skip
/skip/
How catchy is your song? According to data just released by Spotify, nearly half (48.6%) of all people skip a song before it finishes. Even more shockingly, nearly 25 percent move on in the first five seconds.
The finding comes from Paul Lamere of Echo Nest, a music intelligence company just acquired by Spotify. The buyout has given Lamere access to a massive data trove of listening, and some scary skip statistics. “The big surprise for me is how often we skip,” explains Lamere.
“On average, we skip nearly every other song that we play.”
Younger users skip more often, Lamere found, and those on mobile devices are skipping more than those on desktops or laptops. “When we are more engaged with our music – we skip more, and when music is in the background such as when we are working or relaxing, we skip less,” Lamere continued. “When we have more free time, such as when we are young, or on the weekends, or home after a day of work, we skip more. That’s when we have more time to pay attention to our music.”
That said, great songs with catchy openings have a far better chance of survival. “Most of the song skips happen within the first 20 seconds or so of the song,” Lamere continued. “After that, there’s a relatively small but steady skipping rate.”
Written while listening to Mahler.
interesting stuff, thanks.
So the study is saying that the world is filled with people with the attention span of a 12-year old?
Same goes for articles posted online. It’s the nature of the internet.
The sooner people realize that a song need only be 1:20 (or less) for online consumption, the better. You’d get more streams per hour with shorter material- or rather, more songs played per hour depending on which side of the business you work in.
I would actually analogize it to a record digger with a portable record player who has 1000s of records to look through in a limited amount of time. Often diggers know within seconds if the song is for them. With online’s infinite choice and limited time, why waste time on a song that you will likely not enjoy?
indeed! it is the musicians job (and I say this as a musician) to capture the audience. The internet age has made it easier to record and release music regardless of how ready or high quality it is. I love music and love listening to a well done album start to finish, but most songs don’t catch me that way. I skim until i’m grabbed by something that stands out.
Do artists/labels/publishers have access to any of this sort of data? Meaning, of course, specific to their own songs?
Lots of sites, bandcamp, for one, give the poster access to all the stream data: downloads, play counts by day, skips, partial and complete plays. Not all services but some do. Youtube also tells the channel owner how many unique plays and how many minutes total per channel and per video, plus the average minutes per view per video. very good info to see whether your video is just getting clicks or if viewers are engaging with it for the entirety.
As someone cleverly brought up somewhere else I saw this (Reddit maybe?) a TON of those skips are most likely someone looking up a single and the track after it auto playing until you change it. Doesn’t necessarily help anything, but I think it’s less people only giving tracks 5 seconds to hook them, and more people actively picking what they want to hear next.
An exciting and satisfying start gives the song “credibility”, and offers hope that there will be more such moments if you keep listening.
A boring start that sounds like a million other songs loses credibility, and it tossed easily.
That’s how I interpret it.
—Glenn
http://www.reverbnation.com/GlennGalen
Well, there is nothing to take away from this study.
If the song is not what you want to hear “at a particular moment / mood” of yours, then you’d naturally skip it and find something else. The same song you skipped a few days ago, might be what you are looking for a few days later. It has nothing to do with musical qualities of a track or another.
This is clear thinking. Human nature. 🙂
Yes, well put. There needs to be a differentiation between people looking for something they already know and want to hear, and someone who is seeking new music to listen to.
Also, when in a record store buying new records and having a stack to work through, I would start at the intro but always seek through the entire track (one of the great things about vinyl), for me, most of the time the intro was just for mixing anyway so I didn’t put too much emphasis on the first minute of a track either way. But then, this was music made for mixing and not ‘plain’ listening, so the criteria would again be different depending on whether it was a dj or listener previewing the music..
If you skip every song before it finishes, (or specifically, click the next song) then you won’t get commercials on the free version.
Only 48.6%?
Disgusting. People have no respect for music.
It apples to any message – if you don’t get it in 5 seconds you lost!
We are becoming robots.
This is fascinating and can only be good for music. Most people in the world are not ‘serious’ music fans (all though we like to convince ourselves they are…)
Music does not really matter to most. There is a perception that to get ‘into’ recorded music requires a great deal of time and effort – most people are a little intimidated by this notion.
With this ‘old’ audience still out there, with these people we get away with barriers (formats, prices, new unfathomable ‘genres’)
And the big barrier ‘THE INTRO’ – from ‘Yes’ to ‘Brian Eno’ these ‘intro’s’ have gone so far up the artists ‘ar++!’ that it’s moved the ‘music’ so far from reality that most right thinking people have completely LOST INTEREST!
So, this can only be a great thing – streaming means you either catch them in the first 30 seconds or you are screwed!
That’s nothing… 79% of readers of DMN skip straight to the comments before reading the article…
Maybe worth mentioning is that artists get paid in full for every stream over 30 seconds.
That number doesn’t seem unreasonable to me, nor particularly troubling. I suspect a lot of that skipping is people checking out songs or albums to see if there’s something there that interests them. That’s certainly me a lot of the time – checking artists I’ve heard about but not heard, songs that are getting buzz that I don’t know yet, and so on. If I’m listening to Spotify Radio, songs will come up that I’m not into so I skip on. My behavior doesn’t reflect a sad attention span or a lack of respect for artists and has a lot more to do with the possibilities that Spotify offers me.
Really shows it takes a Pro songwriter, musicians, producer to write, compose and create a song worth listening to. Too much of the music today has lost the elements of craftsmanship and dynamics including being way too loud and over compressed. There may be a movement back to dynamics. I know our masters are now no longer overly loud and they seem to get played. “the Doctor.”
I do not see what is to be taken away from this study.
If the song is not what you want to hear “at a particular moment / mood” of yours, then you’d naturally skip it and find something else. The same song you skipped a few days ago, might be what you are looking for a few days later. It has nothing to do with musical qualities of a track or another.
it probably stresses more the diversity of listeners and the inefficiency of music distribution in general (be it on lists, streaming, random play etc etc)
If this study made any sense, it would mean that there is “one unique song” to be written and to be listened to.
But it does not work like that, and even with yogurt, it doesn’t work like that: there are many flavours for different tastes, and people are happy to choose what they like. It does not mean that this strawberry yogurt is better than another or that the other ones are bad.
100% of the DJs skip 99.99% of the tracks before they finish (the 0.01% is the last track).
this is nothing new. we used to call it “needle dropping” or the “eight second ride” .. we often selected on air singles this way. some of us still do. :/
HahnsH is hitting on the bigger issue, in that SPOTIFY DOES NOT PAY ROYALITIES on 48.6% of songs streamed on their service because the stream did not meet the 30 second minimum duration to qualify as a “royalty bearing stream.”
And you think an artist should get paid for 5 seconds of a play?
Actually, why not? And while we’re at it, where do we draw the cut off point? Five seconds? Ten? A minute? The whole song? Should an artist not get paid because the listener skipped five seconds of the outro?
The artist’s (rights holder’s) “job” is done when the listener has the opportunity to listen to a track – regardless of how much they listen to. It’s not like you’re entitled to a refund on a restaurant meal you didn’t finish (most places) and there’s no good reason why Spotify should be entitled to a royalty break on tracks that hadn’t been listened to in their entirety. The user pays for the ability to listen to a track, not for how much they listened to, hence the moment this ability is exercised (by pressing play) the rights holder enabling it is eligible for payment for their service.
Actually, why not? And while we’re at it, where do we draw the cut off point? Five seconds? Ten? A minute? The whole song? Should an artist not get paid because the listener skipped five seconds of the outro?
Simple. It’s governed by whatever agreement is in place between Spotify and the copyright holder.
Because you’ll be diluting the money pool substantially. As for the cutoff? I dunno, I’d be fine with like 20 seconds. Five seconds is pretty clearly in that realm of the person didn’t really want to listen to that song.
HahnsH is hitting on the bigger issue, in that SPOTIFY DOES NOT PAY ROYALITIES on 35% of songs streamed on their service because the stream did not meet the 30 second minimum duration to qualify as a “royalty bearing stream.”
I had to correct that to 35% – I wanted to be fair to Spotify.
Interesting. I would not attribute it to the music as much as the ADD that people have these days more than ever.
If anything, I think the catchier the song, the more likely I am to skip it. That’s why pop music sucks, it is created based on a formula and you always know what comes next. There are no surprises and nothing interesting to “discover” besides the main melody which is usually extremely predictable. Beat is always in 4/4 or 6/8 time-signature, lame-ass chord progressions, barely any “human” element in the instrumentation, cheesy lyrics.. Just awful. It’s not true that people’s attention span is that bad, I recently went to a Chris Thile concert in San Francisco, where he played J.S. Bach’s “Partita No. 1 in B minor” in its entirety and people absolutely loved it! It was heavy stuff and the audience was just your regular mix of young/middle aged folk. Some music is just like junk food – it seems like a good idea at the time, especially if you’re feeling lazy but I won’t recommend it as a steady diet!
If anyone reading these comments is an actual music fan, do yourself a favor, check out Chris Thile.
As arguably the best player of his instrument on the planet, but probably more to the point, what he does has nothing remotely to do with the music “business”.
The headline seems to be wrong. The Echo Nest report doesn’t address the percentage of people who skip songs; it speaks to the percentage of song plays that Spotify users skip. This is entirely different.
It would be more useful to understand how skipping behavior is distributed across the entire population of Spotify users. In other words, what percentage of the population skips 0-9% of song plays, what percentage skips 10-19%, etc? And for each of these ranges, what is the median number of songs played? This is important because the more songs someone skips, the more songs they can skip. Suppose you have 1 hour a day when you can listen to Spotify and all songs are 2 minutes each. If you never skip, you can listen to a maximum of 30 songs. At the other extreme, within that same hour, you can skip 1200 songs after 3 seconds each.
So knowing how big each of these segments are is critical in terms of understanding Spotify usage. It could be that the heavy skipper segment is relatively small and that the light skipper segment is much bigger. Or not. But the key to doing any kind of meaningful analysis is to understand the relative size of these skipper segments.
The headline is wrong, how about Spotify does not pay royalties for 35% of the music streamed on its service. Who else has this no-pay, loop hole deal, Beats? I know Pandora does not, they are obligated to pay royalties as of first note.
Who’d be up for releasing an album containing only very catchy 5 seconds “songs”? 🙂
Why Spotufy won’t pay you…
Does this appear on the artist as 1 stream or as 0 stream ? do artists see that some tracks are being skipped?
Right now, there is no accounting or disclosure regarding streams less than 30 seconds. I assume a stream greater than 30 seconds counts as a full stream but there is no way to tell. As I write that, it wouldn’t surprise me if they pro-rate partial stream royalties too.
Thanks for the reply.
The first 5 seconds mean nothing about the song. Even the shortest pop song. This is clearly misinterpreted data since a user on spotify is seeking specific music. If anything from an artist/genre plays and the user is not expecting it or likes it in general s/he will skip by the time s/he checks out the “now playing” bar (let’s say it takes maximally 5 seconds to check it out). Most users use spotify to play stuff they already know, not to discover unknown music. On the other hand, pandora, itunes radio, rdio, these services offer a random selection of songs smartly selected according to preference criteria and by design the user expects to explore new stuff, so my bet is if you do the same study in those services the data will be skewed much further to the right.
The first five seconds skips are irrelevant and unsurprising. Of course a lot of people are going to skip in the first five seconds.
A song you don’t like, skip it.
A song you’ve heard too many times recently, skip it.
A song that you think was something else and realize it’s not, skip it.
Really, and there are even more reasons as to why you would skip a song. What’s more surprising to me, is the number of people who skip a song in the middle of it. This to me, is much much more destructive. If it’s a song you didn’t know, unless it’s making your ears bleed, how can you have a fair opinion of it unless you’ve heard it from beginning to end ? And if it’s a song you already like, then how fucking short is your attention span that you can’t even listen to a song you like for more than a minute or so ?
Exactly, under 5 seconds clearly means there was no intention of ever listening to that song. Just clicking next until one comes on which makes a person think “yeah that’ll do”, the interesting thing is what happens after that point.
It’s not surprising that mobile devices skip more than non-mobile, if you have headphones in then you’re going to be more picky as it’s the only thing you can hear.
The tiny tail off at the end is not as large as I would’ve thought – but then I listen to alot of electronic music where the end is usually just a gradual deconstruction of musical layers, which I always assumed was a nod to how the producer imagines their music will be played; and lends itself to transitioning to the next track during a mix.
The analysis on which this post is based (http://musicmachinery.com/2014/05/02/the-skip/) has a few more interesting graphs to do with time of day and day of the week that are worth checking out.
Well yeah, if you pause it at the end you don’t have to listen to the annoying ads
interesting, i would love to share my observation about listening time of “chill” dj mixes online.
Since recently mixcoud display the number of minutes listened for each mix and for user account in general. This information is not public and it is visible only for the account owners.
I think, it is very interesting statistics as it finally shows the real appreciation of the music we share online. Indeed , one artist can be exited from the high number of plays of his recordings, it still doesn’t mean that people listen carefully. It happened all the time, we can press play in the beginning and then fast forward or just skip the track. Still from the technical point of view this is counted as play. The phenomena of “highlighting” popular media is often used by media promoters.
So contrary to it soundcloud, mixcloud now provide a number of minute listened per mix and with a number of plays we can calculate the average time that audience spend on a mix.
As for me, I made a quick and ruff estimations, and it looks like my mixes are listened in average of 40 minutes per “play”.
My DJ mixes are generally 60-70 minutes, so it means that at least a half of listeners listen to the mix from the beginning to the end, while another half listen for few secondes / minutes.
Actually in the word of multitasking and quick consumption of information in internet, i think that my statistics are not bad, and I was actually “afraid” to have lower stats.