April 13, 2008 Crowdsourcing for Hits - is it a Mistake?
Two weeks ago I wrote a post titled Create, Validate, Sell. I have been wondering since - could there be a fundamental flaw in the crowdsourcing methods I described to commercially validate music? This may not only be a problem for me, but it could be a serious problem for the record labels and festival operators that are relying upon technology that enables crowds to pick the next “idol”, artist, band or opening act.
Crowdsourcing is the practice of enabling a group (usually a large group) of people to pick a winner, a direction, a strategy, or crowdsourcing can even be used to design something (for example). Faith in crowdsourcing rests upon research that has shown that groups can make better decisions than individuals, even when the individuals are experts.
In 2004, James Surowiecki wrote a book titled Wisdom of Crowds. Just about every venture investor on earth has read this book, and it has been the bible for numerous startups that have wrapped crowdsourcing into their business models. In the music industry you can experience crowdsourcing at work by visiting OurStage, Amie Street, SliceThePie, SellaBand, TheSixtyOne and on many other sites on the Internet. Investors that have been seduced by the potential of the efficiency and effectiveness of crowdsourcing for the next U2, have funded many of these sites.
Here’s the problem - crowdsourcing really works well when the sum of the crowd possesses more knowledge than the expert(s); after all, an expert can never know as much as one thousand people (for example). However, when it comes to songs, ISN’T ALL YOU NEED TO KNOW in the package? Everything - melody, harmony, tempo, pitch, octave, beat, rhythm, fullness of sound, noise, brilliance, lyrics and chord progression - is in the package. If we are moving toward a world where people are more interested in single songs than albums or artists - what else do you need to know about a song to pick a winner? Does the expert have all the knowledge he or she needs to make a decision? Can the sum of the crowd possibly possess more knowledge about a song than the expert(s)?
So, I’m asking your opinion: should those investing in music/songs (like I have) rely on a small group of experts, should we use technology that enables crowds to pick the next hit single, or should we use both?
Reader Comments (23)
Reading your post I recalled a scene in the TV movie "The Temptations" where Berry Gordy conducts a session where songs being considered for release are played and Berry asks the Hitsville USA staff "..if you were down to your last dollar would you buy this record or a sandwich?" According to the scene it wasn't just executives or the producers, but also the musicians, engineers, receptionists, pretty much everyone around at the time. Even though the movie was a dramatization of the story of the Temps, I suspect this scene rings pretty much true as it was based on the book by Otis Williams. It's an interesting example of a form of 'crowdsourcing' from a long time ago.
In his book, Surowiecki was careful to stress the conditions required for the Wisdom of Crowds effect to 'work'. These conditions are neatly summarised on Wikipedia. One of them is independence of opinions: "People's opinions aren't determined by the opinions of those around them". How often are consumers in the music marketplace unaware of the opinions of other people? I'd say just about never. There are always some "information cascades" at work. The industry itself sees to this by always telling us how popular their artists are with other hip people and/or people like us.
So, you may get some benefit from aggregating lots of ratings by crowdsourcing, but if anyone thinks they're tapping the genuine Wisdom-of-Crowds effect by this method, I think they're kidding themselves, or someone.
It is easy to dismiss the "Wisdom-of-Crowds" but it really depends on who the crowd is. If you are just asking random people then the answer could be anything. If you are asking a crowd that you have some relationship with, such as previous buyers or fans, then the answer will be worthwhile. After all the, "Wisdom-of-Crowds" couldn't be much better if they are already customers of yours as they are the ones most likely to buy again.
I started to write a comment here, but after the 3rd paragraph I just made a post instead. Please indulge me by reading On Crowdsourcing for Hits.
Thanks for the thought-provoking post :)
Music is an experience, and the value of that experience will be different for each music listener. For some, music is just a part of a social scene. Those listeners will be drawn to the social networking and crowd sourcing sites because they value that type of interaction. Other listeners value music based purely on the relationships of the fundamental parts (pitch, rhythm, timbre). These listeners will gravitate towards discovery networks that reflect their aesthetic (online radio, magazine reviews, music venues, etc). What everyone wants to know is, "What do 90% of listeners want?" If you know that, you can package a single product to them with very little effort.
I believe, however, that listeners are complex and that a variety of experiences resonate with them. If crowd sourcing consistently delivers the type of music I enjoy, then I will gravitate towards that experience. Personally, the opinion of experts (KCRW.com, Onion AV Club) connects me with the music I enjoy much more quickly than OurStage. Furthermore, I listen to most of my music on a Sonos system and don't have the inclination to listen to new music on the computer. I think the future will include many models of discovery and listening. Bring them all on.
P.S. You can vote for my band on OurStage by going to http://www.ourstage.com/fanclub/dangermaker
Being an expert in music production does not always mean that your an expert in music consumption. I don't think that crowd sourcing with music is much different than testing new material at a gig, if the song stinks then your gonna know about it real quick, you can then write a new version more suited to your audience or drop it, simple.
Thanks for your reply, Bruce. You made an important distinction there that I may have missed in your original dialog, separating the business-minded music seeker from the potential fan. It does make a lot of sense that investors, labels and venue owners will continue to need a pre-qualification mechanism for artists in order to make educated decisions based on popular demand. For them, it is a numbers game.
I also agree that there are flaws with the 1000 True Fans idea on the surface, but I think there is an underlying truth in that artists can make a living doing what they love without needing to sell millions of records. I know that because I know artists who are doing it, artists who you will probably never hear of, who have managed to connect with a handful of True Fans that support their music.
As for every child getting a trophy, artists are still going to have to work for it - i don't expect a fan-o-matic (here's your 1000 true fans, please drive through) to ever exist. But for artists that will do the work, given the right tools to succeed, I think the new music industry will allow niche artists to make a living without compromising towards some industry-approved recipe of what's worthy of consumption.
Thanks for the dialog :)
Clif - we are in agreement. Thanks.
"Fan-o-matic" - now that sounds like a venture to consider.
Having just recently read Surowiecki's book and finding it in agreement with my Anarcho-Capitalist sensibilities, I am always bothered by its critics and its detractors. In my experience, these criticisms often contain a deference to "experts" whether stated or implied.
This always begs the question, "How the hell do you know who the "experts" are? I have yet to find a case where the weaknesses listed for the WOC, don't apply equally to the "expert" selection committees.
I have a few issues with the idea of crowdsourcing for music, personally. First, there's the obvious - that peoples' opinions and tastes are so widely varying that the only thing that can satisfy the entire crowd is something middling and average that neither offends nor thrills. The lowest common denominator, essentially.
You can reason over whether an IDEA is good or bad, and so it's possible for people to tap into that "wisdom" on a large scale - but most people don't use their knowledge to decide whether a song is good or bad. They react to it emotionally. To me, it's more like the wisdom of mobs than the wisdom of crowds - it's not based on rational decisions. They are not answering the questions "Could this song be successful?", but "Do I like it?"
It's pretty simple, if you ask me:
The closer the crowd you use to validate your music is to the crowd that's the actual target audience the better their 'wisdom'.
Ofcourse, finding out if the crowds are similar is the hard part. :)
Crowdsourcing is great in theory, but it conjures up a fantasy world where statistics and demographics behave themselves. Which they don't.
Voters themselves are a demographic - of people who can be bothered to make themselves heard, who think their opinion counts. Unless you can force the whole world to express an opinion on something, your stats are going to be accurate as it pertains to voters ONLY.
People who go on Slice The Pie etc are themselves a demographic of people who want their opinion to be known (as opposed to those who don't). They also tend to be a certain age group. They are also people who are 'actively interested in music'.
The trouble is, this isn't the record/CD/music buying public. It's already become niche.
So if you want to know what kind of music to sell to the demographic of slice the pie, sell them the music they pick. But don't expect the rest of the world to take notice.
I've got through to the last 15, and I still don't believe it all.
The demographic of people on Slice The Pie are:
People who go on the internet AND
Want their opinion heard AND
Can be bothered to listen to lots of music (probably late teens - early 30's) AND
Are predominantly English AND
Listen to Radio One more than any other station
etc etc etc
Although not entirely accurate, you can tell what you're getting at. These ARE NOT the people who go out and buy Leona Lewis. If you really want to know the opinion of the great British public (or the great American public etc) then you need greater coverage than that.
There is also the problem with certain types of music in certain countries. On OurStage I put some of our stuff in Folk. But when I listened to some of the other songs in there, they sounded like country to me - that's because in the states, folk and country are very close and there is a lot of leeway. But in the UK folk music has NOTHING to do with American country, so already some of our stuff didn't fit.
So not only do you have the demographic problem, you also have the interpretations problem, niche problem, and marketing problem.
I'm starting to think the best way of crowdsourcing is simply to release something and see if the crowd will buy it.
By the way, I'm advocating doing the opposite on my blog, Outsourcing your crowdsourcing
:)
You guys are all starting from the premise that public taste is somehow set and eternally galvanised, and that the artist must attempt to locate it so that he can tailor his art accordingly. How sad. The public's taste is on the contrary totally fluid and it has always been up to the artist to give it direction: to focus it so that it can become fixed.
An artist lets nothing but his own taste guide him in his creative endeavour; he views the audience as a hostile whom he must persuade to the superiority of his singular vision. This lecherous running after the crowd reminds me of street walking: it is despicable. Crowdsourcing is not a mistake: it is a disaster.
Prostitution is the bedrock of all professions. Right now I'm waiting tables, being polite to people who need anything but, and earning Affiliate Marketing Dollars based on the amount of merchandise I'm able to move for Amazon...I mean, the restaurant.
Sebastiaan, every time I read your comments I find myself wondering if your definition of "artist" isn't just synonymous with "deliberate obscurity."
Speaking for myself, my motivation as an artist is to give young people access to new ideas. With that goal in mind, "crowdsourcing" and research is just another tool in my arsenal. I want my food to be appealing to people, so I do my research in advance.
I'm also fascinated that a self-professed "Anarcho-Capitalist" would be down with "The Wisdom of Crowds" which bothered me for the same reason Wikinomics did: the crowd does the work, the aggregator harvests the profits.
Yeah, that's fair. And totally different from traditional capitalism, too.
If you want to be artistic why not go and create art instead of mucking about with pop songs. We are working in the realms of popular culture in this conversation.
I am working on an orchestral piece set to eclipse The Rite Of Spring and Turangalila in my spare time. It will be listened to by very few people and not a lot of people will understand it, although it will be 'true' art. If you wish to join me in this obscure reverie you are very much welcome. It will involve poverty and occasional arts council funding, and will probably ruin me. But it will be 'true art' and if the masses don't receive it kindly, should they ever hear about it, it's their loss. My what a smug unheard of composer I've become.
In the mean time lets not forget that even Radiohead's music is pop music, and is nothing without the culture it breeds in. Pop music is inextricably linked to it's audience and potential audience. Without an audience pop music is called 'I have a day job'.
Crowdsourcing can be used to find out who likes your music. It can also be used to find out how more people could like your music. It can also be used to find out what music people like, and writing music like that. Or something in between.
As someone else said, the basic version is, you write some songs, perform them, some don't go down as well as others, you drop them, write some more etc
I think you could well be slightly mental if you always kept in the songs no one likes. Or you could go the whole hog, sack the band, read poetry while herding goats onstage, and dissapear into artistic obscurity. At least you could sleep at night knowing you were artistic.
Music, even pop music, is an artistic skill for which musicians and songwriters have supposedly received some kind of formal training. The public has received no such training and has an infinitely inferior perspective on the matter: they need to be drawn up in incremental steps by the artist. But of course: with the advent of the web everyone is suddenly an expert...
Just because I have only disdain for public taste it does not follow that I must necessarily deal in obscurities: quite the contrary. I think the insistence of the modern songwriter to pander to trends in public taste has been a huge contributing factor in today's lyrical vagueness. If you want to get your musical point across, you had better investigate it from every conceivable aspect, rather than waste your time asking the masses for their opinion. It is and has always been the artist's duty (even in such a humble artistic expression as the pop song) to educate the masses, to lead them onto new ground, to put universal truths into a new light.
Some songs that crowdsourcing would not have recognised as hits:
Good Vibrations - The Beach Boys
Bohemian Rhapsody - Queen
Don't Believe The Hype - Public Enemy
Yes, the public picked these hits themselves without any training, rendering each and every one of these hits worthless. Public taste is a waste of time as they don't know what they're buying, so whenever they do buy something without being told as is the case with the tracks you have illustrated, it just hammers the truth home that whenever the public acts en masse it is usually without any taste whatsoever.
There were much better tracks out in those years that the record industry had promoted heavily, but rather than listening to the arbiters of taste, the stupid public went out and picked what they liked. Idiots.
BTW, you can find some proof positive that crowdsourcing can work brilliantly when applied to music by following this link
Crowdsourcing for the next big hit may not be the right way to think about it. Crowdsourcing for the music that will sell even those tiny niche songs that only a few 1000 will ever buy is the better thought. Crowdsourcing works well in combination with the long tail process. While wavering a bit with the hit idea, after all how often is the crowd being sourced the right combination of people to know the next big "hit"
I also like the idea of crowdsourcing in getting new songs listened to by people actually looking for something new. While allowing those who just want the biggest hits to listen to what has already been sourced rather crowd based or corporate based.
I think if the question posed is whether or not everything about music is already "In the package," then sure... there's no need for crowdsourcing. However, do you really think everything's in the package? I don't.
The music is only one aspect of the music industry. What could be crowdsourced in music? What about Music Publishing. In most of Country and a lot of Pop, the artist doesn't write the songs. What about a crowdsourcing music publishing company that holds contests for the songs it will pitch to the major artists? Wouldn't that be a way to be more competitive than many of the other companies? I'd say so. And where's the flaw in that my friend?
I see what you're saying, and don't totally disagree. I do however believe that crowdsourcing is postured to reconfigure the music publishing industry if the right minds collaborated on the right project...
With respect to who should choose what gets through, I'd say it should be run by a board of panel members who are tasked with giving their unbiased opinions. And perhaps, if enough of the crowd disagreed with their decisions, there could be a built in system of checks an balances.
What about a crowdsourcing music publishing company that holds contests for the songs it will pitch to the major artists?
Hi Wes,
Here's a great video that sums up my feelings on crowdsourcing when it comes to music. The video was created two days ago (June 16, 2009). It's about ten minutes long..
http://blip.tv/file/2250992?utm_source=featured_ep&utm_medium=featured_ep
I believe consumers would be completely incapable of picking great lyrics versus so-so lyrics. If your crowd was made up of people from the music industry, that would be a different story. Consumers / fans - no way. Musicians and songwriters - yes that would probably work.
Check out this post on a revenue model for songwriters.
http://www.musicthinktank.com/blog/songwriters-are-taking-it-on-the-chin-whats-the-solution.html
Best of luck..
Bruce
Crowdsourcing is a category of approaches. The systems that most come to mind involve letting lots of people "vote" for their favorite. Such systems are very similar to polling, because there is little incentive for people to tell you anything beyond what they like.
Compare this to prediction markets, where you reward people for predicting the future and you get VERY different results.
The Iowa election markets have a HIGHLY non-representative pool of participants (heavily white male and fairly well off) YET they are by far the most accurate predictor of elections. They beat out polls that strive to have representative populations - all thanks to the incentives and systems.
In other words, the right incentives could allow even "pop" fans to identify the next big Hip-Hop act, far more effectively than asking "pop" fans to select their favorite hip-hop act.
A delightful discussion Bruce, thanks for starting it!