Book Review of “The Song Machine: Inside the Hit Factory”

Image from Pixabay

Image from Pixabay

It is my completely unscientific theory that the music which often matters most to people is the music they listened to when they were young. From Stravinsky to Springsteen to Taylor Swift, the tunes of your youth will likely stay with you for life. These recordings will always get your attention whenever you hear them and perpetually occupy a special place in your heart from their opening bars to their final fades.

Is there really anyone of any age having any music preference who doesn’t get the chills or at very least tap a toe every time they hear the majesty of the Rite of Spring, the propulsive launch of Born to Run, or the megawatt energy of Shake It Off?

Today’s Music Biz and How It Got That Way

The music, artists, producers and companies who are the subjects in The Song Machine: Inside the Hit Factory (W.W. Norton & Company, 2015), by John Seabrook, are not those that I happened to grow up with. Nonetheless, for interested readers who either did or did not come of age at some point during the past two decades, this highly engaging account of the extraordinary changes throughout the music industry will provide readers with a compelling narrative, cultural history, and business case study. This book further excels as an insightful guide through the music industry’s production processes of writing, recording, marketing, distributing and performing today’s chart-topping tunes.

Like a well-arranged progression of chords, each successive chapter skillfully takes you deeper into the operations of the leaders and innovators of the music industry. It is not so much about the music celebrities’ personal lives as it is about the trajectories of their careers, particularly importance of steadily creating viable hits. Moreover, it carefully examines how smash recordings are well-crafted by everyone involved in their creation to make certain they succeed with global music audiences.

Seabrook illuminates exactly how many of today’s hits, as well as misses, have enough deliberate calculation in the assembly of their beats, lyrics and evocative musical “hooks” to send a rocket to, well, Nep-tune and back. His exposition of the evolution of the “hit factory” takes place beginning early Euro-Pop then on to the Backstreet Boys (and their competitors), and next to the emergence of today’s worldwide stars. He devotes quite a bit of his reporting to how this is done for today’s A-listers such as Rihanna, Katy Perry and Kesha by a small and closely knit group of writers and producers. How and why the leading creatives achieved their prominence in today’s music scene is also finely threaded throughout the book.

Going to a Global Go-Go

As colorfully detailed, the US is often the center of the music industry, with many of its leading participants gravitating towards New York and Los Angeles. There are other key international personalities from Europe and Asia. Sweden in particular had first given a start several of the most influential producers with long histories of innovation in Europe. Later on, they brought their work to the US and achieved even greater commercial success.

Another tectonic disruption, online file-sharing, is explained but not pursued in great depth. Rather, and rightfully so, the author chose to examine how purchasing and downloaded MP3s is now giving way to rising volumes of streaming. He reports on the webwide phenomenon of Spotify’s business model, including its disparate economic impacts upon consumers and musicians. (These seven Subway Fold posts also cover a range of developments involving Spotify.)

Clearly and by definition, factories are places where products are fabricated and shipped.  Their operations must be periodically modernized in order to remain competitive. So too, it has become imperative for today’s music industry to adapt or face decline. The Hit Factory takes readers deep and wide into this unique and worldwide production system where hits by many of the mega-stars’ hits are indeed manufactured. Seabrook’s expert prose conveys the incredible effort, business sense and precision this enterprise requires.

Two Part Harmony

If you have the opportunity to do so, I highly recommend reading both The Song Factory and How Music Got Free (previously reviewed in this August 31, 2015 Subway Fold post), together for a comprehensive understanding of how the multi-billion dollar music industry had fallen and then reinvented itself to rise again. Each book individually, and even more so together, deftly captures this unique world’s intersections of art, science and commerce.

For yet another engrossing historical perspective on the state of the music business set a few decades earlier during the 70’s and 80’s rock era, I further suggest reading a highly entertaining account entitled Hit Men (Crown, 1990), by Frederick Dannen.

Finally, all of the foregoing aside for a moment, have things really changed that much in the pursuit of musical success? Once you have finished The Hit Factory, I urge you to also listen to The Byrds’ 2-minute classic hit single So You Want to Be a Rock ‘N Roll Star and then to reconsider your answer. This song’s sentiment rings as true today as it did way back then.

That’s my theory and I’m sticking to it.

Rock It Science: New Study Equates Musical Tastes to Personality Traits

"wbeem", Image by Scott Diussa

“wbeem”, Image by Scott Diussa

Many people have had the experience of hearing a new song for the first time and instantly being thunderstruck by it. They ask out loud or else think to themselves “Who and what was that I just heard?!” Next, they quickly race off to start Googling away in an attempt to identify the tune that just so totally captured their senses.

But what was it about the listeners’ musical tastes that led to this? Moreover, are their affinities for certain musical styles and artists a reliable indicator of their personalities – – and vice versa?

Researchers at University of Cambridge in the UK and Stanford University in the US have recently devised a new method for predicting musical tastes. Their study was published on PLOS One in a fascinating paper entitled Musical Preferences Are Linked to Cognitive Styles, by David M. Greenberg, Simon Baron-Cohen, David J. Stillwell, Michal Kosinski and Peter J. Rentfrow.

These findings were written up in an interesting article in the August 8, 2015 edition of The Wall Street Journal entitled If You’re Empathetic, You Probably Aren’t Into AC/DC by Daniel Akst. I will sum up, annotate and, well, orchestrate a few questions of my own.

Until the introduction of this new method, researchers traditionally pursued correlating musical tastes with the big five personality traits. These include:

  • Extroversion
  • Agreeableness
  • Conscientiousness
  • Neuroticism
  • Openness to new experiences

For instance, extroverts tend to prefer pop and funk. However, test subjects were asked to rate their preferences according to genre which, in turn, can have many gradations and variances. The article mentioned that rock music can include everyone from Elton John to AC/DC, the latter of whom appear in an accompanying photo to the story. (For that matter, who would have ever expected Springsteen to cover “Highway to Hell” during his tour of Australia in 2014?)

The five traits were also previously covered in a different context in the March 20, 2015 Subway Fold Post entitled Studies Link Social Media Data with Personality and Health Indicators.

Using this new methodology, the researchers turned to whether a person is an empathizer “who detects and responds to other people’s mental states” or a systemizer “who detects and responds to systems by analyzing their rules”. The article includes a link for readers to test themselves to assess their own musically influenced leanings towards one personality type or the other.

Leading the research was David M. Greenberg, himself a sax player, who is currently pursuing a doctorate in psychology at the University of Cambridge. The combination of his academic and musical interests is what led him to this line of research. He and his team were seeking a more precise and measurable “sonic and psychological” factors in their efforts to develop a system to predict musical tastes.

The data was gathered from 4,000 volunteers who were tested for empathy and then were asked to rate 50 songs. The findings showed that:

  • Empathizers preferred R&B and soft rock (“mellow” music), folk and country (“unpretentious” music), Euro pop (“contemporary” music), but not heavy metal. Within genres, they preferred “gentler jazz”, as well as “sadder, low energy music”.
  • Systemizers preferred “more intense music” including “punk and heavy metal”.

In the future, this research might be helpful to music streaming services like Spotify to further improve their song recommendation engine. (See also the August 14, 2014 Subway Fold post entitled Spotify Enhances Playlist Recommendations Processing with “Deep Learning” Technology.)

Mr. Greenberg is also interested in researching the reciprocal of his research findings in regards to whether particular types of music can raise empathy or systemizing levels.

My questions are as follows:

  • Might this research also be helpful to a startup like Reify which is developing augmented reality apps for music as covered in the July 21, 2015 Subway Fold post entitled Prints Charming: A New App Combines Music With 3D Printing?
  • Is this research applicable to the composition of music scores for movies, plays and TV shows as storytellers and producers seek to heighten the emotional impacts of certain scenes? (The December 19, 2014 update Subway Fold post entitled Applying MRI Technology to Determine the Effects of Movies and Music on Our Brains discussed Flicker: Your Brain on Movies, a book by Dr. Jeffrey Zacks that, among many other things, covered this type of effect.)
  • Is this research applicable to marketers in developing their ad campaigns aimed at specific demographic groups?

Is Big Data Calling and Calculating the Tune in Today’s Global Music Market?

music-159870_1280-1The extraordinary degree to which big data apps and analytical services are now affecting the marketing, economics, talent development and the popularity of new tunes, has just been thoroughly and expertly explored in an article in the November 2014 issue of The Atlantic entitled The Shazam Effect, by Derek Thompson. This report covers this phenomenon across a multitude of musical genres and commercial venues. I highly recommend checking out this piece in its entirety for a true sense of this ongoing revolution in terms of the leading participants and the fascinating issues concerning business and creativity.

The following is my own summary, annotations and commentary upon just some of the key – – forgive me – – players, market data and open issues worth, well, noting.

I.  Today’s Key Music Business Data Players:

  • Shazam on its surface is an app that helps users to identify a particular song or melody. To date, it have been downloaded half a billion times and is searched 20 million times each day. It can identify emerging songs with breakout potential months in advance. While users enjoy its ability to readily identify a song, the music industry engage it as and early radar array. As well, it assists in identify unknown performers for talent scouts and agents.
  • Pandora and Spotify data sets are used by concert promoters and performers to shape touring venues and set lists. (X-ref to this August 14, 2014 Subway Fold post entitled Spotify Enhances Playlist Recommendations Processing with “Deep Learning” Technology.) One of Pandor’s executives, Eric Bieschke,  is quoted that his online service is not driven by a singular algorithm, but rather “a meta-algorithm that directs all of the other algorithms” to enable users to select songs and artists from the vast  troves of music across the Web.
  • Next Big Sound is a dedicated music analytics firm that gathers, blends and assesses relevant data streams from Spotify, Instagram and other online sources. In turn it sifts through this to identify 100 possible music stars to emerge during the next year. They are currently achieving a success rate of twenty percent. The company also offers a subscription based customizable search tool called “Find”  that will gather and assess selected data flows from Twitter, Facebook and other social platforms. They have found performers’ Wikipedia pages to be valuable predictors.
  • iHeartMedia (previously known as Clear Channel¹) uses Shazam to gauge the virality of new songs and Nielsen Audio deployment of tech called Portable People Meters to track individuals’ radio listening, (X-ref to this July 31, 2014 Subway Fold post about Nielsen’s data and analytics work entitled New Analytical Twitter Traffic Report on US TV Shows During the 2013 – 2014 Season.) HitPredictor, a subsidiary of iHeartRadio, accurately forecasts hits prior to their release by playing them for a large online test audience in order to solicit their feedback.
  • Billboard Top 100 (BT100) combines point-of-sales sales data, download music numbers and Nielsen’s listening metrics. One result is that songs remain on the BT100 longer. As a result of this “the relative value of a hot has exploded”. Thus, the top 1% of recording artists earn 77% of all recorded music sales, while the top 10 selling songs have increased their capture of the market by 82% during the last decade. This is indeed a market where the revenue rich continue to get richer revenues.

II. Current Market Influences and Trends:

  • Wisdom of the Crowds:  Before the advent of big data, music company execs largely relied on their own instincts in choosing artists and products to promote. Now with the advent of these sophisticated apps and services, they are relying on upon a group f principles known as the wisdom of the crowds². Very simply stated, large and diverse groups of people, such as the web-wide millions using these services, is more likely to make more accurate decisions and forecasts than smaller groups and/or experts in the relevant field(s).
  • The Long Tail Effect:  As noted above, there is an intense and very small concentration among artists for whom big data and analytics is producing economic rewards.³
  • Social Media:  Some, but not all to the same degree, of these platforms are now the major drivers in marketing new artists and their new music. They might even be more influential than the tradition of drawing audiences to live concerts.
  • Radio Airplay:  This mainstream media, while maintaining its ongoing relevance in the music business, likewise replies just as heavily on all of the social media and data analytics in order to “connect all these dots”. The Wisdom of the Crowds also plays an integral part of radio programming.*
  • Overproduction of Repetitive and Bland Music:  Music industry people whom Thompson approached for this article expressed concern that the data-driven nature of today’s market is producing a “clustering” of music in different genres and, in turn, noticeable levels of sameness and copycat acts. Nonetheless, he further writes that research shows that listeners very often seek out familiar music they have heard many times before.
  • Effects Upon Musical Artists: Notwithstanding the prior point, musicians and composers are aware of this phenomenon but generally have limited its effects upon their creative output. As well, some will add variations and imperfections to their live performances in order to keep them sounding fresh. (X-ref to this August 11, 2014 Subway Fold Post entitled The Spirit of Rock and Roll Lives on Little Steven’s Underground Garage about how, among other things, this is one of the basic tenets of  Garage Rock.)

III. Ongoing Issues:

  • At the very heart of all of this activity is, as precisely framed by Thompson “What do people want to hear next?”
  • While the music business is significantly benefiting from the accuracy of all of this data and calculation, is it likewise producing “better”, more diverse and imaginative music for audiences to consume?

My own additional questions include:

  • Despite the Long Tail effect, are artists is the much longer end of the curve still accruing some demonstrable benefits from big data insofar are being heard by larger audiences online and in concert?
  • Based upon the monumental amounts of past, present and future data about music and the music industry, could deep learning and other artificial intelligence (AI) methods be used to produce genuine hit songs in multiple genres, without any human intervention? Alternatively, is the human touch always needed in the musical arts? If the answer ever turns out to be “not always”, what are the implications?
  • Could analytics and AI produce a new genre of music that is not necessarily a hybrid? That is, are there sounds, rhythms, arrangements, styles, tablatures and so on that have not yet emerged and can be entirely machine synthesized?
  • The article mentions that Led Zeppelin’s iconic Stairway to Heaven was never played much after its initial release and that it never landed on the BT100. As an experiment to test the validity and accuracy of today’s music data apps and services, what would happen if many such great hit were retroactively tested? Would any be proven to be hits that never should have occurred according to today’s tech and, conversely, are there obscure songs from years ago that would produce results indicating they should have been hits? Could or even should, such results be used to further fine tune, if not develop new musical data methods and metrics?
  • What other new opportunities will arise, based on this merger or art and science, for entrepreneurs, artists, talent scouts and agents, established music companies, and concert halls?

December 19, 2014 Update: 

Adding to the big data strategies and implementations for three more major music companies and their rosters of artists was a very informative report in the December 15, 2014 edition of The Wall Street Journal by Hannah Karp entitled Music Business Plays to Big Data’s Beat. (A subscription for the full text required a subscription to WSJonline.com, but the story also appeared in full on Nasdaq.com clickable here.) As described in detail in this report, Universal Music, Warner Music, and Sony Music have all created sophisticated systems to parse numerous data sources and apply customized analytics for planning and executing marketing campaigns.

Next for an alternative and somewhat retro approach, a veteran music retailer named Sal Nunziato wrote a piece on the Op Ed page of The New York Times on the very same day entitled Elegy for the ‘Suits’. He blamed the Internet more than the music labels for the current state of music where “anyone with a computer, a kazoo and an untuned guitar” can release their music  online regardless of its quality. Thus, the ‘suits’ he nostalgically misses were the music company execs who exerted  more controlled upon the quantity and quality of music available to the public.

________________________

1Right Off the Dial  (Faber & Faber, 2009), by Alec Foege, chronicled the causes and effects of the company’s rapid rise in the commercial radio industry. I found this to be an eye-opening and very informative account of rapid consolidation within a specific sector of the mainstream media. I recommend it to anyone interested in this company and topic.

2.  For a well reviewed and highly readable treatise on this, I also very highly recommend The Wisdom of
the Crowds by James Surowiecki (Doubleday, 2004). Also here is a Wikipedia page summarizing some of the main points of the book.

3.  The definitive and superlative book on one of the most interesting outgrowths of online commerce is The Long Tail by Chris Anderson (Hyperion, 2006). Furthermore, I also highly recommend his incredibly diverse and entertaining show on NPR called Studio 360.

*  I still that think Bruce Springsteen’s take on the state of music radio in 2007, Radio Nowhere, deserves a click and listening here. Besides, it’s an exhilarating rocker.

 

Startup is Visualizing and Interpreting Massive Quantities of Daily Online News Content

"News", Image by Mars Hill Church Seattle

“News”, Image by Mars Hill Church Seattle

Just scratching the surface, some recent Subway Fold posts have focused upon sophisticated efforts, scaling from startups to established companies, that analyze and then try to capitalize upon big data trends in finance¹, sports² , cities³, health care* and law**. Now comes a new report on the work of another interesting startup in this sector called Quid. As reported in a most engaging story posted on VentureBeat.com on November 27, 2014 entitled Quid’s Article-analyzing App Can Tell You Many Things — Like Why You Lost the Senate Race in Iowa by Jordan Novet, they are gathering up, indexing and generating interpretive and insightful visualizations for their clients using data drawn from more than a million online articles each day from more than 50,000 sources.

I highly recommend a full read of this story for all of the fascinating details and accompanying screen captures from these apps. As well, I suggest visiting and exploring Quid’s site for a fuller sense of their products, capabilities and clients. I will briefly recap some of the key points from this story. Furthermore, this article provides a timely opportunity to more closely tie together seven related Subway Fold posts.

The first example provided in the story concerns the firm’s production of a rather striking graphic charting the turn in polling numbers for a Democratic candidate running for an open Senate seat in Iowa following a campaign visit from Hillary Clinton. When Senator Clinton spoke in the state in support of the Democratic candidate, she address women’s issue. Based upon Quid’s analysis of the media coverage of this, the visit seemed to have helped the Republican candidate, a woman, more then the Democratic candidate, a man.

Politics aside, Quid’s main objective is to become a leading software firm in supporting corporate strategy for its clients in markets sectors including, among others, technology, finance and government.

Quid’s systems works by scooping up its source materials online and then distilling out specific “people, places, industries and keywords”. In turn, all of the articles are compared and processed against a specific query. In turn, the software creates visualizations where “clusters” and “anomalies” can be further examined. The analytics also assess relative word counts, magnitudes of links shared on social media, and mentions in blog posts and tweets. (X-ref again to this November 22, 2014 post entitled Minting New Big Data Types and Analytics for Investors that covers, among other startups, one called Dataminr that also extensively analyzes trends in Twitter usage and content data.)

The sample screens here demonstrate the app’s deep levels of detail in analytics and visualization. As part of the company’s demo for the author, he provided a query about companies involved in “deep learning”.  (X-ref to this August 14, 2014 Subway Fold post entitled Spotify Enhances Playlist Recommendations Processing with “Deep Learning” Technology concerning how deep learning is being used to, well, tune up Spotify’s music recommendation system.) As this is an area the writer is familiar with, while he did not find any unexpected names in companies provided in the result, he still found this reassuring insofar as this confirmed that he was aware of all of the key participants in this field.

My follow up questions include:

  • Would it be to Quid’s and/or Dataminr’s advantage(s) to cross-license some of their technology and provide supporting expertise for applying and deploying it and, if so, how?
  • Would Quid’s visualizations benefit if they were ported to 3-D virtual environments such as Hyve-3d where users might be able to “walk” through the data? (X-ref to this August 28,2014 Subway Fold post entitled Hyve-3D: A New 3D Immersive and Collaborative Design System.) That is, does elevating these visualizations from 2-D to 3-D add to the accessibility, accuracy and analytics of the results being sought? Would this be a function or multiple functions of the industry, corporate strategy and granularity of the data sets?
  • What other marketplaces, sciences, professions and products might benefit from Quid’s, Dataminr’s and Hyve-3D’s approaches that have not even been considered yet?

________________________________
1.  See this November 22, 2014 post entitled Minting New Big Data Types and Analytics for Investors.

2.  See this October 31, 2014 post entitled New Data Analytics and Video Tools Affecting Defensive Strategies in the NFL and NBA.

3.  See this October 24, 2014 post entitled “I Quant NY” Blog Analyzes Public Data Sets Released by New York City.

*  See this October 3, 2014 post entitled New Startups, Hacks and Conferences Focused Upon Health Data and Analytics .

**  See this August 8, 2014 post entitled New Visualization Service for US Patent and Trademark Data .

Music Visualizations and Visualizations About Music

There is a rare neurological condition called synesthesia where a person’s senses become linked in unusual ways. People affected by this may perceive, among other things, numbers as having specific colors or shapes as having tastes. Such people are termed “synesthetes” and this condition is not always seen as a detriment by them. This phenomenon has been studied and written about extensively in scientific literature as well as used as a plot element in fiction writing.

A number of inventive and strikingly beautiful visualizations of music have appeared online that can best be described as “sythesthetic”. That is, they are video creations to show what the music looks like to the graphics artists who have created them. Here is a sampling of some:

There are also two new tools available to provide innovative perspectives on:

All of these links produce big musical fun and I highly recommend clicking through to entertain and enlighten you eyes and ears.

Spotify Enhances Playlist Recommendations Processing with “Deep Learning” Technology

Music fans across the Web are now using music streaming services such as Spotify and YouTube for more than downloading sites according to a report in the July 3, 2014 edition of The New York Times entitled Downloads in Decline as Streamed Music Soars. Moreover, streaming is continuing to gain momentum for a number of reasons, not the least of which is that it is less expensive to access. Correspondingly, downloading sales are in decline. This article contains all of the details about this very significant shift in the online music marketplace, including the benefits to consumers and the concerns of musical artists.

A related follow up report entitled How Spotify is Working on Deep Learning to Improve Playlists was posted on Gigaom.com on August 5, 2014 that helps to explain how Spotify is trying to maintain its technical advantage. As reported here, Sander Dieleman, and intern at the company, has developed a method based upon deep learning (a branch of artificial intelligence), that parses large data sets in a new manner aimed at getting newer and lesser known recommendations into Spotify’s users’ playlist recommendations. Underlying this data processing is an examination of the acoustical properties of the user base’s song preferences.

The science behind the recommendation engines used so successfully by Spotify as well as Netflix and Amazon has come along many light years in it sophistication and accuracy since its earliest incarnations on the Web. For a comparative historical perspective, I also, well, recommend checking out an article from the December 1997 issue of WIRED magazine entitled Pattie about the work of Dr. Pattie Maes, a professor at MIT who founded (and later sold to Microsoft), a company called Firefly. Its technology was was one of the first efforts used to create software “agents” to scour the web for user-defined preferences for prices and products.

December 19, 2014 Update: 

Presenting an even stronger case that you-ain’t-seen-nothing-yet in this field was an engaging analysis of some still largely unseen developments in deep learning posted on December 15, 2014, on Gigaom.com entitled What We Read About Deep Learning is Just the Tip of the Iceberg by Derrick Harris. These include experimental systems being tested by the likes of Google, Facebook and Microsoft. As well, there were a series of intriguing presentations and demos at the recent Neural Information Processing Systems conference held in Montreal. As detailed here with a wealth of supporting links, many of these advanced systems and methods are expected to gain more press and publicity in 2015.