, track, and manage artists using machine learning. The platform pulled data from social networks and streaming music services, including Facebook, Twitter, Instagram, Apple Music, Spotify, YouTube, and SoundCloud, to find hidden talent.

Asaii offered two products specifically: a music management dashboard for A&R representatives to quickly scout and manage talent, and an API for music services to integrate a recommendation engine into their platforms.


"Our machine learning powered algorithms finds artists 10 weeks before they chart," the startup's website states. "Our algorithms are able to find the next Justin Bieber, before anyone else," another page claimed.

The acquisition will enable Apple to bolster its content recommendations to users, and help it compete with Spotify's efforts to work directly with smaller artists and music labels, according to the report. Apple Music and iTunes are likely to benefit from Asaii's machine learning algorithms.

Asaii was founded in August 2016 by Sony Theakanath, Austin Chen, and Chris Zhang, who have collectively worked at Apple, Facebook, Uber, Salesforce, and Yelp previously. All three individuals now work on the Apple Music team at Apple, as of October 2018, according to their LinkedIn profiles.

In an email to customers shared by Music Ally last month, Asaii said it would be shutting down operations on October 14, 2018.

Last month, Apple announced that it completed its acquisition of Shazam, a popular music recognition service that can identify the names and lyrics of songs and music videos. Shazam could be more tightly integrated into Apple products and services as a result, ranging from Apple Music to Siri.

Update - Oct. 15: Startup incubator and Asaii investor The House has confirmed the "recent acquisition."

"As the first investors in Asaii, we are incredibly excited by their recent acquisition by Apple where they will have the opportunity to dramatically scale their impact and continue building out their vision for the future of the music industry," Cameron Baradar, founder of The House, told Music Ally.

However, Apple declined to provide us with its usual statement confirming an acquisition. MacRumors also received an anonymous, unconfirmed tip today claiming that Asaii's co-founders received standard job offers from Apple as part of a deal worth "way, way less" than $100 million, so it may be an acqui-hire.

Update - Oct. 15: TechCrunch's Ingrid Lunden has also been unable to receive Apple's usual statement confirming an acquisition, leading her to report that the company has not acquired the assets of Asaii. Instead, she says Apple merely hired the startup's three co-founders, consistent with the tip we received.

Article Link: Axios: Apple Has Acquired Asaii, a Music Analytics Platform 'Able to Find the Next Justin Bieber' [Update: No]
 

pavinder

macrumors regular
Jan 16, 2009
105
22
So the algorithms will detect things within songs which, based on musical structure/chord sequences/whatever and chart history, contain elements deemed likely to become popular?
In other words, looking backwards.

This is probably going to make things even more formulaic than they already are...welcome to the homogenised world of unsurprising, generic, non-innovative "music".
 
Last edited:

Yod4

macrumors newbie
Sep 12, 2014
14
32
So the algorithms will detect things within songs which, based on musical structure/chord sequences/whatever and chart history, contain elements deemed likely to become popular?
In other words, looking backwards.
Looks like it’s worse than this. The article mentions that “The platform pulled data from social networks like Facebook, Twitter and Instagram [..] to find hidden talent”. This likely means that at least part of the algorithm weighs things like demographics, social trends, social media interactions etc. It’s got nothing to do with music (Hence the part about the next “Justin Bieber”).
 

Brandhouse

macrumors 6502
Aug 6, 2014
428
674
So the algorithms will detect things within songs which, based on musical structure/chord sequences/whatever and chart history, contain elements deemed likely to become popular?
In other words, looking backwards.

This is probably going to make things even more formulaic than they already are...welcome to the homogenised world of unsurprising, generic, non-innovative "music".
Isn't that what all the music of today sounds like already?
 

karnsculpture

macrumors regular
Feb 19, 2006
136
46
London, UK
A music executive in the 70s, Peter Meisel, used similar methods to dictate what his producers and acts should record (including Giorgio Moroder and Frank Farian) - he analysed what was successful in great detail. This did not work in the long-term, it lead to a lot of very similar-sounding disco and rock records that were hits but eventually the public bought into what was different. This is nothing new.
 

pavinder

macrumors regular
Jan 16, 2009
105
22
Isn't that what all the music of today sounds like already?
Pretty much, although I still believe there's a lot of human choice involved (even though the taste of these people is questionable).

Once it's automated (and "justified") by machines though, we're headed for a complete absence of humanity.

That said, you could argue that judging by the state of music today, anything would be an improvement - as long as they program some randomness into the algorithms.

I think it's about time for the ideas of The Dice Man to be programmed.
 

loekf

macrumors 6502a
Mar 23, 2015
622
208
Nijmegen, The Netherlands
So the algorithms will detect things within songs which, based on musical structure/chord sequences/whatever and chart history, contain elements deemed likely to become popular?
In other words, looking backwards.

This is probably going to make things even more formulaic than they already are...welcome to the homogenised world of unsurprising, generic, non-innovative "music".
Well, with autotune everybody can sing. I even suspect some popular artists talk using autotune.

And please, get rid of the labels. Artists should try to get things sorted without labels in this era of Spotify etc.
 

JSt83

macrumors member
Jan 6, 2014
56
64
You guys get that that whole Bieber line is just them selling the quality of their algorithm, right? Since he's probably the most successful social media discovery. It's pretty dumb to assume Apple bought the company to fulfill some marketing slogan of theirs.

Isn't that what all the music of today sounds like already?
No it's not. There's lots of amazing music being released, never before so easily accessible.
 
Last edited: