There was a funny story[1] (transcript[2]) on the NPR program On The Media a couple of weeks ago about someone[3] who writes and records dozens of songs per day to put on Spotify (playlist[4]). Many songs deserve to not be played.
There are also hundreds of "cover" songs on Spotify that, from what I can tell, only try to earn plays through mistaken clicks.
For a good example, search for Thrift Shop -- you might be inclined to play "Thrift Shop" by "Thrift Shop" on his hit album "Thrift Shop." It has 675k plays already.
AC/DC has an interesting twist related to cover bands -- the band only has one song (through an old soundtrack) on Spotify. About a dozen cover bands have filled the gap; the top tunes for each of them collectively have over a million listens.
I don't think it's about misclicks only. Another reason they get plays is because they automatically get queued and go on after the real deal when you search for a specific track.
I don't know, "Seaching for Sugarman" is a pretty good Rolling Stones approximation. Adam Sandler made a pretty legit music career out of stuff kind of like this.
But why listen to a "Rolling Stones approximation" when you can listen to the Stones.
As far as I can tell, a large of portion of bands aiming to make it are indeed good or bad approximations of existing pop bands and mover over, entirely aimed at satisfied the fairly uniform taste of mainstream audiences.
Most are poor approximations of this but a few are indeed good approximations of this. But it doesn't matter.
The problem is that none of these develop anything creative or more specifically, they aren't serving any taste pool but this wide mainstream (this is arguably necessary for their success - they are practicing at being good at specific skill, being alternative wouldn't be the same practice).
So even, especially with even, bands being "almost good", there's still no reason for me to listen.
In the old days one had no choice - whatever was on the radio was what you heard, and it was very difficult to hear the whole CD/tape/ancient-media. I used to have a 3-4 song rule - if I didn't like 3-4 songs on the album I wouldn't buy it. It was very hard to adhere to, and breaking the rule almost always resulted in disappointment.
MP3 changed that, but not because of pirating - because listeners could start sampling the whole product to see if they wanted to buy it. Tech improvements keep making that easier. Bands will have to work harder to produce quality work for their fans (including having a good handle on what 'quality' and 'fans' means for them), instead of producing one good song and ten fillers and calling it an album.
It's not very (at the track level) - but at the same far too (at the 'release' level) - accurate.
"Wayward Wind" by Gogi Grant (off "Sweet Memories") is listed -but- the track itself is listed as one of the most popular if you search for her, just not off that particular album. There's a metric ass-ton of releases which probably explains why one particular release has zero even when the track is clearly very popular.
This is very annoying on Spotify itself as well. If you use the 'radio' feature (basically random tracks based on some genre/artist), it often repeats the same track over and over, from different releases.
I see there's a track popularity figure in the API [1] so maybe these songs have zero popularity. And then I can only assume they've crawled every track as the search API [2] doesn't support filtering by popularity.
The "Knights of 'New'" on reddit have to wade through a lot of junk, but everyone else who uses the website benefits from their efforts.
This is a similar deal for Spotify. Some percentage of these songs are good, it might be less than one percent, but if you can get 500 people with decent ears to listen to 1 song each, you can move the needle.
I'd argue this is not as you say, but rather another step towards decentralised marketing.
Because Spotify has its own ranking system and social media/last.fm scrobbling, while a bad song will merely be skipped over, a good undiscovered song with get saved to playlist and exposed via networks to a much greater extent than it had a chance of being before.
Yeah, when that guy was covered recently in the TLDR Podcast, there was no mention that he was basically spamming, and we shouldn't be glorifying people who are making a huge mess while scrounging around for change in the couch cushions of the internet.
I'm gonna be "that guy" and complain again that developers (web developers in particular) really need to learn the basics of scaling. There is no point having your app publicly available if you're not prepared for small spikes like this.
Edit: At the time of posting, the service was 503ing.
Learning to be a developer is still a hodge-podge, unstructured process, scaling is definitely not taught and often talked about in extremely abstract terms because a lot of authors have no clue themselves what is and isn't scalable.
A small VM can take huge amounts of traffic thrown at it. If you happen to be running basic PHP. But install wordpress and it's a whole other story.
Often until you've been hit by a spike like this you won't even know what is and isn't scalable.
That's because scaling is not in the domain of developers, as much as Heroku/AWS/etc. strive to blur the lines. It's in system administration and architecture.
Scaling, like security or maintainability, is a really fluffy and cross cutting concept. It's hard to get it right without having a gradual increase in load. So talking about the "basics" is a bit unfair, I think.
Is there a place to learn how to scale? Even just the basics or a getting started or something?
As someone else pointed out, learning to program is often done on your own, so where would someone go to learn this on their own outside of the hot seat?
Is there a service that tests the scale of web apps? I'm not thinking of software, rather a service that web developers could use to test their app with simulated "real" traffic..
Perhaps their budget (both in terms of time and money) doesn't allow for it...? Don't default to the web developer incapable or disregarding scalability.
I am the exact same way. I absolutely love this idea. This site will see much use from me.
I'm wondering why this isn't a Spotify app instead, though. They wouldn't have to handle the load, and you wouldn't have to navigate to a separate site.
What's a revelation about this? I recall taking out esoteric library books in college that had the vintage checkout cards in the back. Last checkout for the book was 20 years ago. There's a reason they have a music category called "Pop."
I tried to do a book report on The Three Musketeers in high school, only to discover the local library had discarded its only copy because no one had checked it out in years. (This was a few years before Amazon and Project Gutenburg, with no bookstore in town. Nowadays I own two different translations in print and have another on my phone.)
I just checked, it's not there. But it is on YouTube and my old Geocities page has been mirrored. You could find it if you wanted too, but it's not worth the effort.
Storage is cheap with digital music. Even assuming 10 MB/song, you're looking at 40TB, less than $4k/mo on S3. It's not trivial, but it's probably a drop in the bucket compared to their other costs. Realistically the file sizes are much less than 10MB.
4M? Interesting. One of the recurring themes of "long tail" discussions is that everything is requested at least once every month or so. Would like more consideration of why that's not the case here.
I was in a band in my college years that played this "indie / experimental" instrumental rock. We only played maybe a dozen shows that were sparsely attended. There's one show we played that had at least 50-100 people, but that was because we opened for a semi-popular local band. In 2003-2004, just about when I left the band, we recorded an album in on of the band member's bedroom. I doubt more than a handful of people have ever listened to it.
This album is on Spotify. If it has any listens it's probably because I listened to it, or maybe another one of the former band members did.
Album is _Hurt Seconds_ by The Shy Trafficker. I'll warn you that it's a long poorly recorded slog. There's a couple songs with some vocals but most of it is instrumental. Instrumental rock tends to be pretty unpopular I've found so I felt like the "warning" was needed.
I really wanted to listen to the first random song, as well! What is the "I have Spotify" button for? It doesn't work for me... I was hoping that would let me log into my (US) account without being in the country, but nothing happens when I click it.
I got a cool looking jazz trumpet piece, but it seems to want me to sign up for a Spotify account. I guess I should have known, I'm just too used to sound cloud.
[1] http://www.onthemedia.org/story/100-songs-day/
[2] http://www.onthemedia.org/story/100-songs-day/transcript/
[3] http://www.moternmedia.com/bio.html
[4] http://spoti.fi/1gnQ7nd