How Does a Band Become Famous? | Alex White | Next Big Sound | Pandora | Berklee Onsite

For almost as long as
I can remember, I’ve been fascinated with a very specific
and certain line of questioning– how does a band become famous? I mean, how the hell does a band
go from playing in their garage to headlining a nationwide tour? And can we possibly unlock
the black box of an artist on their trajectory from undiscovered
to the biggest artist in the world? And if we can, is it possible
to understand which artists are going to become popular next? But how did this all begin? Well, like many of you in this room, I
wanted to– at one point in my life– be a rock star, except
I very quickly learned that I didn’t like performing the songs
I’d written in front of other people. And that’s a fatal career
flaw if you’re a rock star. And so, if I couldn’t be the rock
star, I wanted to sign the rock stars. I wanted to be the record
mogul in the corner office with the gold plaques on
the wall signing bands. And so the summer after freshman
year, I interned at Motown Records in New York City, stapled
SoundScan reports together, have them on the executives
desk by the time they came in every Wednesday morning. And throughout that summer,
I would take five minutes whenever I could to talk to the senior
vice presidents in the corner offices with the gold plaques who
were signing these bands. I would ask each of
them how they got there, expecting them to say that they started
as an intern and worked their way up. And that’s what I was ready to do. Except none of them said that. They all said that they started a
record company, or a management firm, or a distribution company,
and got bought by Universal. And that’s how they got
to the corner office. That was honestly the first time I
ever considered starting something. And it was really in response to
realizing that where I wanted to be, everyone had started a record
label and got bought by universal. So the year is 2005, and
I started researching about starting a record label. If you know anything
about the music industry, 2005 was literally the worst time to
start a record label in 50 or 60 years, in the history of the
recorded music business. The economics were terrible. Napster and the pirating
sites were running rampant. And everyone was concerned about
the future of the recorded music business as a whole. But I had an idea. I wanted to create a website– instead of creating my own label– that let anyone create their
own fantasy record label and sign the bands they thought were
going to become popular to that label. And we would track how early
on you sign those bands. So if you signed Esperanza Spalding
second, she was signed 10,000 times, you get points based on how many
people signed the band after you did. And this was the idea that I
was obsessed with for years. I thought this idea was so good it
would revolutionize the entire music industry. And I was so afraid that if I
told anyone about this idea, that they would steal it and they would
change the music industry without me. And so for three years, I told
zero people about this idea. Yet every class I took– I’d be scribbling notes about
the idea and relating it back to the next big sound. Bought the domain name in 2005. And for three years, nothing happened. The first person I decided to tell
about Next Big Sound was David Hoffman. He’s now my co-founder. I took him out for breakfast three
years after I thought of this idea. I chose him, because he was
the only person at Northwestern I knew who knew how to build websites. And I presented the idea,
except I didn’t lay it out very cleanly like this. I had three years of kind of
Beautiful Mind scribblings. And I just kind of dumped all
that on him over breakfast. He said sounds kind of interesting. Good luck with that. I’ll see you later. Now this was the first person
I’d told about the idea that was going to change the music
industry, and he wasn’t interested. Well, he must be dumb or
something else must be wrong. So I called him back
over the winter break. He was in Hawaii. And I said, we have this
entrepreneurship class I know we’re both taking
this winter quarter. I want to be in your group and I
don’t care what project we do– even though I really
cared what project we did. He said fine, I’ll be in your group. And I managed to convince him and
Samir Rayani, and a fourth co-founder Jason Sosnovsky, who’s
no longer with us. He’s still alive, but just
not with the team anymore. [LAUGHTER] Managed to convince them
to do my idea, which was this fantasy sports for music idea. The name of the course, or the
professor, was Troy Henikoff. And it was winter quarter. And by the end of the quarter,
we’d shaped it together into something that was really
exciting and compelling. We had mock-ups. We had a basic version of the site. We had the game dynamics figured out. We had a business model. We had a 60-page business plan that
no one ever read except our professor. And everything was ready to go. It culminated with a pitch that
we thought went really well. We were approached afterwards by some
folks in the audience, really excited, handing us business cards. And we went out for drinks
afterwards, the four of us– Bat 17 in Evanston. And we all cheersed. And I told them that I wasn’t taking
classes spring quarter senior year. And my not-so-secret plan
the whole time had been to start this business spring quarter. Because everyone knows starting a
company can be done in one semester. [LAUGHTER] And I said, you guys want
to do this idea with me? And what will it take for you guys
to work on this company all summer? Now they had great internships lined up
at LinkedIn and AT&T and Sports Agency. And they said, if you can
pay for our the money we were going to make during the summer,
we’d much rather work on this with you and continue our idea. So we figured out it was about
$30,000 that we needed by June 1st, in order for them to back
out of their internships and cover summer costs for all of us. I had signed an employment
agreement with a consulting firm. I was supposed to start in
September, but that seemed ages away. And so the next morning after
waking up from the celebrations, I had to raise $30,000 by June 1st. Except I was 21 and knew no
one with any money whatsoever. All my friends were broke as hell. And parents had just been
putting us through college. So I called Troy. I said, Troy, great news. We’re going to keep doing this. I just need $30,000 by
June 1st, how do I do that? He’s like, did you get a couple
of business cards yesterday? I was like, yeah. He’s like, start there. So I emailed them. And basically, for several months, I
was putting on a baggy oversized suit and going down into Chicago– because
I thought you had to wear a suit if you wanted to raise any money– and talking to these people. And I was so excited about this idea. And every conversation
kind of ended with, it’s really interesting,
exciting, not for me, but you should talk
to x, y, and z person. And so each meeting led to
three more coffee meetings. Basically, for two months I was
bouncing around Chicago, literally– like jittering around Chicago, going
to five or six coffee meetings a day. I’d realized that at Starbucks, if I
was paying, passion ice tea was $2.35. That was the cheapest one I could get. If they were paying, and
they made that clear upfront, I would get a latte or
fancier kind of drink, because I knew I had several coffee
meetings ahead of me and had no money. And we got into this program
down in Champaign, Illinois, called IllinoisVENTURES 10, iVENTURES10. And that came with a $25,000 investment. That came through on June 2. So it wasn’t $30,000 on
June 1, it was $25,000 on June 2, which was close enough. Everyone quit their summer plans. And we piled in the car and we
moved down to Champaign, Illinois. Has anyone been to
Champaign, Illinois, before? Maybe people are from here. We got a four bedroom place
for $900 for the entire summer, which was like a few dollar– a passion ice tea a day, per person. [LAUGHTER] And that allowed us to keep a
majority of the money, which will come in later on in the story. So we’re down there, Samir had
just switched into computer science and wasn’t really clear about if we
could actually build this website. David had switched from
print design into web design, so was just getting his sea legs there. I’d never raised money or sold software
or sold advertising or done anything. So we were all just
figuring out as we went. I should say that right
before we went down, we had a validating experience where
we won the Northwestern Business Plan Competition. And so, we won the best undergrad
pitch and that came with $1,500 in one of those giant oversized checks. We had the check in our office. And there’s something about human
beings and seeing giant oversized checks where you can’t help but ask
if you’ve tried to cash that. So everyone who came into
our office asked that. And as soon as they did, we
would break into applause, and recognize that literally
every person does that. Anyway, we cashed the check. And in order to do that,
we needed a bank account. So in order to open a bank
account, you need a corporation. So we became The Next Big Sound, LLC– an Illinois LLC– which we spent
about a few hundred dollars forming. And then almost immediately
with the $25,000 we raised, we had to dissolve that corporation
and become a Delaware C corp. And we spent $1,000 of
the money paying someone to try to build the
core kernel of the site. And that didn’t end up
working or going very well. So that first $1,500 we made, spent
$500 forming and dissolving an LLC, and to a developer where it didn’t work. So off to a great start. We have the 25k, of which
$900 is for summer rent. I’m just orienting you
into our bank account. We moved down to Champaign. We build out the site. We’re ready to launch. I’ve been waiting for this, as you
know now, for over three years. The second we launch, I’m
waiting for the press to fly in, the TV appearances, the fame, the
fortune, and the music industry to hail us as the geniuses that are
changing the way that artists are identified and music industry works. And a funny thing happened
when we launched– nothing. Zero of those things happened. And no one seemed to notice. Now we got some press and our friends
and family were super excited. And they would tell us that they
loved it and used it all the time, even though we could see in the
numbers they weren’t really using it. [LAUGHTER] And the summer wrapped up,
and we moved back to Chicago. David and Samir had one
year left at Northwestern. I was needing to go to New York
to start my career and job. And that was where I found us– and again, we were at the same bar– said, what will it take for you guys to
quit Northwestern, me to quit my job? We decided if I could raise
$150,000, then we would do all that. I don’t know why we picked that number. So then my task became going around
Chicago, meeting with everyone I could, and raising $150,000 for a streaming
music site, in the fall of 2008, as a 21-year-old,
first-time entrepreneur. Anyone remember what
happened in the fall of 2008? Lehman Brothers collapsed and the
financial apocalypse began right after I called my consulting
firm and quit before I started, returned the $10,000 signing bonus,
and then the whole world went to hell. And so I moved in with
three good college friends who lived downtown
in Lincoln Park in Chicago, and slept on their couch. I was couch dude. And I would pack things up
and put it in the closet and didn’t really have much stuff. And my rent was $0. And we had the bank account that we
had from the summer with, fortunately, a lot of it left. And I had calculated it
out how long I could last. David and Samir were back in school. How long could I last with
the money that we had? I calculated it would cost $20 a week in
groceries, $40 every weekend for beer. [LAUGHTER] And doing that math, I could
make it to maybe April or May, if nothing happened. Now I got an opportunity to go out
on the road tour managing a band. This was the band. They’re called, Sing It Loud. And I was 21-years-old and doing
all of the social digital stuff. And as a 21-year-old, I was the
responsible adult on the team. They were all 18, 19, 20. We were opening for three other bands
on a two-month-long nationwide tour– Hit The Lights, Forever
the Sickest Kids, and Cobra Starship’s
2008 Sassy Back Tour. I had a one-way ticket to San Francisco. I was living on a couch,
so anything– you know, this adventure seemed like fun. I knew their manager from my time
at Northwestern booking bands. And he said, I’m looking for
someone to go do this crazy thing. Do you know anyone? And I said, I’ll do it. He’s like, you’re way too
qualified for this job. I was like, you don’t
really know what’s going on. I just quit my job. And the thought of going out on the
road seems like a fascinating learning experience. My daily expenses will be covered. And I’ll get to see the whole country. So I flew one-way to San Francisco. They played the Fillmore
West that night. Backstage, the rider was peanut
butter and jelly, chips and salsa, and a 12 pack of Budweiser
everywhere we went. There was a string of House of
Blues dates around the country. Went down the coast
line, Vegas on Halloween over through Texas and Florida,
and up to the eastern seaboard. And playing big iconic
awesome venues the whole way. And it was one of the most fun
two month periods of my life. And I never want to do it ever again. [LAUGHTER] And that is when the last shred of
me that wanted to be a rock star completely died. We were in a van and trailer. We were driving. We would sleep, you know, eight
of us to a hotel room, merch guy, day-to-day tour manager, and the band. And saw the whole country, slept very
sporadically on different things. And it was a fascinating experience. I have nothing but the utmost
respect for touring musicians, but it wasn’t something that
I could handle or ever want to do again, as I said. But it allowed me to see the
live side of the business. That was from November and
December, basically, of 2008. So 2009 comes around, and The
Next Big Sound gets written up in the New York Times, January 4, 2009. Everyone in my friend group and
family thought I was rich and famous, and we were off to the races. I had moved back to
the couch in Chicago, and we were running out of money. And we were about to shut
the company down that spring. And we applied to a program called
Techstars in Boulder, Colorado. Has anyone heard of Techstars before? There’s a program here
in Boston actually now. We’d applied the year before
to Techstars and didn’t get in. We applied with the same team and
the same idea, and got in this time. And it comes with $18,000 and
a three month accelerator boot camp-type thing, where they surround
you with mentorship and office space and legal assistance, and
a lot of blocking and tackling to get your business off the ground. There was 1,000 companies that applied. We were one of ten selected. We moved out to Boulder, Colorado. This journey is taking us from
Chicago to Champaign, Champaign back to Chicago, around the US for the
tour, out to Boulder, Colorado. I would have obviously
moved to Timbuktu, if there had been some financing
there for The Next Big Sound. And we tried everything that
we could think of at this point to get Next Big Sound
up and to the right, and get those metrics
and usage happening. And it was growing. We had thousands of bands and users,
but we had no way to make any money. And we were tired of waking up. We weren’t leaping out of bed like we
used to be and staying up all night. And I’d been pretty well
convinced after a year and a half of trying this,
that this idea was not going to immediately
revolutionize the music industry, as I had thought for three years prior. So on the first day of
Techstars, David Cohen, the executive director comes out
and gives a warm introduction. And he says, we invested in all
of you because you have what it takes to be great entrepreneurs. We didn’t invest in you just
purely based on your idea. In fact, we invested in some
of you in spite of your ideas. [LAUGHTER] And we’d been having this
debate, how the hell are we going to tell David Cohen and the
executive team that just selected us out of thousands of companies that
we’re switching our idea on day 1? And David said, we have what it
takes to be great entrepreneurs. And I was like, we’re 22 years old. How does anybody know? And so after he said that, that gave
us the confidence to go up to him. And we said, you know
that idea we applied with? We don’t want to do that anymore. Now it turns out, they really
actually had liked our idea. He said that’s fine. What are you switching to? We said, we have no idea. He said, that’s a huge problem. It’s like, we know. This is May 7, 2009. We’re in Boulder for
the three month program. And we don’t know what
we’re going to do. David and Samir were finishing
up– they’re still at Northwestern. It was a long year. And Northwestern’s on a
quarter system and very late. And so the midterms happened
end of May, and they didn’t walk at graduation till June 24. So for the first third
of the program at least, they’re flying back and forth to take
midterms and finals and graduate. We came back to the
question that we’d been obsessed with for a very long
time, and fascinated with. The most exciting part of The
Next Big Sound original idea was, if we can have a
million people signing the bands that they think are going
to become popular in the future, well guess what? That band is popular. And we can sell things back to them. And we can basically
have an army of A & R talent scouts out there,
scouring the bars, looking at the data around
who’s going to be popular next? So we said, we all
know who Lady Gaga is– and she was number one in 2008– but who was the first person
to listen to Lady Gaga? And what were the key inflection points? And can we unlock that
black box of how a band goes from 0 to 60, and from playing in dive
bars to headlining a global arena tour? And so we looked at the biggest website
at the time for music, Remember Myspace? So all it had was a play counter in
the top right corner of every page, but you couldn’t tell if those million
plays happened in the last week, month, quarter, two years, or how those
shifted or changed over time. So Samir built a crawler to hit Akon’s
Myspace page every five minutes, and we let that run overnight. Half a million plays between 2:00 AM
and 8:00 AM on Akon’s Myspace page on June 5, 2009. I knew from my time
in the music industry that no one was paying
attention to these numbers. And I knew from my time
on the road with the band that we were desperate to
understand, of all the things we’re doing from the
tour– which is very expensive– to press to
interviews, what, if anything, was moving the needle? And what we realized then was,
we couldn’t go backwards in time and see who the first person
to talk about Lady Gaga was, but if we started tracking
every artist online– June 5, 2009– in five
or seven years, we’d have five or seven
years of historical data that we could go back to
to understand and build models about how an
artist becomes popular, what the key inflection
points are along the way, and help the music
industry, and every artist, understand what they can do to reach
as wide an audience as they can, and build a sustainable business
and career for themselves. So we still celebrate June 5. We call it day-to-day– D-day– every year. And we now have coming up on– actually, it’s next week– eight years of historical
information that we can go back to for every artist. And so we have an insane amount
of data that we’ve captured. And I’m going to talk to you about some
of the things we’ve learned along here. But I just want to spend a few minutes
talking about the process from May 7, when we arrived at
Techstars, to June 5, when we collected that first data point. We didn’t know what we were going
to do, but we knew the problem area that we were fascinated with. And I had contacts from
the music industry, so every week, for those first
few weeks and through the duration of the program, I had a weekly
call with artist managers. And those started with
very big broad questions. What’s the biggest problem
that you’re facing? Now if you want to hear an
earful, ask a band manager what the biggest problem they’re facing is. But we started scratching the
surface with this Myspace data and with every five minutes. We would read it, and you would start
to see really interesting behavior that makes a ton of sense. Andrew Bird, I remember, had most
of his activity in the mornings. And Akon had most of his
activities at night, and especially on the weekends– kind
of the opposite behavior. So we spent a week or two building
these beautiful hand-drawn graphs. And we would send them
to 10 band managers. This is 2009– spend hours building
these, and got no response back. It’s like, is this not
incredibly valuable information? I would have killed
for this on the road, and imagine building your
artist’s career would be– this will be critical. Turns out, they were on
BlackBerrys and traveling. And in 2009, you couldn’t
open PDFs on BlackBerrys. So we said screw it. The next week, we just did plain text– Decemberists, 5,000 plays this
week, x 1,000 plays last week, percentage change. And we got back 10 out of 10– holy cow, this is awesome. Can you add in this person? Can you start tracking this band? Can you copy the band members? And that was how we kind of
knew we were on to something that was interesting and real. And I’m telling this story just to
illustrate how it, unequivocally, was not a blinding flash of insight,
Eureka moment, like you read about. Like, my first idea– that
was a blinding Eureka moment. For three years, I didn’t
tell anyone and then it didn’t change the music industry. This was very much a slow, methodical
process, a thousand conversations, and we are still in the middle
of digging in and understanding what the most useful thing we
can do is, in terms of helping artists make sense of all this data. So we were plugging away all through
those three months of Techstars. And we launched on Demo
Day, August 6, 2009. Turns out, the mentors have a list
that they send around each week around how the teams are doing– a good, bad, and ugly categorization. We were the youngest team
by five or ten years, and we were 10 out of 10 on the ugly
list every single week of the program– I found out after the fact– until the final pitch day,
when we leapt to number one. I don’t recommend this
strategy, but it worked for us. And we came out– for the weeks leading
up to Demo Day, we tried everything we could to raise money. We thought if we could raise $300,000,
that would be more money than we would ever need for this business. Just remember how hard it was– impossible– to raise
$150,000 a year earlier. So we asked for
$300,000, and we ended up closing a seed round of $900,000 led by
Foundry Group out in Boulder, Colorado. Jason Mendelson was on
the board for six years. And we ended up staying
in Colorado, specifically, to keep working with him every week as
we’d done through the entire Techstars program. And I found myself suddenly with
a bedroom, and a small salary, and all the trappings of a real business
that was getting off the ground. So the three of us David,
Samir, and I didn’t know anyone when we moved to Boulder. So we lived in a three bedroom
house on College Avenue. We hired our first employee. They didn’t know anyone. So we moved into a four bedroom house. We hired another employee, an engineer. He didn’t know anyone, so we
moved into a five bedroom house. We hired a third employee. They didn’t know anyone. We moved into a six bedroom house
near Chautauqua, right south of here. And for three years, we had
very little work life balance, living with your coworkers. We had a separate office, but we
were all obsessed with the work that we were doing. And we started selling this
product into the music industry, starting with managers and small labels. And ultimately signing three
of three major label deals. We built the business to about a million
a year in revenue and nine employees, but the last three did not live with us. [LAUGHTER] And in 2011, we went out
to raise our Series A. So we closed our seed
round in September of 2009, and announced our A Round
$6.5 million in January 2012. And I didn’t know until we went out that
everyone was shocked at how much we’d done on so little money. And it was because we were cheap to
keep alive and, from my couch days, I was able to visit friends and stay on
couches in every city that I went to. I know more about– I just bought a new couch– I knew everything about couches– for my apartment in New York. And we saved $50,000, in 2011
alone, just on my hotel– not going to hotels. I’m super cheap, if you can’t tell. That ended up being critical in
getting us to where we got to. So we raised $6.5 million A
Round from great investors, and moved nine people to New York City. We doubled to 18 people. We expanded into Next
Big Book and selling what we were doing to book publishers
like Harper-Collins and Macmillan. We started selling to brands,
so helping Pepsi and AmEx and other big brands make smarter music
investment decisions, and marketing and partnership decisions. And grew the business through
acquisition in July of 2015, without any further financing. So this category has changed a
lot since we started tracking data about the music industry. Switching gears into the
industry and category as a whole, and some of the things I’ve learned– every year, we released this
state of the industry report. And data analysis in the
music business was not a lane that existed, that I
could say I wanted to go into. Now it’s become– you know, the
course I’m teaching at NYU every year is Data Analysis in the Music Business. It’s the first grad level course around
data analysis in the music business. On conferences, they now have
data tracks or data panels. And this was not something
that was talked about regularly in the industry when we started. So a lot of what we were
doing is just beating the drum of the importance of using
data to make smarter marketing and sales decisions, and treating these numbers
as legitimate information sources. And so in 2012, we tracked 93 billion
new plays, 5.7 billion new fans, across all of the
sources we were tracking. The following year it was over 6 billion
fans across Facebook, Twitter, YouTube, SoundCloud, and others, and
223 billion new streams. We were trying to legitimize
this as a category, because one of the objections we would
get is– from the music industry– why the hell do I want to
know about my Myspace plays? I’m trying to sell records. It’s like, wait a minute. This is the way that people are
increasingly listening and interacting with artists. And we want to use this
information to help you better understand what you’re
going to sell in the future, and how all the marketing
and promotion you’re doing is relating to social
streaming and sales revenue. Following year, 434 billion new plays– and in the first half of
2015, we tracked 1 trillion– I’m not going to read
that number, it’s huge– across every stream on Pandora,
Rdio, Spotify, SoundCloud, Vevo, Vimeo, and YouTube. And we decided that was about enough. We didn’t need to keep beating the drum
that this was an important category. I think you’ve seen from
industry reports and trend lines that streaming really is starting to
move the needle in terms of revenue growth after several years of decline
and flat in the recorded music industry. So one trillion new plays. And most of these plays come
with geographic information, or demographic information. They’re daily. They’re really measurable. And how the hell do you start to
make sense of one trillion plays? You need a data engineering team
to store and manipulate the data. You need a data science
team to understand, and make heads or tails of all this information. You need a great front end and web
team to build the products that surface this information. And that’s really what
we’ve been wrestling with. So I wanted to share and spend the
rest of the time talking about some of the constructs we’ve used
at Next Big Sound at Pandora to make sense of all this information. And the foundational image that we’ve
leaned on has been the data pyramid. And this has been incredibly helpful for
customers, for partners, for employees, for investors, for board
members, for acquirers. Because when you talk about data
analysis in the music industry, people want to jump
all the way to the top and predict which artists
are going to be popular, predictive analytics around
revenue streams in the future. But I’m going to show you how none
of that is even possible or works, if you don’t have a rock
solid, data foundational layer, and the other
layers, leading up to it. So just to show you what we’re dealing
with, this is Kesha’s Twitter page. And her Twitter handle is @KeshaRose. And it used to be @keshasuxxx. And she just switched it one day. And we needed to update our system so
we weren’t pulling follower accounts, and @ mentions, and retweets
from the wrong account. And this is her Instagram, iiswhoiis. And no computer can really understand
that iiswhoiis and KeshaRose are the same thing. But when I’m Kesha’s
manager or record label, and I type in KeshaRose
in The Next Big Sound, I expect to see all the data for
Kesha right there in one place. This is one of the things
that we’ve invested millions of dollars in understanding
and building this network, and connection and understanding which
artist is which artist across all of these different networks. And these networks change all the time. That’s the next topic– and keeping up with the artists
themselves, because they are artists. So I’ll give you one example. Nicki Minaj– millions
of Twitter followers– one day, her graph goes
from millions to zero. So we freak out. We think it’s a data confidence issue. We test all our systems. We talk to the infrastructure
and database engineers, and are diagnosing this,
and tearing our hair out. And I was driving home from work. And I hear on the radio an
interview with Nicki Minaj, who had just deleted her Twitter account. [LAUGHTER] Now why she would have
done that, I still didn’t understand after the
interview, but she was sick of it, and just deleted it one day. Of course, it came
back a few weeks later, when the promotional
cycle started again. [LAUGHTER] But for a time,
that wasn’t a computer issue. That was entirely a human
being and an artist issue. So one bit of technology we’ve
built around this problem is called an endpoint
voting system, which means if her Twitter page
links to her Instagram. And we know that’s her Instagram. That’s a vote of confidence
that that’s the right one. Or if it links to the website or
the SoundCloud, that’s how we do it. We also have the best technology
at solving this in 2017– human beings, in a kind
of Mechanical Turk, contractor way, looking through the
ones where we are a little more suspect. So computers plus human
beings is how we solve this. And our data sources have
come and gone over time. Myspace was the first source we started
tracking, then iLike, then iMeme– I don’t know if they’re
even on there– a bunch of sites that don’t exist anymore. After we were acquired by Pandora,
some of Pandora’s competitors stopped feeding us data as well, which
is why some of these still exist, but we aren’t seeing data from them. And those are two examples from the
data foundational layer about just how difficult it is, and the
investment constantly required, to make sure that the rock
solid data foundation is there in order to do any of
the interesting things we want to do higher up the pyramid– so turning data into
information, those raw data feeds we have coming in
from the different partners. Want to show you about the
information layer, which you can think of like a dashboard. So in 2009, when we launched
the site on Techstars Demo Day, the site looked like this. If you’ve seen, it
looks an awful lot like this. And that was on purpose. That was our inspiration. We wanted it to be like
a for bands, where artists and managers
could put themselves in and any other comparable artist, and
see how they were comparing across. And use that to identify
opportunities that they could– maybe I should be on this other site,
or maybe they’re over-indexing here, and we should follow on
board, or our numbers are better or worse than these
artists on my label or on tour with. And it worked really well. And artists were very– and managers in the music industry– started emailing us, you know,
who the hell are you guys? We’re really interested
in this information. We’ve been trying to track it ourselves. We have interns every day going
through and tracking these numbers, but you automate it and deliver
it every week to our inbox. And one quick story on
the information layer. I was on an airplane, and seated
next to me were four gentlemen. And they were talking about how to
array a set of microphones in an arena. Said this is an
interesting conversation. Excuse me, gentlemen, I
couldn’t help but overhear. Are you guys musicians? And the gentlemen right next to
me said we’re traveling mistrels. He said, are you in a band? What does that mean? And he’s like– he was Kevin
Cronin from REO Speedwagon. And being the intrepid entrepreneur,
I said, I have a music analytics site. And I pulled it up on the
crappy airplane Wi-Fi. And I typed in REO Speedwagon. And his first question was, how
does that compare to Journey? [LAUGHTER] And humans and musicians
and the industry is really hungry for this information. He was fascinated. We typed in 100 other bands. And the most important
thing about all this, you know, if I just
tell you that you have 1,000 plays on x service yesterday,
you don’t know if that’s good, if that’s not good, how
that rates you compared to the entire rest of the ecosystem. And so a big part of our role, we feel,
is adding context to this information. And that can come in the percentiles
that I showed you in the profile page, or that can come in the comparing
yourself to another artist, or the modules at the top of
the page where you understand which networks are reacting, and how
that compares to what we would expect based on an audience of your size. So this was the first version. The second version of Next Big
Sound for about three years had a black background. And we wanted this to be
like a Bloomberg Terminal for the entertainment business. We even went so far– it’s not in
this one, because we killed it. We had a ticker going across
the bottom about your artist and how much new activity
they’d gone up or down. Really gimmicky and super
annoying, but that shows you how far we were willing to go. We studied Bloomberg Terminals. And we thought it was
a huge opportunity. This was our paid-for product. We called it Premier. And it came out in 2011. And was the most powerful suite of
music analytics software on the market. And this is really what our customers
were using for years and years. We were ingesting their proprietary
information, physical sales, streaming information, e-commerce
numbers, marrying that alongside all the social streaming
and video data we were tracking, and helping them understand
which artists were reacting and which marketing and
promotional events were driving the numbers that they cared about. We moved away from the
black background when we started doing a lot of customer visits. We moved to New York City to be closer
to our customers in the industry. We could’ve gone to LA, but
we wanted to live in New York and hire a lot of the data engineers
out of finance, and live there. And we were walking around, and we
see these printouts of these graphs on their table and stapled to the wall. Like, you were never
supposed to print that. Why did you print that? Well, I have a desktop and I
needed to bring it into a meeting. Oh. Or, I wanted to show
so-and-so, and I didn’t know how to do it via the account. And it turns out, we were
just crushing ink cartridges around the music industry. And it shows up horribly if you try
to print that, as you can imagine. And so we moved away from the
black background to the background that you see today. Where we’re going next is where we’ve
been after since the very beginning. It’s just taken us years and
years to climb the data pyramid. And that’s the language
we talk about internally is every time we tried to move up the
pyramid, something breaks lower down, and we have to shore up the
foundation and move again. So the third engineer we ever hired
was hired as an intelligence architect to specifically do the highest order
thing we could do with the data. And almost immediately, we’re
like, what are you working on? We have contractual obligations for
Sony or Universal that we have to do and they were pulled back in. Or, this is breaking and
we need your help here. And so wanted to talk a
little bit about where I believe BI and data
analytics is headed and where, specifically, where
we’re leading the charge. So this is in beta right now. These are automated
insights and intelligence we view as automated insights
and recommendations, at scale, for every artist in the world. And again, back to that
mission of helping artists understand and make use of
all the data that’s out there. So on September 19, Fifth
Harmony saw a large spike across Vevo, YouTube, and Wikipedia. And they’d posted a video– this was last September. So what this is is an
automated, what looks like a handwritten English
sentence paired with what looks like a hand-drawn graph. It was [? Carl ?] who
drew that, and then we’ve been automating
[? Carl ?] with his help. And then our best guess at the event
that drove that spike in activity. I’m going to flip through a
few examples of real alerts that I’ve gotten as we’re
beta testing this system. So on September 17,
deadmau5 added around 2,000 new Instagram
followers, 1.8 times more than the expected average
over the last 40 days. And he tweeted, you know what’s neat? My set in Tokyo and my set
in Atlanta are actually on the same day at the same time. I have invented time travel. So New Politics updated
their Spotify playlist, and they saw a spike in
Facebook Talking About metrics. That’s one of the things
that we can’t stress enough is how interrelated all
these different services are. You have consumers who, the first
time they hear about a new band, one might go to Spotify, one might go
to YouTube, one might go to Pandora, one might go to Wikipedia. And really capturing all
of this and understanding when I do x, what is my
audience going to do, and how do I welcome them
and greet them and make sure I’m optimizing all of
that as much as I possibly can? On September 19, The 1975 added
3,000 new Wikipedia page views. And they won a Mercury Prize. So people might have either been
looking up the Mercury Prize or the band themselves. And the last one is Migos. Fans get this tweet to 5,000 retweets
and we will release our mixtape. So a kind of gimmicky contest,
but that drove, of course, a spike in Twitter activity. And it worked. And one of the things our
most sophisticated customers are doing– there’s no
silver bullet around do this and we know this will happen. It’s all about rapid experimentation,
measurement, learning, iterating, and driving your career this way. And looking at what works
for other artists like you, maybe a step or two
ahead in your career, and using that to guide your
plans for release and marketing. So this section, I just
wanted to unpack a couple of pieces that go into intelligence. And the two pieces are artist stages,
which we looked at a little bit with Esperanza Spalding and
Branford Marsalis and John Mayer, and event modeling. And I think these will be
interesting to give you a peek into two of the building
blocks towards intelligence. There’s a lot more that we’re not
covering around deliverability and how we get the right alert to
the right person in the right format, and a whole bunch of other data. But again, the five stages are
undiscovered, promising, established, mainstream, and epic. We looked at every artist
across every metric that we track and clustered them
around these different areas. Now I’m going to walk
through, stage by stage, around what the kind of average
activity for that artist at that stage that we see. So in the undiscovered stage,
over the last six months, artists typically see
10 Facebook page Likes– give or take– 5 new Twitter followers,
200 YouTube channel views. And I think it goes without saying that
the vast, vast majority of the million plus artists that we track are
in the undiscovered category. And the exciting part, back to
the questions at the beginning, in reverse engineering the Billboard
charts, which undiscovered artist is going to be tomorrow’s epic artist? Promising 120 Facebook page Likes, 25
Twitter followers, 7,000 plus YouTube channel views. And I’m going to stop
at the end and give you a resource if you want to
dive deeper into these, but I’m just trying to give you an idea. These are three of the metrics that
we’re spotlighting, but keep in mind these stages incorporate Instagram
and Vevo and Pandora data, and others as well. At the established stage, 2,500 Facebook
page Likes, 550 plus Twitter followers, 6,500 YouTube channel views. The mainstream stage– 15,000 Facebook page Likes,
4,500 Twitter followers, 5.4 million YouTube views. You can see these numbers
really starting to add up. And lastly, these are the numbers– you guys can read– at the epic stage. And why do we need this information? Well, without this as
one of the building blocks for recommendations at scale– remember, we’re doing this in an
automatic fashion and not by hand– we would be recommending the top
events to every single artist. So we would say that every artist
should play the Super Bowl, and SNL, and the other events
that we know reliably draw a huge spike in social
streaming sales revenue. So we need to understand what works
for artists in your stage, what other artists are in that stage,
what are maybe a stage or two ahead, and hey, we’ve
never seen an artist go from undiscovered to mainstream
without first doing x, y, and z things. So this is a classification system
that we use behind the scenes to better understand
and meet our mission. And this is the URL if you’re interested
in seeing what the numbers are. We have an updated table that updates
automatically with the new information and what you need at each level
across each of the different sources at the bottom of this page. The event modeling is another component
that I wanted to briefly touch on. On March 29, Galactic appeared on Conan,
and saw a 335% spike in Wikipedia page views compared to the previous week, and
a 250% increase in iTunes album sales. And we looked at similar
artists who had appeared on Conan versus similar
artists on the same night who had not appeared on Conan,
just to make sure and prove it was statistically significant. What we found– this is in 2013– was that artists who
appeared on Conan see an average lift of these metrics
across the different sources. That’s the expected based on dozens and
dozens of artists that appeared there, and the spike in Wikipedia page views
on Conan compared to Fallon and Kimmel and Leno and Carson– obviously, we have
to update this as the musical chairs rotate– but no one had ever
quantified this before. And it made the cover
of Billboard Magazine, because Nielsen SoundScan was
a weekly CD sales information way to capture that. And when an artist appears on late night
TV, it’s normally during release week when a lot of things
are hitting all at once. And it’s hard to unpack and be
precise about what was causing what. And so you really need
that daily information in order to understand and
quantify the bump in activity. So back to our original question. That’s great, Alex. Which artists are going
to become popular next? You can hear it in the
name, Next Big Sound. And notice it’s Next Big
Sound and not song or artist. It’s really about what’s
the next big movement. But before we can get into who
is next, we need to understand what does popular even mean? One last question here in
our line of questioning. When does an artist break? This was my experience,
in 2012, going around to hundreds of very smart
people in the music industry asking them to define
mathematically what success meant. And I got a lot of blank
stares, and a lot of– you have a data scientist on your team. Why don’t you tell us
what success means? Which was interesting, and what we
found, after all these conversations, success, of course, is
different for different people. Some artists want their
successes playing at a venue where they have always idolized, or
being number one on a certain chart, or hitting a revenue goal, or
getting certified gold or platinum. So we built a model where you could
sub in different success criteria– and we patent this in 2014– around the likelihood for success. And you can sub in success criteria. So the one we built into the
product that I’m going to show you is what’s the likelihood
that within 12 months an artist who’s never appeared
on the Billboard 200 chart will make their first
ever appearance there. We call it our Billboard 200 score. We can debate about Billboard as a proxy
for success, but it was one of the– there’s 50 plus years of data. And as you can hear from that
questioning and fact pattern, we need a lot of precision when
we’re building these models. I don’t know what any of this
means though, by the way. Just trying to look smart. So this is a tool we built called Find. And it pulls up all
the data that we track and you can filter in a
bunch of different ways. And this is what we’re looking at
as new Pandora artist station ads, new in the last seven days. And these look like a lot of the
top hits and top artists overall. These are top spins,
new in the last 30 days. And almost a million plays
a day for “Shape of You.” Remember, we’re interested
in the next Ed Sheeran. If you’ve heard “Despacito,” the
song of the summer, potentially– that’s what’s driving these top two. And you can intersect and
triangulate data sources together. This, by the way, is not
available externally. It’s an internal tool that we use
at Pandora working with our clients. But I’m looking for artists
that are reacting virally on Pandora in terms of number of
spins, and have between 100,000 and a million page Likes on Facebook. And here’s the billboard score
and the top seven artists. So just what we’re looking
at here, 59.4% of the time, we see the momentum that Dua Lipa has
across all the services that we track. 59% of the time that artist will, within
12 months, hit the Billboard 200 chart. So it’s not deterministic–
this is a hit and I know it– world of the
music industry, traditionally. It’s a much more probabilistic model
saying, based on everything we know and all the new data flowing into
the system every day, it will update. So she actually has an album coming
out today on National Donut Day. It’s unrelated, but Dunkin’
Donuts told me that. Although it’s probably always
National Donut Day at Dunkin’ Donuts. But this will update
based on the reaction her new album has in the marketplace. And these aren’t cold numbers
of a bunch of statisticians just writing down numbers. These are real human beings
listening to the music and deciding, do they want to interact, do they
want to share it with their friends do they want to download
it to their collection, do they want to stream it on
repeat over and over again, or is it not connecting with them
in the way that other music does. That was what we just looked at
around our mathematical attempt to reverse engineer the Billboard
chart, and just some numbers for you. Our 2016 Artist to
Watch predictions list– we published in December of 2015– 12 of the 25 artists on
that list, 48% ended up hitting the Billboard 200
for the first time in 2016. And 12 of those songs, 48%,
appeared in the 100 top thumbs songs on Pandora in 2016. So in December, making the
prediction for the year ahead. So it’s not perfect, but it’s a
lot better than random guessing or even educated guessing. And we use this to release
Artist to Watch 2017, which is up and out of music festivals. And working closely with the
Pandora music industry team, combining the knowledge of the industry
with the data behind these predictions. So we’ve covered a ton of ground today. And thank you for your
time and attention. I just wanted to talk a little
bit about what I’ve learned, and what my journey
has been like, and what I’ve seen work, which is find an area
of the world that you can obsess about. And these things or whether you are
a musician, an entrepreneur, both, or whatever you’re up to. And finding that world and data
analysis and collective intelligence in the music industry was not
a sexy area when we began. I don’t know how sexy it is now,
but I think it’s a lot better. I’m biased. Remember for three years I didn’t
talk to anyone about my idea and I made no progress? That’s not how you change the world
is by keeping things to yourself, so start finding and talking about
what you’re doing and recording and what you’re working on
with the smartest people that are like-minded that you can find. And then, once you find that area
to obsess over and start talking, just run as fast as you possibly can. That’s all any of us can do. We can’t– as long as you’re
running as fast as you can, that’s all you can give. Thank you so much for your time
and attention this morning. [APPLAUSE] [MUSIC PLAYING]

, , , , , , , , , , , , , , , ,

Post navigation

9 thoughts on “How Does a Band Become Famous? | Alex White | Next Big Sound | Pandora | Berklee Onsite

  1. Lmao what a waste of time but thank you for the data info! I really can careless for this atm see you at the top

  2. speed it up to 1.5x – 2x.
    What are the helpful tips? Haven't found any yet… not wanting to waste the entire hour.

  3. Misleading title but the talk was somewhat interesting. His story about how he got to launch his idea and find success was great!

Leave a Reply

Your email address will not be published. Required fields are marked *