Lone Gunman

New and poten­tially dis­rup­tive music stream­ing ser­vices are hav­ing a hard time break­ing into the mar­ket, with many ana­lysts blam­ing their busi­ness mod­els and oth­ers blam­ing the con­trac­tual demands from labels for the trou­bles encoun­tered. There are also com­plaints about the roy­al­ties paid to artists and poor rev­enues of exist­ing ser­vices.

Michael Robert­son–founder of MP3Tunes and MP3.com–attempts to lift the veil on the indus­try by look­ing at some of the (you could safely say “unrea­son­able”) con­trac­tual demands placed on music stream­ing ser­vices by record labels:

Gen­eral deal struc­ture: Pay the largest of A) Pro-rata share of min­i­mum of $X per sub­scriber, B) Per-play costs at $Y per play, C) Z per­cent of total com­pany rev­enue, regard­less of other busi­ness areas.

Labels receive equity stake: Not only do labels get to set the price on the ser­vice, they also get par­tial own­er­ship of the company.

Up front (and/or min­i­mum) pay­ments: Means large amounts of cash are nec­es­sary to even get into the game. […] This fur­ther sti­fles inno­va­tion in ser­vices and busi­ness models.

Detailed report­ing, includ­ing monthly play counts: Pro­vid­ing addi­tional reports unre­lated to pay­ment, includ­ing over­all mar­ket share of sales in var­i­ous cat­e­gories. […] The labels effec­tively offload their busi­ness analy­sis (and the cost of such analy­sis) onto the music services.

Data nor­mal­iza­tion: With­out stan­dard nam­ing con­ven­tions and canon­i­cal meth­ods for ref­er­enc­ing artist, tracks and albums, the ser­vices are left to try and match artist, track, album names pro­vided by one label with those of another. It’s incred­i­bly inef­fi­cient, as each ser­vice must undergo this process separately.

Pub­lish­ing deals: Once you’ve signed deals with the labels, you then need to cut deals with the pub­lish­ers. […] Although you may have the rights to stream from labels, you some­time can’t get the rights to stream from the pub­lisher, or worse, even find the publisher.

Most favored nation: This is a deal term demanded by every major label that ensures the best terms pro­vided to another label are avail­able to it as well. This greatly con­stricts the abil­ity to work out unique con­trac­tual terms and fur­ther lim­its busi­ness models.

Non-disclosure: This is the main rea­son music ser­vices, not the labels, have been get­ting heat from the artist com­mu­nity. Music ser­vices can’t defend against accu­sa­tions about low artist pay­ments because they pay the labels who don’t dis­close what they’re pay­ing to the artists.

It’s worth not­ing that while Michael Robert­son is a trust­wor­thy writer and likely to have access to peo­ple who know this infor­ma­tion (if this isn’t first-hand infor­ma­tion any­way), he’s also likely to har­bour some resent­ment toward record labels from his busi­ness ven­tures. Still, even with­out a solid ref­er­ence I felt that this was too inter­est­ing to just pass up.

Let’s assume that if you had a few mil­lion pounds, you could prob­a­bly buy your­self some hit songs from a song­writer, some stu­dio and musi­cian time, plenty of mar­ket­ing, and almost cer­tainly get your­self a pop career.

The ques­tion is; is it eas­ier to get your­self into this posi­tion (a mil­lion­aire pop star) via pure luck (by enter­ing the lot­tery) or by enter­ing a com­pe­ti­tion like the X-Factor?

We don’t know how many peo­ple apply for the X-Factor, but based on 10,000 peo­ple at a sin­gle Lon­don audi­tion, we could con­ser­v­a­tively esti­mate 40,000.

Although the X-Factor mar­kets itself on the win­ner receiv­ing a “£1 mil­lion record­ing deal”, recent infor­ma­tion about the con­tract has sur­faced that shows “the vic­tor may only receive £1 mil­lion after at least four albums” (note the ‘may’ and ‘at least’; we’ll ignore these for now and assume they will after four albums).

If we look at the num­ber of albums released by win­ners of this type of show (X-Factor, Pop­stars, Pop Idol), we find that less than one in five win­ners (to date) have released four or more albums.

And we can’t even expect this to improve; plot­ting all the chart posi­tions (for sin­gles and albums) for all of these win­ners, over time, shows a dis­tinct down­ward trend:

So, 1 in 40,000 appli­ca­tion odds com­bined with 1 in 5.5 “four albums” odds gives total odds – of enter­ing the X-Factor, win­ning and becom­ing a mil­lion­aire because of it – of about 1 in 220,000.

The chances of win­ning the lot­tery (with aver­age jack­pot win­nings of £2,053,984) is 1 in 13,983,816. You would there­fore need to buy £64 of tick­ets for a slightly bet­ter chance of win­ning the jack­pot than becom­ing a mil­lion­aire through win­ning the X-Factor. £64 may seem like a lot, but prob­a­bly doesn’t com­pare to the cost of trav­el­ling to/from the audi­tions, tak­ing a day off work to spend a full day there (with food and drink), etc.

Of course, if you funded your own career, you’d also get a much higher per­cent­age of earn­ings, wouldn’t be locked into a lengthy con­tract, and wouldn’t suf­fer from the stigma of being a real­ity star win­ner. So you’d prob­a­bly even have a longer career than these win­ners, as plot­ted below (each bar rep­re­sents a dif­fer­ent win­ner from one of these real­ity shows). As a win­ner, you have a 55% chance of hav­ing a pop career of less than one year, and a 36% chance of less than six months.

With the cor­rect choice of music and by giv­ing the per­cep­tion of progress cus­tomers on-hold in a tele­phone queue under­es­ti­mate the time they have been kept wait­ing and will stay on the line longer before hang­ing up.

Newsweek sum­marises a num­ber of research stud­ies that have looked at the psy­chol­ogy behind tele­phone queues and on-hold music, not­ing the dif­fer­ent reac­tions cus­tomers have when con­fronted with hold music, recorded apolo­gies or esti­mated wait times.

Though it hardly seems pos­si­ble that the Muzak (the term is often used gener­i­cally, but Muzak Hold­ings LLC is a real com­pany) pumped into malls could actu­ally influ­ence shop­pers, the truth is, alas, that it does. James Kel­laris, a mar­ket­ing pro­fes­sor at the Uni­ver­sity of Cincin­nati, says that music can have an impact on a wide array of cus­tomers’ behav­iors, chang­ing their per­cep­tion of time, con­di­tion­ing them to asso­ciate a song with a brand, or lim­it­ing their abil­ity to crit­i­cally ana­lyze a poten­tial pur­chase due to musi­cal dis­trac­tion. “When shop­pers are exposed to music in a store, sales resis­tance decreases,” he says via e-mail. Our brains have a finite band­width for tak­ing in and pro­cess­ing infor­ma­tion, and clog­ging that band­width with music is some­times enough to pre­vent us from mak­ing ratio­nal pur­chas­ing deci­sions, or wor­ry­ing about the time.

The arti­cle also notes how we have the rather excel­lent Erik Satie to thank for the muzak phenomenon:

[Satie] devel­oped a very cyn­i­cal atti­tude toward the lis­tener. Satie was so obsessed with the idea that music could no longer com­mu­ni­cate to the audi­ence, he con­cluded that music in the 20th cen­tury was des­tined to be a vac­u­ous, com­fort­able appa­ra­tus best used as a back­ground for other activ­i­ties, much like a favorite chair.

via Mind Hacks

Walk­ing through the Lon­don Under­ground I usu­ally don’t give much thought to the des­ig­nated busk­ing areas. How­ever, the scheme, started by Trans­port for Lon­don in 2003, is sur­pris­ingly involved, as I dis­cov­ered after read­ing this pro­file of Mike Mut­tel, an Under­ground busker.

Muttel’s offi­cial busk­ing license, good for one year, hangs vis­i­bly from a lan­yard around his neck. It took six months of rig­ma­role to obtain that license, in which time he applied, audi­tioned for a panel of four or five Lon­don Under­ground staff mem­bers and agreed to a manda­tory police back­ground check. The process didn’t cost any­thing, but took tal­ent, patience and a lit­tle luck (audi­tion judges are not required to have back­grounds in music). Still, of the 400 buskers that audi­tion each year, 80% pass. Now that he’s in the sys­tem, Mut­tel is not required to re-audition; he just re-applies for his per­mit every year. He has been busk­ing for almost three years. […]

Of the 28 or so total pitches at 21 Tube stops through­out cen­tral Lon­don, some argue there are really only half a dozen ideal spots: two at Green Park, two at Tot­ten­ham Court Road, one at Pic­cadilly Cir­cus and one at Leices­ter Square. If a busker shows up late for a spot, the pre­vi­ous busker is enti­tled to stay for the next two-hour time slot. Unsur­pris­ingly, this can get messy.

Two great arti­cles on cur­rent research into how artists and songs become hits:

Group Think looks at research pre­dict­ing musi­cal hits using “geo-aware query strings” from file-sharing net­works such as Gnutella.

The geo­graphic loca­tion of an emerg­ing artist is the key to pre­dict­ing their suc­cess […]. “If an artist has the poten­tial to be suc­cess­ful, peo­ple will first start notic­ing them in the small geo­graph­i­cal area where they live and per­form.” In fact, a poten­tial pop star will typ­i­cally enjoy thou­sands of down­loads a day on a local level, while remain­ing rel­a­tively unheard of on a national level. A large diver­gence between local and global pop­u­lar­ity, called the Kullback-Leiber diver­gence, is a strong indi­ca­tor of star poten­tial. The algo­rithm mea­sures the K-L diver­gence to pro­duce a short list of poten­tials, of which 15 to 30 per­cent will go on to reach national pop­u­lar­ity within weeks.

Tak­ing a dif­fer­ent approach, The Anatomy of a Hit Song shows that what makes many of us like a cer­tain song isn’t its sound; it’s the ‘out­side influ­ence’ of our peers lik­ing the song.

While [the researcher] could pre­dict which songs would be pop­u­lar after an ini­tial round of feed­back, he said it’s ini­tially dif­fi­cult to guess which songs will become pop­u­lar and which will be despised strictly on their own mer­its. He cites the per­for­mance of the song “Lock­down” by 52metro, which ranked right in the mid­dle among the 48 avail­able tracks by lis­ten­ers who had no social con­text. How­ever, in two sam­ples sub­jected to out­side influ­ence, it came in first place in one trial and 40th in the other.

As the arti­cle states, these find­ings aren’t strictly con­fined to music; the the­ory likely applies just as much to books, movies and TV shows.