3 ms·

Every time I read an article about Shazam, I'm deeply skeptical of the claims. Example: http://www.forbes.com/sites/parmyolson/2014/08/20/shazam-hits-100-milli

by davidu 11y ago

Every time I read an article about Shazam, I'm deeply skeptical of the claims. Example: http://www.forbes.com/sites/parmyolson/2014/08/20/shazam-hits-100-million-active-users-as-it-gears-up-for-ipo/ http://www.forbes.com/sites/parmyolson/2014/08/20/shazam-hit... Article makes it seem like Shazam has metrics like a social network, but my experience, and everyone I've ever seen is that it's rarely used -- and simply to quickly identify a song. This latest one in The Atlantic implies people use it far more frequently then I think they do, and I'm again, very skeptical. And even if they do, they don't shazam hits they know, or songs about to be hits, but they shazam that random song they can't identify -- not sure that's an indicator of hit'ness. From user numbers, to capital raised, to stories like this, each time I read about Shazam, I say to myself "Gosh, I really love the app, but can it really be this big and lucrative?" I feel like I'm missing something. I hope I'm wrong.

girvo 11y ago
I use Shazam as a kind of "music bookmarking" service. I hear a track I don't know, and I Shazam it to identify the track and it's saved into the list for later. That's pretty valuable data for Shazam.
Grue3 11y ago
How is this surprising? Shazam tracks searches and location. If a lot of people search a specific song in a specific location, that means they 1) liked the song, but it's new to them 2) the song is nevertheless played quite often. Both of these are good predictors for a viral hit. Because nobody uses Shazam on well-known songs, this filters out a lot of noise.
jasode 11y ago
>, they don't shazam hits they know, or songs about to be hits, but they shazam that random song they can't identify -- not sure that's an indicator of hit'ness. One of the Shazam folks gave a 10 minute presentation at O'Reilly conference.[1] She gave an example of how early Shazam datapoints correctly predicted that Katy Perry's song would beat Lady Gaga's song. (The singles were released at the same time.) Perhaps she was cherry picking clear correlations but it doesn't seem too far fetched that the early trends it sees from users is a reliable prediction of a future Billboard chart position and sales. [1]https://www.youtube.com/watch?v=mcTPvxo8SXY https://www.youtube.com/watch?v=mcTPvxo8SXY deep link at 3:32 that compares KP and LG: https://youtu.be/mcTPvxo8SXY?t=3m31s https://youtu.be/mcTPvxo8SXY?t=3m31s
mintone 11y ago
Think about the kind of data that they would collect. I personally rarely use Shazam, but when I do, it's more often then not a song I don't know at all rather than trying to remember a song's name. So someone like me hears a song and wants to identify it. Then another person, and another. So you have three people in a specific location at the same time looking to identify a song. That would suggest that those people like what they hear. You know what the song is so you know interest has piqued. If you cross reference that with radio plays at that time you have a new metric that you couldn't measure previously - reaction to a radio play. It's incredibly interesting because of the type of data it collects, which to be honest I think the article explains explicitly and very clearly!
- pbhjpbhj 11y ago
  It's only just struck me that it can also catch copyright infringement as location of public performance of a work that's not authorised can be established - especially if more than one person used Shazam to ID the work. Interesting. Perhaps this was obvious to everyone else. Personally I've never used it.

by davidu 11y ago

girvo 11y ago

I use Shazam as a kind of "music bookmarking" service. I hear a track I don't know, and I Shazam it to identify the track and it's saved into the list for later. That's pretty valuable data for Shazam.

Grue3 11y ago

How is this surprising? Shazam tracks searches and location. If a lot of people search a specific song in a specific location, that means they 1) liked the song, but it's new to them 2) the song is nevertheless played quite often. Both of these are good predictors for a viral hit. Because nobody uses Shazam on well-known songs, this filters out a lot of noise.

jasode 11y ago

>, they don't shazam hits they know, or songs about to be hits, but they shazam that random song they can't identify -- not sure that's an indicator of hit'ness. One of the Shazam folks gave a 10 minute presentation at O'Reilly conference.[1] She gave an example of how early Shazam datapoints correctly predicted that Katy Perry's song would beat Lady Gaga's song. (The singles were released at the same time.) Perhaps she was cherry picking clear correlations but it doesn't seem too far fetched that the early trends it sees from users is a reliable prediction of a future Billboard chart position and sales. [1]https://www.youtube.com/watch?v=mcTPvxo8SXY https://www.youtube.com/watch?v=mcTPvxo8SXY deep link at 3:32 that compares KP and LG: https://youtu.be/mcTPvxo8SXY?t=3m31s https://youtu.be/mcTPvxo8SXY?t=3m31s

mintone 11y ago

Think about the kind of data that they would collect. I personally rarely use Shazam, but when I do, it's more often then not a song I don't know at all rather than trying to remember a song's name. So someone like me hears a song and wants to identify it. Then another person, and another. So you have three people in a specific location at the same time looking to identify a song. That would suggest that those people like what they hear. You know what the song is so you know interest has piqued. If you cross reference that with radio plays at that time you have a new metric that you couldn't measure previously - reaction to a radio play. It's incredibly interesting because of the type of data it collects, which to be honest I think the article explains explicitly and very clearly!

pbhjpbhj 11y ago
It's only just struck me that it can also catch copyright infringement as location of public performance of a work that's not authorised can be established - especially if more than one person used Shazam to ID the work. Interesting. Perhaps this was obvious to everyone else. Personally I've never used it.