How A&Rs use data to scout and evaluate artists

By: Julie Knibbe

Published: 2020-10-26

This is the third installment in a five-part series about how music professionals use data across multiple industry sectors, including A&R, artist development, live events and royalty management.

As a disclaimer, this piece focuses on recorded-music A&R, and less so on A&R functions in publishing, live events and other sectors.

Navigate to the rest of the series below:

The digitization of the music industry has created many challenges for artists & repertoire (A&R) departments, who are responsible for scouting and signing new artists:

1. The amount of new artists to uncover seems practically infinite. About 12 million creators publish music worldwide. Figures from Alpha Data show that only 1% of them account for 90% of streams, meaning that the vast majority of artists are undiscovered.

2. The amount of available information to gauge an artist’s potential has increased by more than 10x over the past 10 years, thanks to the availability of analytics on social media and digital streaming platforms. Whereas there used to be a few commercial signals such as ticket sales, record sales and chart placement, A&Rs can now draw upon dozens of KPIs to evaluate an artist’s ability to acquire and retain an audience.

Thankfully, big-data technologies can help A&Rs as they work to scale their analysis to increasingly large amounts of data. In this article, I’ll cover how A&R use data analytics tools:

Before diving in, a note about concerns of algorithms replacing human tasks: In our case, data analytics tools are not meant to replace an A&R employee. Automated A&R analytics tools will come and go — remember Musicmetric, Asaii and Next Big Sound? — but with the major labels signing several hundred acts a year, A&R activity in general has only gotten more fervent amidst the abundance of data, not less.

This points to the futility of the “gut versus data” debate that frequently pervades the A&R world. As Chaz Jenkins, CCO at Chartmetric and former A&R, puts it to me: “A&R has always been about data.”

In the old days, an A&R scout would attend a gig and know that they should sign a band right then and there, based on their intuition. However, that intuition is likely informed by market data that they acquired through several years of experience — from local listening trends, to previous success stories in a given genre, to trendsetting venues and their respective capacities.

Such data points have historically been only implicit, living only in an A&R’s brain. Put another way, in many cases, intuition is just data that has yet to be unlocked and operationalized in a way that’s more accessible and scalable to others. The main difference with the current climate is that data points that were previously difficult to unlock are now explicit, standardized and broadly available, at a scale that is just too big for a single person to process all by themselves.

How A&R use data to scout artists

To source new talent, A&Rs combine online tools with their own referrals. On average, based on my conversations with A&Rs, only around 50% of talent discovery comes from digital tools, as network and word-of-mouth are still important in vetting an artist.

“We detect talents thanks to a mix of digital tools and human expertise,” Marie-Anne Robert, General Manager Marketing & Artist Services at Believe, tells me. “We are lucky to have an incredible tank of talented artists with TuneCore. However, we also think that something is more important than data detection and data prediction: Artist needs. So when reaching artists, we always focus on understanding what their needs and aspirations are. Two artists having similar trends on streaming platforms might have very different needs, and we need to bring them adapted answers; this is why Believe has developed a range of specific offers, from TuneCore to Artist Services.”

In fact, the quality of a given deal between an artist and a label, distributor or other partner can be boiled down to a balance of needs. Aside from artists having their own distinct requirements, every A&R is also looking for different criteria in an artist, depending on the strategy and business structure of their employer. An artist who is a good fit for one label might not be for the other.

For instance, small indie labels will sign artists for whom they have the right network to help them take their career to the next level; the majors will be more likely to source artists whose music has the most potential to scale to larger, more mainstream audiences. The extent to which these different parties rely on digital tools also depends on their distinct goals, on the resources they have to evaluate a certain number of artists and on the personalization options available in the tools themselves.

I distinguish between two main ways to scout emerging artists:

A. “Push” approach: Using charts to find breaking artists

In the “push” approach, analytics tools suggest lists of artists to watch, based on their own metrics and on their own set of monitored assets (e.g. charts, playlists and radio stations), with filtering options that A&Rs can customize.

Digital tools like Chartmetric, Instrumental and Soundcharts offer users the ability to construct “watchlists,” i.e. algorithmic charts based on a set of custom criteria. Here’s the example of Chartmetric’s A&R dashboard, which offers the option to order artists by popularity growth, fastest follower growth on Instagram, fastest monthly listener growth on Spotify and so forth, with additional filters for country and genre:

Such watchlists can give A&Rs hundreds of artists to look at, which is not actually making the former’s job much easier. To make these lists more relevant and easier to navigate, digital tools are adding more options to automatically filter out results based on a certain set of factors (e.g. previously evaluated artists, signed artists in a given label, a certain minimum threshold of social media growth).

Soundcharts recently launched custom “blocklists,” enabling A&Rs to hide songs, artists or even labels from their lists so that they can focus on artists they haven’t already heard of:

The American rapper Arizona Zervas emerged from such lists last fall. His song “Roxanne,” released in October 2019, quickly grew into a viral hit thanks to playlisting support on Spotify, becoming the first track by an unsigned, 100% independent artist to top Spotify’s U.S. Top 50 chart since early 2017. The feat helped him easily top several watchlists, including Instrumental’s blanket watchlist for best artist growth, and quickly started a bidding war among labels.

[Pictured above: Screenshot from Instrumental’s website demonstrating the kinds of data the tool monitors for artists like Arizona Zervas.]

B. “Pull” approach: Monitoring trusted trendsetters

In the “pull” approach, A&Rs define their own trusted channels to monitor new artist mentions. Independent of data analysis, A&Rs are already proficient and knowledgeable in their own markets, and closely watch the trendsetters in their local genre or scene.

Therefore, from the platform’s perspective, another approach to streamlining the A&R process is to help A&Rs automate the monitoring of their trusted sources:

Online digital tools have not completely streamlined such techniques, as they present a lot of technical challenges. Identifying artist mentions on videos, websites and blogs in particular is a tough problem to solve, because it can give a lot of false positives that add a lot of extra work to sort out. For instance, any artist with too common of a name like “Regard” or “Topic” would get a lot of irrelevant results.

On the other hand, analyzing radio and playlist programming is a lot easier. Chartmetric has already released a watchlist of artists breaking on over 1.4 million Spotify playlists; while extensive, the list of monitored playlists is not yet customizable. Giving more personalization options to A&Rs to tweak the inputs for their discovery and monitoring tools can help take A&R to the next level.

How A&Rs use data to evaluate artists

Discussions about A&R and data are often skewed towards the very first phase of the process, namely artist discovery. However, where the real craft of A&R happens is when they evaluate artists they’ve heard about and decide whether or not they are a good fit to sign them to a given deal.

Signing an artist is an investment, and A&Rs make them carefully. A lot of fact-checking happens to understand the story behind the numbers. In the evaluation process, A&Rs look for artists who have the ability to retain the attention of their audiences, and who present clear cases for why their audiences want to follow them.

As mentioned in our first piece about using data to pitch artists, music-industry professionals across the board focus more on growth and engagement than on raw numbers of views or followers. The primary KPIs they study are often in the form of ratios, such as the ratio of Instagram comments to followers, of YouTube channel subscribers to views or of Spotify followers to monthly listeners. In all these instances, the general consensus is that the higher the ratio, the better the engagement.

Professionals will be looking at whether an artist’s growth and engagement are organic and sustained, as well as how their engagement compares to those of similar artists. In practice, they will check if growth builds incrementally over time and is consistently spread out across networks. An artist with a genuine audience is usually growing evenly across social networks and streaming platforms.

[Pictured above: A good sign of organic and sustained growth. Arizona Zervas’ audience has been growing evenly across networks, with a common inflection point around December 2019.]

As the aforementioned Chaz Jenkins from Chartmetric tells me: “Many numbers are easy to fake, and it is just as easy to spot fake numbers. Tools like Chartmetric that combine data from various sources are very efficient at detecting fakes.” Given that we’re seeking consistent growth across networks, inorganic growth might look like sudden spikes in YouTube views or Instagram followers, but no corresponding increase in Spotify listeners.

For example, looking at the virtual influencer Lil Miquela’s audience growth earlier in her “career,” one notices an imbalance in growth among her various social media channels, with sudden spikes in 2017 and 2018 that do not spread across networks (pictured above). This implies that the company behind the project, Brud, used paid methods to boost her early followership to get their name out and to help attract venture-capital investment in virtual influencers.

What artists can do to “pass the A&R test”

A lot of elements of the A&R process fall out of artists’ control, especially if their artistic style doesn’t match a given label’s current roster or strategic goals. However, there are a few things artists can do to make sure their profiles stand out:

Conclusion: Predicting long-term success requires a human touch

Long-term success for an artist depends on many factors, including their own mental health, the quality of their team, their business structure, their budget, their creativity … many elements that data cannot measure, or for which information is not easily accessible.

When talking about data in A&R, the first thing that comes to mind is a hypothetical, magical tool that can predict an artist’s success. While predictive algorithms are now good at delivering short-term forecasts with a good level of confidence, they currently fall short on longer-term predictions.

This is because data is simply measuring the output and results of multiple intangible factors that cannot necessarily be expressed in numbers. Therefore, data alone isn’t enough to explain success, and is even less suited to forecast success on its own.

“Algorithms and automated tools can’t assess personal relationships very well, nor whether the artist manager is good, whether the agent really genuinely believes in the artist, whether the artist really wants to get involved in promotion,” Jenkins tells me. “These are all critical success factors in an artist’s career, and that makes the human touch indispensable.”

There’s also an important, and sometimes detrimental, relationship between data and risk. While A&R is getting more data-driven than ever before, data analytics tools can’t detect anything that is not already on their radar; in most cases, an artist needs to have successful data points around their music to get noticed in the first place.

This has led to a trend in the music industry where labels and other partners sign artists later and later in their careers. Artists that are signed today have often proven their potential independently first (with artist management companies taking on the bulk of the work in early “artist development”), then record labels come into play later on to help them take their careers to the next level.

This “cold start” problem could potentially present a major issue for the industry if A&R ultimately becomes too conservative. The more risk A&Rs take, the more leverage they ultimately get in breaking major artists.