January 28, 2010 The Song/Artist Adoption Formula - 2010 Update
This in an update to a previous post.
To the extent that a recording artist (versus an entertainer) is the sum of his or her songs, I am going to stipulate that song-adoption equates to artist-adoption.
I effectively use this formula when working with industry startups and artists to concisely communicate (usually on a bar napkin) the challenges that artists face as they attempt to obtain marketplace traction for their songs.
I have updated the formula (below) to recognize the importance of placing unknown songs into a series of songs that are familiar to listeners (the Adjacent Song Factor).
Fans = L * OFR * SSR * RR * ASF
Fans = Listeners * Optimal Frequency Rate * Social Situation Rate * Resonation Rate * Adjacent Song Factor
- Listeners - a song obviously needs as many listeners as possible.
- Optimal Frequency Rate - a song needs maximum spins (plays) within a compact span of time.
- Social Situation Rate - a song benefits from maximum socialization during that same time period.
- Resonation Rate - the percentage of listeners that a song easily resonates with.
- Adjacent Song Factor - the frequency rate in which a song is placed into a series of familiar songs.
The formula stipulates that for a song to obtain maximum traction, all the variables in the formula have to push up and max out. If you plug the formula into a spreadsheet and play around with scenarios, you will notice (it’s all multiplication), that a single low variable sinks a song (this is important). In other words, you need ALL the variables to work for you to maximize the conversion rate from listeners to fans.
Here’s an extended description of the variables:
Listeners (L)
Listeners (L) is the variable that equals the number of listeners (not fans but receptive listeners) that have frictionless access to your song via a download, a music stream, a broadcast, or by way of receiving your CD.
Optimal Frequency Rate (OFR)
It’s often stated that falling in love with a song is a complex process. For the purpose of this post, I am going to speculate that a song needs to be heard by the average person at least 10 times within 60 days to make a shallow (but lasting) memory imprint. Therefore, 10 spins within 60 days equals the Optimal (maximum) Frequency Rate of 100%.
Less spins over a longer time period equates to a lower Optimal Frequency Rate.
Social Situation Rate (SSR)
Once again, the imprinting/socialization process is complex. Most (young) people need social cues (signals from others) to believe in (adopt and evangelize) a song. When people spin songs in a vacuum (think about the lone iPod user with headphones on), they are less likely to have an imprinting experience than during a shared/social listening session.
Social settings (where social cues are gathered) range from listening to songs with friends, to hearing songs at a club or party, to sharing/playlisting/promoting songs to friends online. In a perfect world, 100% of a song’s early spins would occur within a social situation; this would equate to a Social Situation Rate of 100%.
All social situations are not created equal. If you want to be more specific, assign varying weights to different social situation types.
Resonation Rate (CR)
Resonation Rate is the subjective component of the formula. Listeners are going to love your song(s) along a spectrum. A percentage of listeners (this would be the resonation rate) are going to adopt your song, while others won’t give it a second listen.
Adjacent Song Factor (ASF)
A recent study has shown that listeners easily tire of screening unfamiliar songs. The more often that a song is played within a playlist or stream of familiar songs, the higher the Adjacent Song Factor is going to be.
Now in simple terms…
You need a ton of listeners; a lot of spins within a compact time period; spins that occur within social situations have more impact; you obviously need a great song; and your songs are more likely to be received when sandwiched between pre-existing hits. Sounds like radio doesn’t it?
Reader Comments (26)
Wow! This is a well articulated concept that helps me visualize what it is I need to work on.
Thanks
Everything here rings true to me, but it also leads me to ask: does any one channel really provide all of these?
Most new artists are locked out of radio, so would a popular music blog that champions a band by writing about them at least once a week replace the role of radio in supplying OFR? Seems to me like it could (in fact, I think I've witnessed this in the rise of some local acts in my city). What else?
Also: this formula seems useful when you keep it scoped to the demographic most likely to find your music appealing. It's easier to visualize what those variable values actually are if you know something about the audience you're struggling to reach.
I have been doing a lot of reading lately and a lot of studying, trying to equate how to leverage social media to properly form a community. The more I dive deeper and deeper into the music business world, the more I am seeing how that magical world of 'art' doesn't change the fact that it is a business. I am myself a musician and spent most of my 'music career' (for lack of a better term) thinking that if I write good music, the fans will just come. But as this equation shows, it is just as anything else in the world that all of the little pieces need to be in place, and need to be acted upon proactively in order for traction to be achieved.
Bruce,
This is really excellent. And it explains Beatlemania in 1964:
1) Everyone was listening to radio (L), and
2) The Beatles songs were being played constantly. (OFR)
3) Kids were playing songs for each other socially in groups, because records were relatively expensive and many kids could not afford them; plus, they couldn't COPY the records! (SSR)
4) They were wonderfully crafted and performed and had a huge "resonation rate". (CR)
5) The Beatles songs were sandwiched in on the Top 40 radio playlists. (ASF)
One quibble: the Pandora Study you quoted was only a 60/40 split of people who had a problem with strings of unfamiliar songs. Your sentence made me think *all* listeners have a problem with strings of unfamiliar songs.
But really, great job. It shows all new artists what we are up against, and points the way to gain popularity...even if it is a hugely difficult road to climb.
And I think it also explains why even veteran publishers cannot accurately predict what will be a HIT!
--- Glenn Galen
http://www.GlennGalen.com
And I think it also explains why even veteran publishers cannot accurately predict what will be a HIT!
And it also points the way towards a very interesting future where individuals or small teams of "hit makers" can have the connections, money, placement and access to data/feedback sufficient to run the whole circus from a small office somewhere. Old school, heavily networked label execs teaming up with a younger group of tech heads could be a serious force -- perhaps we haven't seen the last of the Dinosaurs.
It's a very Interesting post as are so many on the MTT. This one got me thinking about introverted (in style, genre or content) artists whose music might be listened to less socially.
When I plugged in a few numbers, though, I found an anomaly. Now, I know the formula is intended as a rule of thumb rather than as mathematical truth. But in case it's of interest, here's the problem:
Suppose:
A equals the number of times someone heard the song alone in 60 days.
S equals the number of times someone heard the song socially in 60 days.
Then,
SSR = S/(A+S)
Say you have 10 listeners, each listen 5 times alone (A = 50) and 5 times with one other listener (S = 10*5*2 = 50) in 60 days. To keep it simple, assume RR and ASF are 1, then Fans = 10 * 1 * (50/100) * 1 * 1= 5.
Now imagine you lose 30 of those solo listens. What happens?
You have 10 listeners, each listen 2 times alone (A = 20) and 5 times with one other listener (S = 50). Again, assume RR and ASF are 1, then Fans = 10 * 0.7 * (50/70) * 1 = 5 fans.
The fan number doesn't change, when in reality, it should decrease. However, it may be difficult to tweak the original formula without making it too complex to be useful!
Frank,
At first I thought you were not serious. After reading your comment a few times, I realized you were really working it. If you are using Google Docs can you post a link to your adaptation?
Also (not sure if this helps) - my suggestion in the post was to "assign varying weights to different social situation types"?
You are correct, this post is more of a general concept than a true science experiment. It would be great to continue to refine it over time, and without mushrooming the complexity.
-Bruce
Hi Bruce,
That's what happens when you start posting formulae on a music blog! Thanks for pointing out your comment about weighting the situation types. I missed that the first time round.
I don't really have any separate calculations to the ones I posted but please let me know if you'd like me to clarify anything.
The basic problem remains the same no matter how you calculate SSR. Suppose you have two identical songs each with identical social listens (by situation and number of listens), the same number of listeners (L) and the same RR and ASF.
Suppose the only difference is that the first song has been listened to more by people on their own (and more times overall since there is no difference in their social listens). What happens for the first song is that OFR goes up and SSR goes down. Depending on how SSR is calculated, the two changes may cancel each other out and each song may have an identical number of fans according to the formula.
What I think should happen is that OSF should be higher for the first song (which it is) but the social measurement should be the same for both songs.
How about looking at the frequency of social situation plays, rather than the ratio to total plays?
Here is one possibility:
Replace SSR with Social Situation Frequency (SSF), where
SSF = 100% if the song is heard by the average person at least 10 times within 60 days in a social situation.
Seemingly brilliant suggestion. Next time I update this, I will revisit and I will probably send you an email to discuss prior to posting. This is why I post this stuff - to get great feedback. Can you contact me using the contact form here..
Thanks!
-Bruce
I like the formula. I like the concept. Still one question remains for me...
How many fans does it take? I mean obviously the more the merrier, but where is the break point? Your previous post was making the point that 1,000 true fans can do the trick under the right circumstances. So how many fans (as described in this post) do you need to get to 1,000 true fans (or is that number really 5,000, or maybe 1,000 true fans per band member)? I would love to see a study (or do one myself) specific to how many fans does it take to reach a given success point (measured in financial terms).
I think with that information one could figure out how many listeners at a minimum one would need to reach, how often to get the tune played, in what settings and so on. It is my opinion that having a concrete target like that would help artists to reach milestones. I believe a lot of artist look at the task of gaining traction as just simply overwhelming. This would give them a chance to break the elephant down into bite sized pieces.
Thanks for the post. I will be thinking about it for a while.
Tom Siegel
www.indieleap.com
Hi Bruce,
Thanks. You're welcome! I've just sent you an e-mail with a few more comments. Also, I have just been thinking about your variable ASF. If the ASF is low, you might get fewer listeners and you may get fewer listeners returning to wherever it was that they heard it (and therefore fewer repeat listeners) but these will be taken account of by the variables L and OFR.
The remaining question is: does ASF matter, independently of its effect on L and OFR? I.e., when they do hear it, will people enjoy a given song more when the songs around it are familiar? I'm not sure that necessarily follows from the study you cited (a href ="http://www.musicthinktank.com/mtt-stats/pandora-satisfaction.html" >full study).
Mind you, you could reason that if ASF is high it gives the impression that a song has a higher critical or commercial status, which may, on average, encourage more fans. Who knows, that may be one reason for the results of the study.
Finally had a chance to read this. Thanks for laying this out. I haven't played around with the math, but you point to factors which contribute to a song's success, which is useful.
The Beatles comment by Glenn Galen resonates with me because I am old enough to have been a part of that. During their years of releasing albums, there was always at least one Beatles song in heavy rotation. So they were a constant cultural presence while they were a band. It makes no sense to me that these days labels sometimes allow three years between albums. In contrast, the Beatles were releasing multiple albums per year. Of course, few artists are capable of putting out that many quality songs, but during that time Motown artists were able to do it with the aid of songwriters, the Beach Boys were releasing quite a few albums, etc. So maybe if you don't have a songwriter who can write that much material, or you don't want to turn to professional songwriters to put out that many hits for an artist, you should think twice about signing him/her/them.
The other thing that was important was the social setting. There are certain songs from then that trigger very specific memories of where I was when the song came on the radio. So it isn't the song so much as the friends and activities I was doing at that moment.
The fact that music listening is much less a social activity than it used to be has greatly lessen the overall value of music, I think. When music listening is something you are likely to do by yourself, you aren't creating memories of music and your friends.
Bruce,
I have a fundamental quibble with your formula. Is it 'De-scriptive' or 'Pre-scriptive'?
In other words, are these the common attributes of successful songs that you measured ex post facto? Or are these the inputs that actually lead to success? There is a big difference.
Let me give a hypothetical example...
7 of the top 10 prize fighters in the heavyweight division are:
1. Well-conditioned (can bench press 350, dead lift 500+, etc)
2. Well coached
3. Trained at a world-class gym for more than 3 years
4. Fought at least one top-20 contender in the last year.
One might think that this is a prescriptive formula for being a top prize-fighter. If you do all of these things, you will be a champion.
But this is really just descriptive data. It cannot be used to describe the probability of success by doing these things. The piece of data that is missing is how many people accomplished all 4 of the objectives, but DID NOT become a champion prize fighter.
What if a million boxers fit the description, yet only 7 of them became top prize fighters? Would this be considered a prescription for success? What if only 100 boxers did less than all 4 things, yet 3 of them were in the Top 10 rankings for prize fighters?
To determine whether or not your data is 'De-scriptive' or 'Pre-scriptive', consider doing a 'look back' analysis to see how many artists had high scores in the attributes you call out, but did not achieve 'success'. I'd love to see you do a follow up on that, and publish the data you already posses as well as any new data.
You know me well, Bruce, and I'm not here trying to poke holes in your theory, willy-nilly. But I want to make sure that this formula is held as accountable as it should be to proper statistical methodology before Artists go off and ascribe to it with expectations of success. If it cannot be, that's cool, but it should probably be sited as such.
I'm a huge supporter of yours and of MTT. This has been a thought-provoking post for me and others (which is why I am writing a response).
In case I wasn't clear, I am not in any way suggesting that your formula is incorrect, only that it may need more rigorous analysis before it should be recommended as a recipe for success. I applaud you for consistently striving to break down the music business into tangible pieces for Artists to understand.
-Jed Carlson
ReverbNation.com
Jed
I have a sentence above (in the post) that says that I use this formula on a "bar napkin".
It would be easy to say that it's all common sense once you read it, but yes I will agree that a lot of research / data analysis needs to be done on the formula. Right now, it's just food for thought.
On the same subject, I find myself leaving this comment a lot lately: this entire industry needs more case studies and research data. It would be great to have some university or some researchers survey numerous artists on the factors listed above, income, success, etc. The data produced via data mining and algorithms does not tell a complete story. Someone needs to (telephone) survey 500 to 1,000 independent artists (and beyond just the artists that are using RN) that are actively performing and selling music. I can get this done, but I need a list of artists (from multiple sources) and some collaboration (from someone skilled at mining for answers) on what it is that we would be surveying.
-Bruce
Amen to that, Bruce.
I agree that ideally this should be conducted by an academic research team of some sort (to avoid any possible bias, or perception of bias). If you find said researchers, let me know so we can help.
re: the bar napkin comment. I've seen your 'bar napkins' before. They can be fairly complex and quite compelling. ;) I would never consider the fact that you wrote it on a bar napkin to suggest that I was necessarily seeing a casual concept.
Few have the passion you do for unpacking and understanding what underlies today's music business.
The exposure part of it has some basis in advertising. There have been studies of to see how much exposure to an ad leads to interest, recall, and then declining interest.
I didn't take the song/artist adoption formula as gospel. Just as a place to start for more discussion and research.
Is there a science to predicting optimum song/artist success? There's probably at least a little more to be learned than we know now.
Jed,
You raise very important points.
It's like the articles on "How to Earn $1 Million" that only take a detailed look at the actions and traits of people who have earned $1M. They really need to look also at people who tried and failed to earn $1M.
You might find that both sets of people had similar traits and took similar actions, but only some earned the $1M, perhaps from simply being in the right place at the right time and a great situation landed in their lap!
Re: the research. You could look at Billboard charts and bands in 1964 to find other bands in heavy rotation on the radio, with compelling songs, but who did not ultimately succeed as The Beatles did.
But immediately you start defining "compelling songs" by after-the-fact performance. I.e.,if the band did great, they had compelling songs. If not, then the problem was the songs.
The subjectivity of the CR is indeed a real hangup with using the formula for prediction.
Part of the question is the extent to which exposure leads to popularity.
We know that heavily promotion results in some artists/songs becoming popular. Not in all cases, but there are enough examples of so-so artists/songs still doing well that we suspect promotion is the deciding factor.
We have also seen songs/artists that are great but are not widely known because of lack of exposure/promotion.
A few years ago it was easier to see a direct relationship between marketing dollars and return on that investment because you could estimate how many CDs you could sell. Now that fans are buying fewer CDs, you can have a popular artist, but not necessarily have the right formula to generate income. For example, I've been watching someone who is very popular on YouTube, but that hasn't translated into a lot of income yet. Her fan base loves her, but is too old to get money from parents for merch, CDs, and shows, but too young to have a lot of money themselves to spend on that sort of thing.
@ Glen. If an artist assumes he or she has a great song (high conversion/resonation rate - most artists assume they have this covered), then he or she really only has to worry about / predict the other factors in the formula.
Bruce, very interesting and constructive premise, thoughtful. Difficulty these days, accurate as your essay may or may not be, is the danger releasing a song in today's climate of piracy. How can it stand that test long enough for it to stand your formula test, in order for it to become a successful business product?
This is the problem our publishing company is currently having. Sure would like to know how to get around it.
Reality is created by agreement. So take the SSR and compute it as SSR2. That gives it an arbitrary extra weight for the concept to be workable (not as an absolute, of course. It's purpose is to motivate, not triangulate - not to be read and taken literally).
Art is for people. OTHER people. So the more people that an artist can get to experience his music at the same creates a 'shared experience' and reality.
I love the simplicity of the formula. Artists, especially independent ones, need to be told constantly to get the music out there - way out there - to as many people as possible as often as possible.
Or get some people to do it for you.
All the best,
Jon
Suzanne's question about how the extent of exposure affects popularity sparked a memory for me. Many, many years ago, perhaps around the original "dot com" explosion, I read an article about a college radio experiment. (I think it was online.) I have tried off & on over the years to find details on when & where this happened, and IF it happened, but have not turned anything up. But anyway, here's what I remember about the story:
A college radio station offered free pizza or some other incentive to students to sit & listen to music for several hours & rate the songs, score them on a scale of 1 to 10. I believe the students had to be regular listeners of the campus station in order to participate. It seems like they had 50 songs or so, maybe even 100, a mixture of tunes in heavy rotation on the campus station & commercial radio, some from lighter rotation, & some that had never received any airplay anywhere. After they had a couple hundred surveys completed, maybe more, they analyzed the results. Not surprisingly, the heavy rotation songs scored the highest, the zero-airplay songs scored the lowest.
So they took the 5 WORST scoring songs and put them into really heavy rotation on the campus station. They also took the top 5 scoring songs out of rotation completely. They had the DJ's make glowing introductions of these 5 worst scoring songs as well, if I remember correctly. After 90 days they ran the listening & scoring event again, with all the same songs as the first time. All 5 of the worst scoring songs were now in the top 10, the original top 5 were somewhere in the middle. The conclusion was that exposure and positive recommendations equaled acceptance from a broad audience.
As I said, I have tried to verify the story, and since I have only put minimal effort into the search, I have nothing. It seems that when I saw this story I was taking a break from my music career, so I don't think it came from someone trying to sell me radio placement services or some such thing. But I do know that I read it somewhere, and it has always made me wonder about the exposure factor. It does seem to make a lot of sense, as does this formula of Bruce's.
Anyway, my 2 cents.
Cheers!
Clark
@Clark Colburn,
Your hypothesis has been proven. Here is the most recent conclusive proof:
http://www.princeton.edu/~mjs3/salganik_dodds_watts06_full.pdf
The 'Herd mentality' is not a myth. People tend to like what they think is popular over what they would prefer if they were the only one to hear it.
This is no different that the concept of 'decision markets' that have been around for a while predicting the outcomes of presidential elections better than polls by asking a subtly different question than polls:
Polls: "Who are you going to vote for?"
Decision markets: "Who do you think is going to win?" (or, sometimes put as 'who do you think others will vote for'?)
This subtle shift from asking about personal preferences to asking people about the trends they see represents a monumental shift in the accuracy of prediction, and the music business would do well to ascribe to it whenever possible.
Each time a song is played on repetition, or endorsed by a DJ, the trend of 'what others must be thinking about this song' is reinforced. And the ultimate outcome is that humans would rather be on the bandwagon than be forced to independently synthesize such unstructured data like music.
We are creatures of convenience.
-Jed
@ Jed-
Thanks for that link. It's a typical academic/scientific read but the truth it contains is powerful. And your comment "People tend to like what they think is popular over what they would prefer if they were the only one to hear it" is another truth.
Music biz coaches are always saying "make great music & the rest will follow" (if you buy their $40 book/DVD/whatever, that is) and it has been proven again & again that it takes so much more than great music. An artist needs to get at least a couple of taste-makers on their side, find a way of getting or creating ongoing exposure, and have somewhat catchy music just for a chance that the "herd" will even notice they exist. Plus so many other factors that can't be ignored, and frequently can't be controlled.
I've given up trying to figure it all out, at least in terms of having so much data that I contract "analysis paralysis." Ha ha! My new, simple strategy is this: write, record & perform the music that is true to me; put it in front of people that either express interest or are known to like my genre; make new friends; repeat until I croak. Of course, each of those steps has smaller components (but not many), but that is the overall plan. I have struggled so much trying to figure out where the "puck" is headed, and lost ground in the process I think, that I need to just Keep It Super Simple for a while.
Cheers,
Clark
Wow! I am new to this site and stunned by the intelligence of this discussion. I am going to live here for the next few months!
That said, and I guess what the dozens of other posts are about, it how to positively impact each of these factors. I am also curious how all this works in the Folk niche where the audience is older and has different social networking habits.
This is a great way to explain the factors required for song adoption. It sparked a few ideas for me on opportunities I can create for myself and my music. Thanks for the inspiration - sometimes you need to hear what you already think you know.
KG