How did Trippie Redd go from a virtual unknown to stardom? What took Logic from cult success to an artist who fills arenas? How do labels identify the superstars of tomorrow?

The answers to questions like these have traditionally involved factors like touring, endless promo runs, and lucky breaks. But now, artists and the labels that build their careers are finding a new way to move up the ranks: data.

Numbers and analysis have long been powerful tools in the music business (think of album sales, BDS counts of radio spins, and the Billboard charts that crunch those numbers). But with more of music consumption becoming digital, there is an ever-increasing amount of information available. Increasingly, numbers are being used to discover potential breakout songs, help advance artists’ careers, and even to determine the sound of the music itself.

Ankit Desai worked on analytics at Universal Music in Sweden before breaking off to run Snafu Records. Desai came to the record business after studying economics, which meant when he first arrived at Universal, “I was the only one with a background in data, analytics, and economics in a very creatively driven company.”

Analyzing information from Spotify, Desai discovered something interesting: a much larger than normal percentage of Logic’s “1-800-273-8255” listeners were adding the song to personal playlists and playing the song repeatedly. Those stats convinced the label to invest more marketing dollars into the song, and it paid off: the track ended up making it to No. 3 on the Billboard Hot 100 and it went quintuple platinum. 

Desai discovered a similar anomaly when he analyzed Tove Lo’s streaming data. One of her songs was making inroads to EDM fans in Sweden. So two months later, they released “Heroes (We Could Be)” with EDM artist Alesso, which went platinum and made it to No. 1 on Billboard’s U.S. Dance Club Songs chart. 

Data is becoming a primary way for labels and other tastemakers to find their next stars. Shav Garg is the co-founder of Indify, a company he calls a “music data platform.” Music pros use his company to figure out who the next hot artists are, and they were extremely early in noticing artists like Khalid, who the company first featured in the fall of 2015, nearly a year and a half before his debut album.

“We can go down to the level of, if something's really trending in Europe, we see it trending in Germany, but more specifically it's trending in Berlin.” - Jacob Fowler

Garg can tell a lot about an artist before hearing a note of music. If they are gaining a consistent or growing number of followers per week on YouTube, Instagram, and Spotify, it’s worth paying attention. That kind of “cross-platform growth,” as Garg calls it, is a key metric. Another one is engagement—how many likes is your content getting? Are people actually watching the videos you post?

Gerard Rice is also paying close attention to streaming and social media numbers. Rice is the commerce manager at the record label Same Plate Entertainment, a joint venture with Sony that is home to Marlon Craft and TOBi, and he gives weekly data reports to the label’s artists. If an artist on the label wants to do cooking videos (a real-life example he shared), does that overlap with the interests of their social media followers? Does an artist retweeting lots of fan praise of a new album actually cost them followers? (Yes, as it turns out). Given where an artist’s online following lives, where does it make most sense to route a tour? Which streaming service is giving the artist most of their plays? In that last example, Rice says, he and the artist’s team have to make a decision: double their efforts on that streaming platform, or put money into bumping up stats on the other ones?

One other way many labels are trying to juice their streaming numbers is with viral videos. When a song has a challenge attached to it, the increased attention leads to greatly increased streams, and of course more money. But how does a song go viral in the first place? Enter Devain Doolaramani. The college senior manages viral video makers like The Bailey Bakery and Nick & Sienna, most of whom have their biggest audience on the platform TikTok. Record labels approach Doolaramani with requests for his clients to make videos that use their artist’s tracks. Sometimes the labels just want him to enhance a trend that already exists, but other times they ask him to make the whole thing up from scratch.

His biggest success came from the former. Someone had created a lyric prank video using Russ’ “Civil War” several weeks prior to Doolaramani starting his campaign. “So we decided to jump on that idea and it ended up absolutely tearing it up,” he remembers. There were around 340,000 videos made on TikTok, he reveals, and the challenge spread to Instagram and Twitter. 

The impact was immediate. Streams on the song, a month old at the start of the campaign, shot up from 20 million to around 48 million. And many YouTube comments referred to the prank videos. 

Milo Stokes has helped guide Trippie Redd’s career from its earliest stages. (Many reports refer to Milo as Trippie’s manager, but “at this point, I'm more of a partner,” he says. “I don't really like titles, I like relationships.”) Stokes discovered Trippie when the rapper had only 9,000 social media followers. He fell in love with Trippie’s sound and look, and built a team for the rapper out in Los Angeles. Stokes used the company he co-founded, Create Music Group, to look for new opportunities, and they began by working with Vine stars like Nash Grier, but quickly moved into the music space, helping artists collect YouTube royalties.

Stokes’ experience with YouTube helped him see who he needed to connect with to help Trippie get attention. “Early on, we worked with artists such as Famous Dex and even Cole Bennett,” Stokes explains. “These were different channels that were really tapped into their fanbase and that subculture that was brewing up.” Even Trippie’s collaboration with a pre-”Gummo” 6ix9ine, “Poles1469,” was in part driven by these same concerns. “6ix9ine, that was an artist that we all thought grabbed our attention and had potential to do a lot of cool stuff in the market as well,” he says.

“It might be weird to say, but looking at data is a very creative thing,” Stokes explains. “You can see analytics, but if you have a perspective or if you are really asking questions as you're looking, you can see patterns and movements of audiences and how consumers are behaving. [It’s about] being creative and crafty on how you can be a part of that stream and get in that mix with your content.”

Data also has implications in areas beyond individual artists’ careers. Crunching the numbers shows, for example, that Latin artists over-perform on Facebook. So if you’re marketing a Latin artist, or are a Latin artist, Facebook would be the place to spend your dough. 

“New platforms emerge, and people’s appetite for data changes as well,” says Sung Cho, founder of the music data analytics company Chartmetric. “A great example is Twitter. People cared so much about Twitter in the past. Now, not many people care about that, and people care less about Facebook friend counts as well.”

Chartmetric’s “music data storyteller” Jason Joven says that data can demonstrate whether promotional efforts such as TV appearances are working.

“[Consumers] don't even know who they're listening to half the time. They're on Today's Hits on Apple Music or whatever. So when we see a spike in Wikipedia views that correlates with a performance on [The Tonight Show Starring] Jimmy Fallon, that's usually a sign for labels, managers, and artists to realize: ‘Something about what I said in that interview made people really curious about me.’”

Flipp Dinero’s “Leave Me Alone” received a sudden burst of interest in the spring, leaving the employees of the track’s distributor, The Orchard, scrambling. After some searching, they discovered that the culprit was Odell Beckham Jr, who had posted a short video of himself dancing to the track on Instagram. 

“That's really difficult to find,” explains Samantha Moore, The Orchard’s director of analytics. “On an Instagram Story, that data isn't as robust as some of the other sources. We can see on Spotify that obviously something happened and immediately react. But we have to do a lot of fact-finding missions to find a specific reason.”

Sammy Pisano, the company’s senior manager of sales analytics, remembers a similar event. Luh Kel’s “Wrong” was getting a lot of plays in the Philippines, seemingly out of nowhere. They discovered that a cover version of the song on YouTube by a Phiippino duo led to people in that region to search for the original. The marketing team jumped on the opportunity, promoted the song, and it ended up as one of the Philippines’ top songs on Spotify.

Being able to find and dig into data has led to artists and labels to attempt to capitalize on potentially viral moments. But that same data also allows them to find out who’s listening. Information from streaming services is anonymized, but artists, labels, and distributors can still find valuable insights.

“We can go down to the level of, if something's really trending in Europe, we see it trending in Germany, but more specifically it's trending in Berlin,” explains Jacob Fowler, The Orchard’s SVP of Engineering & Product. 

Demographics are also crucial. Older audiences are “slightly less adventurous,” Moore explains. So if an artist’s listeners skew that way, they may need more outreach than if the fanbase is young. 

One thing almost everyone makes sure to point out is that whatever data’s uses, it shouldn’t enter the equation until after the music is made. Otherwise, as Jason Joven points out, “You’re going to end up in this loop [where] everything sounds like Drake and Ariana Grande,” since everyone would only want to chase what’s already popular. 

“I feel like people will start to use that information to try to craft songs.” - Gerard Rice

Samantha Moore agrees. “I really have an adverse reaction to trying to use data to change content,” she explains. So while numbers may help people in the music business find new audiences for songs, or capitalize on sudden bursts of attention, even the people on the front lines of that process agree that their job begins after a song is created. At least for now. As data comes from more and more sources (TikTok and other new social media platforms, country or region-specific streaming services, and virtual assistants like Amazon Alexa), the potential to use it earlier in the process may become too enticing to resist. Moore cops to doing “large scale analysis around what genres might work well together,” though she’s adamant that doesn’t mean that collaborations are designed as a result. But they might be soon. 

And there’s a precedent for that. After all, that factored in to how Netflix designed House of Cards: they knew that their subscribers liked David Fincher movies, films with Kevin Spacey, and the original British House of Cards. They looked at data to figure out that the overlap from those three categories virtually guaranteed a hit. It’s not impossible to imagine that a similarly designed collaboration between, say, a reggaeton artist and a K-pop artist could be thought up in a boardroom and then brought to life. The line between using data to promote and using it to create may not stay clear forever.

In fact, Gerard Rice thinks that line may have already been breached. He told me that the next trend in data will be “measuring the song analytically”—that is, looking at the constituent musical elements of a song (Pandora, for example, already does an elaborate version of this). From there, he imagines people analyzing what kinds of songs end up on which playlists (or are liked by different types of fans), and designing new tracks with the desired elements.

“I feel like that's going to be a thing in the next few years, if it’s not already,” he says. “I feel like people will start to use that information to try to craft songs. Definitely not all artists will do that, but I could imagine that happening.”