<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Josh Hayes on Medium]]></title>
        <description><![CDATA[Stories by Josh Hayes on Medium]]></description>
        <link>https://medium.com/@josh_7639?source=rss-7a06e0067f24------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*rkrcNOPM8Vlu0YKU-JZCJw.jpeg</url>
            <title>Stories by Josh Hayes on Medium</title>
            <link>https://medium.com/@josh_7639?source=rss-7a06e0067f24------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 07:02:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@josh_7639/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[The Next Era of A&R Tools]]></title>
            <link>https://medium.com/chartmetric-blog/the-next-era-of-a-r-tools-a20c1b6ad5e3?source=rss-7a06e0067f24------2</link>
            <guid isPermaLink="false">https://medium.com/p/a20c1b6ad5e3</guid>
            <category><![CDATA[product-updates]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[music]]></category>
            <category><![CDATA[music-business]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Josh Hayes]]></dc:creator>
            <pubDate>Wed, 09 Oct 2019 18:13:19 GMT</pubDate>
            <atom:updated>2019-10-09T22:39:36.159Z</atom:updated>
            <content:encoded><![CDATA[<p>Chartmetric’s new A&amp;R tools leverage machine learning to shortlist 1.7M musicians into the high-potential artists that interest you most.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*qCmqniKmZQHmx-mJ" /><figcaption>Is it a pair of binoculars? Is it a robot? Either way, it is certainly a metaphor. (Photo by <a href="https://unsplash.com/@markusspiske?utm_source=medium&amp;utm_medium=referral">Markus Spiske</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><p>One of the most inspiring parts of the music industry is when passionate, new artists get their break, make it big, and express their sound. For many industry-people, the joy and excitement around that journey is what got them into the music industry in the first place. For others, that is what fuels them day by day.</p><p>The digital streaming era has made it easier than ever for new artists to make and release music. At the same time, it has made it harder than ever for managers, marketers, and A&amp;R scouts to keep track of what is happening across the entire digital music landscape. Increasing amounts of data on unprecedented amounts of aspiring artists make it hard to identify the stand-outs.</p><p>Finding a rising star today is like Where’s Waldo on steroids.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ejs-XtVZSvbI1ww8" /><figcaption>With artists like Lil Nas X redefining the music industry, identifying the next star is a daunting, but exciting, challenge. (Photo Credit: <a href="https://peopledotcom.files.wordpress.com/2019/07/lil-nas-x.jpg?crop=0px%2C237px%2C1800px%2C946px&amp;resize=1200%2C630">Getty Images</a>)</figcaption></figure><p>A&amp;R scouts and managers know that identifying the next big artist requires industry savvy, forward thinking, and hard work. Doing so requires skills that cannot be replicated with artificial intelligence (A.I.). We are nowhere near a point where A.I. could replace the talented people in A&amp;R.</p><p>But A.I. still has a big role to play. One of the main goals of artificial intelligence is to tackle bogglingly large-scale problems. And a “bogglingly large-scale problem” is exactly what the music industry is facing right now. For example, on Spotify alone <a href="https://www.musicbusinessworldwide.com/nearly-40000-tracks-are-now-being-added-to-spotify-every-single-day/">40,000 new tracks</a> are added every single day. That says nothing about the talent emerging on Soundcloud, Apple Music, YouTube, Deezer, Amazon, Instagram, and TikTok.</p><blockquote>Chartmetric’s goal is to build tools to empower A&amp;R scouts, not replace them.</blockquote><blockquote>We don’t want to tell you who <strong>the</strong> next big artist is. We want to help you focus your expertise, time, and energy in the most promising directions.</blockquote><h4>“What does A&amp;R empowered with machine learning look like?”</h4><p>Imagine if a tool could help you filter the overwhelming landscape of aspiring artists down to a “short list” of the artists that have what it takes to succeed. Imagine further that you can choose to cast either a wider or a narrower net. You could decide how much of an artist’s history to take into account: Do you look for artists that are literally just starting out? Or do you target musicians that have already started showing a consistent ability to release tracks and gain listeners?</p><p>Whatever strategy you favor, our new Predictive A&amp;R tools make that hypothetical a reality, and you can use them right now, over at <a href="http://www.chartmetric.com/">Chartmetric.com</a>!</p><blockquote>These tools don’t just look back at the past, they make active predictions about the future.</blockquote><p>Our new A&amp;R tools filter through 1.7M artists to return the top results for artists that are most likely to have a big career break in the next 7 days.</p><p>On our site currently, you can customize the type and breadth of your search, sort by region and genre, and filter artists that are signed by major labels (in case you’re looking to sign talent or want to find an up-and-comer to collaborate with). Checking out an artist’s full profile, including direct links to their music, is always just one-click away via Chartmetric’s artist page.</p><p>You can see what it looks like here:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F365399968%3Fapp_id%3D122963&amp;dntp=1&amp;url=https%3A%2F%2Fvimeo.com%2F365399968&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F821126872_1280.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1920" height="880" frameborder="0" scrolling="no"><a href="https://medium.com/media/9825fdfe1e8a6fe553a595a0ab1c4c55/href">https://medium.com/media/9825fdfe1e8a6fe553a595a0ab1c4c55/href</a></iframe><p>Alright, so that seems pretty cool. But how did we actually build it? How does it work?</p><h4>Per usual, even the most seemingly complex machine learning processes can be broken down into fairly simple and logical steps.</h4><ol><li><strong>Define Successes: </strong>We have to identify artists that have “made it” so that we can learn from them.</li><li><strong>Identify Common Patterns: </strong>We identify patterns shared within those artists’ histories. You can also think of this as establishing “profiles” for what breaking artists look like.</li><li><strong>Compare New Artists to Those Success-Patterns: </strong>After that, we can set up a pipeline to grab artists’ recent stats and see when they start to exhibit patterns like the artists we saw that already “broke.”</li></ol><p>That short description belies a lot of work, of course. So let’s briefly go through each of those steps to better understand what the process looks like.</p><h4>Step One: Define “Successful” Artists</h4><p>The very first thing we need to do is create a category of artists that we want to identify with our model.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*vxUu3UpEQoRJCL_e" /><figcaption>Even though I’ve never played a sold-out show, my mom still says I’m a success. Because we had to draw a line somewhere, we did not consult my mom when defining successful artists. (Photo by <a href="https://unsplash.com/@roccocaruso?utm_source=medium&amp;utm_medium=referral">Rocco Caruso</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><p>There’s a famous, and often cited, example of a ML project to classify images with tanks from images that did not have tanks. By inadvertently feeding the model pictures of tanks taken on sunny days and tankless pictures from non-sunny days the researchers had actually trained a model to distinguish sunny days from non-sunny days. For a more recent but less evocative example, you can read about how this <a href="https://www.kaggle.com/c/noaa-right-whale-recognition/leaderboard">2nd place Kaggle winner</a> realized that while building a whale-classifier he had actually inadvertently built an <a href="http://felixlaumon.github.io/2015/01/08/kaggle-right-whale.html">ocean-waves-classifier</a> instead. Oops!</p><p>The takeaway? ML models will lock on to the commonalities of what you feed them. So you need to be incredibly thoughtful, rigorous, and critical when working with your data.</p><blockquote>This is the hardest, and most crucial, stage of the entire process. The quality of your machine learning model’s results is a direct consequence of the quality of the data you put into it.</blockquote><p>So we need to pick “successful” artists with a lot of care. I’ll skip a lot of the nitty-gritty here because it gets technical quickly and we could write a full-length article just on this process… but here are the major takeaways:</p><ul><li>We wanted to get a sense of the likelihood for an artist to break, so classifiers seemed like the way to go. That means we had to break all artists into just two groups: Successes and Still-Aspiring-Artists.</li><li>We looked at all artist histories that we have in our database over the past 3 years. It didn’t matter if the artist broke 4 months ago or 2 years ago. Everyone was welcome to compete.</li><li>We exclude artists that are already successful based on a number of factors. We don’t want our model to “discover” Ariana Grande. That would be about as revolutionary as saying you “discovered” the sun… No one would be impressed.</li><li>We set a numeric threshold that we considered to be a notable-level of success (“Upper Threshold”). Any artist that started below a certain point (“Lower Threshold”) and then passed the Upper Threshold within a Specified Time Frame made it through the first round of filtering.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/841/1*tdarin4Ak_hqDfnQ8zL7Qg.jpeg" /><figcaption>Only artists that started below the lower threshold and went above the upper threshold made it through the initial filters.</figcaption></figure><ul><li>We wanted artists that “made it” and then <strong>stayed</strong> there. The second stage of filtering excluded a number of undesirable patterns: “15-minute of famers” that fell back into obscurity quickly, artists that rose and fell repeatedly, and many more.</li><li>Because the best upper-threshold for success is subjective (largely varying based on the type of artists you want to find), we created multiple versions. Users are able to pick the thresholds that they feel are the most effective for their use-case.</li></ul><h4>Step Two: Identify Successful Artists’ Shared Patterns</h4><p>Whenever we identified a success case, we looked at the 6 months of historical data leading up to that point. We gathered together a number of performance metrics from Spotify, Instagram, and each artist’s release schedule as well.</p><p>The historical trends tended to look something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-_7q6O6FLLrcNtH2fWlNkg.png" /><figcaption>An example Instagram history from an artist that grew from obscurity into stardom.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fGTAE6f1y7UWn4ZIToGDdA.png" /><figcaption>An example Spotify history from an artist during their big career break. Notice the track release that kept snowballing to more and more listens.</figcaption></figure><p>The “secret sauce” to make this method excel is that we do not want to be so strict in identifying success cases that we end up requiring artists to conform to restrictive paths to success. If we did that, we would essentially become so narrow-minded that we would ignore up-and-comers just because we didn’t like <strong>how </strong>they were growing. In today’s shifting music market, and with so many new emerging mediums like TikTok, we want to be as open-minded as possible.</p><p>For example, take a look at the difference between how Billie Eilish and Lil Nas X grew their careers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*40rDuLSdGuiQpCIhd3Yvdg.png" /><figcaption>Lil Nas X almost grew as much in one month as Billie Eilish did in one year. Yet both are undeniably mega-successes.</figcaption></figure><p>Already we can see that we have a lot to learn from the stories of both Lil Nas X and Billie Eilish. They had very different paths to success. Similarly, we probably have a lot to learn from hundreds of thousands of other artists’ experiences.</p><p>The most accurate way to make predictions is to look at a wide range of cases, and then start paying attention to which patterns most reliably lead to positive outcomes. That way when we see new cases, we can compare them to what we have already seen and then assess how confident we are about making a prediction.</p><p>That is exactly how our machine learning models operate.</p><p>We spent a lot of time and energy identifying between <strong>8,000–10,000</strong> instances of artists’ career break-outs to learn from. But we also compared to <strong>tens of thousands</strong> of cases where artists did not have career break-outs, to find the patterns that distinguished each group.</p><blockquote>Chartmetric’s database holds more than 271 million data-points on artist streaming data going back to 2017.</blockquote><blockquote>We drew on all of it.</blockquote><h4>Step Three: Compare Patterns of Current Artists</h4><p>Most simply, we start by looking at a current artist. Then we think back on the masses of artists that we’ve learned about, and compare how similar they all are.</p><p>Out of all the artists that had patterns similar to this current artist, how many of them had a big career break in the near future? 10% of them? 100% of them? Or was it more 50/50? The more artists that went on to be successful, the more confident we can be that this artist will too.</p><p>Here is an illustrative example:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1012/0*rFR3aSmqETRmBYIa.gif" /><figcaption>Our models identify artists like this one, who mirror the success patterns we’ve seen in the past. A surge in listenership immediately after a new release is a great sign.</figcaption></figure><p>Based on what we’ve seen in the past, we can give each artist a symbolic “thumbs up” or “thumbs down” and we can also say how confident we are in our assessment.</p><p>The graph above looks pretty promising, right? The recent release as well as a surge in listenership gives us more confidence that we can give this artist that symbolic “thumbs up.”</p><p>Let’s compare that to another artist with stats that are less positive:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DIpXTFraaCKH0lTQNRzjVg.png" /><figcaption>This artist could go either way — recapturing their momentum, or losing it.</figcaption></figure><p>This artist is at a cross-roads of sorts. Lots of artists peak and then fall back. Other artists have moments where their stats level off, but then they release again, or tour, or do a collaboration and then their stats take off and rise again even higher.</p><p>This artist could go either way. So if we were smart, we would have more reservations about predicting one way or the other. Because our models order artists by how certain we are that they could continue to rise, in-between cases like this one get pushed further down the list. Eventually, if our certainty is low enough then the artist gets dropped off the list completely.</p><blockquote>One of the most handy features of a classification model is that we <strong>intelligently</strong> infer the certainty of our predictions based on real-world data — not on anecdotal evidence or our personal biases.</blockquote><p>That’s essentially the definition of data-driven decision making.</p><h4>Check Out Brand New Up-and-Comers!</h4><p>We deploy these models every single day to filter the overwhelming number of artists into a shortlist of artists that are making waves in the music scene.</p><p>You can further refine each day’s results by filtering on region and genre, as well as excluding or including artists signed by major labels. Beyond that, it is easy to learn more about any artist with one click by heading to our artist pages, or you can go directly to their social media and streaming platforms.</p><p>Just remember, we did not build these tools to <strong>tell you</strong> which artist is going to be the next big hit. We built these tools to <strong>help </strong>filter the millions of artists in the way <strong>you</strong> think is best so that <strong>you can discover new sounds, new talent, and new stories.</strong></p><p><em>Predictive A&amp;R is a paid feature, but you can grab a </em><strong><em>free trial </em></strong><em>for Chartmetric to explore many of our services </em><a href="https://app.chartmetric.com/signup"><em>here</em></a><em>.</em></p><p><em>Or if you dig data as much as we do, then let </em><a href="https://blog.chartmetric.io/@jasonjoven"><em>Jason Joven</em></a><em> keep you in the loop on the music industry’s latest data gems by tuning into our podcast: </em><a href="https://chartmetric.transistor.fm/episodes"><em>Your Daily Data Dump from Chartmetric</em></a><em>.</em></p><p><em>Additionally, we always love to get feedback from you at hi@chartmetric.com.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a20c1b6ad5e3" width="1" height="1" alt=""><hr><p><a href="https://medium.com/chartmetric-blog/the-next-era-of-a-r-tools-a20c1b6ad5e3">The Next Era of A&amp;R Tools</a> was originally published in <a href="https://medium.com/chartmetric-blog">Chartmetric</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Using AI to Make Meaningful Comparisons Between Musicians]]></title>
            <link>https://medium.com/chartmetric-blog/using-ai-to-make-meaningful-comparisons-between-musicians-4bd5a0e536cc?source=rss-7a06e0067f24------2</link>
            <guid isPermaLink="false">https://medium.com/p/4bd5a0e536cc</guid>
            <category><![CDATA[music]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[product-updates]]></category>
            <dc:creator><![CDATA[Josh Hayes]]></dc:creator>
            <pubDate>Sun, 30 Jun 2019 01:50:24 GMT</pubDate>
            <atom:updated>2019-10-09T18:30:25.015Z</atom:updated>
            <content:encoded><![CDATA[<p>How unsupervised machine learning and natural language processing (NLP) can identify artists’ genre neighbors</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4dyX6D-jv9BIUBbR" /><figcaption>Similarity and Difference (Photo by <a href="https://unsplash.com/@adairwen?utm_source=medium&amp;utm_medium=referral">Earvin Huang</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><p>I love comedy. It is fun and lighthearted, but surprisingly often the jokes highlight deeper insights. Let me give you an example.</p><p>Flight of the Conchords, a music-comedy duo made up of Jemaine Clement and Bret McKenzie, opened their <a href="https://youtu.be/qFcj24A6W3g">HBO Live special in London</a> by introducing themselves and recapping their career trajectory.</p><blockquote>Bret starts, “We are one of the biggest bands in New Zealand now…”</blockquote><blockquote>Before the crowd can cheer, Jemaine interrupts, “…in terms of the number of members.”</blockquote><p>The result is the first belly laugh of the night. It’s a great joke. Punchy. Self-deprecating. Ends with a surprise twist.</p><p>More than that, the joke also hits to a important but easily overlooked insight:</p><blockquote>How you compare musicians drastically influences what you see.</blockquote><p>People from all over the music industry have to decide how to compare musicians every day. How one chooses groups to draw comparisons against can radically change the insights and take aways.</p><p>Identifying a comparison group is never more tricky than when it comes to music genres. Where does one genre stop and another begin?</p><p>To try to visualize the problem, we pulled all the top artists from our database and measured the co-occurrences of genre tags. We put the results in a network graph, where each node is a genre and the links between nodes are co-occurrences. The more times that genres show up together, the stronger the nodes are attracted to each other. Try it out, with just a small sample of genres:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Fjosh_chartmetric%2Fembed%2Fpreview%2FRzLjME%3Fheight%3D600%26slug-hash%3DRzLjME%26default-tabs%3Djs%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;url=https%3A%2F%2Fcodepen.io%2Fjosh_chartmetric%2Fpen%2FRzLjME&amp;image=https%3A%2F%2Fscreenshot.codepen.io%2F3341989.RzLjME.small.b5cf6542-b4cd-4cc4-803a-f33cb944ec54.png&amp;key=d04bfffea46d4aeda930ec88cc64b87c&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/81b0674a481b7e5f41e1c76743ae5330/href">https://medium.com/media/81b0674a481b7e5f41e1c76743ae5330/href</a></iframe><p>This visualization gives us a high-level perspective on how hard it is to cleanly separate genres. Some genre groupings do seem to naturally emerge (like how Swing and Bebop appear to be off-shoots of Jazz Blues), but more often genres overlap and blur strongly together. Pop and Hip Hop are highly connected to many nodes — so no natural distinctions emerge between them.</p><p>Even though it is tricky to draw boundaries around genre, it can be extremely helpful because it contextualizes an artist’s performance in relationship to their competitive peers. In response to this challenge, we decided to build a tool that helps people make meaningful comparisons between artists. To build the best tools though, we have to know why people want to compare artists at all.</p><h3>Why Compare Artists?</h3><p>Comparing artists is both one of the most common and the most useful techniques to get perspective on artist performance. Usually when people make comparisons between artists, it is because they want to know more about a specific artist (we’ll call that the “target artist”). At a glance, we can tell how the target artist is performing by comparing them:</p><ol><li>across the music industry in general.</li><li>against their direct competitors.</li></ol><p>In both these scenarios, we need to see artists that are “similar” to our target artist. But if you really dig into this, you’ll realize that we often need to have different reference groups for each of those insights.</p><blockquote>In both analyses, we may say that we want to compare an artist to similar artists… but each relies on a different definition of “similar.”</blockquote><h4>Two ways to identify “similar” artists:</h4><p>We built the “neighboring artist” feature on Chartmetric to help you quickly assess artist performance. <a href="https://blog.chartmetric.io/new-features-cities-page-neighbor-artists-instagram-follower-demographics-amazon-music-ba5eb669daf8">You might be familiar with it already</a>. Going forward, this feature will have two options for grouping artists, and it is worth looking into how the two methods offer different insights.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*RA3fys1edJcREgjL" /><figcaption>Two is often better than one. That’s why “Neighboring Artist” allows you to define neighbors in two ways. (Photo by <a href="https://unsplash.com/@hannynaibaho?utm_source=medium&amp;utm_medium=referral">Hanny Naibaho</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><h4>Method 1: Universal Sorting on Similar Performance</h4><p>The original option with our “Neighboring Artist” feature allows users to specify a single metric, and we return a list of artists that perform similarly. Often this is incredibly helpful for determining how an artist is doing across the music industry in general.</p><p>Let’s look at a quick example. Let’s say that I want to understand the Spotify performance of <strong>Kendrick Lamar</strong>.</p><blockquote>I can see that Kendrick Lamar has 23.76 Million monthly listeners on Spotify… but I may wonder how successful that actually is.</blockquote><p>By comparing my target artist to artists with a <strong>similar metric</strong> (i.e., monthly listeners on Spotify), I can get an idea:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GzBHVeUk9kopt8GQ_H3Bgg.png" /><figcaption>We can understand Kendrick Lamar’s Spotify performance more by looking at other artists with similar monthly listeners. (Screenshot from Chartmetric.com)</figcaption></figure><p>Kendrick Lamar is right between P!nk, ZAYN, James Arthur, and Demi Lovato. Each of these artists is a power performer, so now I feel like I’ve got a better idea of what “23 million” monthly listeners means. It seems to be a pretty elite tier of performance.</p><p>What’s more, I can compare these artists across other metrics, too. Kendrick Lamar has a similar number of monthly listeners as James Arthur and ZAYN, but he is doing much better on our measure for <a href="https://blog.chartmetric.io/chartmetric-originals-cross-platform-performance-cpp-scores-and-ranks-38df7d3a8566">Cross-Platform Performance (CPP)</a>. He’s trailing P!nk and Demi Lovato on CPP, but he’s leading this pack on Spotify Popularity. I can easily see the nuance of how these artists are situated.</p><blockquote>Comparing artists by the similarity of their performances works especially well for popular artists because each neighboring artist is often well known.</blockquote><p>Pretty cool.</p><h4>Method 2: Cluster Artists By Genre of a Similar Type</h4><p>However, that first method is not always the right comparison to make because it compares performance across genres without accounting for contextual clues. Let’s take a look at another example, where comparing by performance is less helpful. Let’s say I want to look at how <strong>John Prine </strong>is performing on YouTube channel views:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zru-tZSNvnWZJ0woyO47ZQ.png" /><figcaption>Artists around John Prine, sorted by performance only, are not that instructive…</figcaption></figure><p>We have a wide range of artists across different genres and regions, and it is not immediately clear how to make sense of this information. Personally, I’m not familiar with any of the artists around John Prine. We have an Australian j-rock/j-dance band, a South African band doing African rock and Afrikaans, and an American artist doing “strut.”</p><blockquote>I thought “strut” was what I did through my grocery store vegetable section, but apparently it’s also a <a href="http://everynoise.com/new_releases_by_genre.cgi?genre=strut">music genre</a>.</blockquote><p>Unlike the Kendrick Lamar example above, I don’t gain much insight from comparing John Prine to artists with similar performance. Because I am not familiar with the other artists, they are not a great reference point. I suppose this could be due to my own music ignorance (I’m a data-nerd after all — a certain degree of social ignorance isn’t too surprising). But no one can be expected to know all artists. Even the most savvy music aficionado will eventually run into comparison artists with which they are unfamiliar.</p><p>Without knowing the comparison artists, I cannot get much perspective on my target artist. Look at the other performance metrics, like the Cross-Platform Performance rank. Is John Prine ranked higher because he is a great artist? Or is his higher CPP rank due to his staus as a more mainstream artist, while the other artists are in more obscure genres? The answer to these questions is not immediately clear.</p><p>How could we improve this comparison?</p><p>Even if I don’t know every artists’ name and performance, it can be helpful to see comparison-artists that are in the same genre as my target-artist. If I could see how other ‘folk, rock, singer-songwriter’ artists are performing on YouTube, I could more accurately make sense of how John Prine is doing. What I want is data on <strong>artists that are of a similar type to the target artist</strong>.</p><p>If we were to compare artists within a single type, then we would want see results like these:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zjbvtFBAZ-1qcl1Ko8rVWQ.png" /><figcaption>Now we see artists that are similar to John Prine by genre and performance.</figcaption></figure><p>Now we can see that John Prine is one of the top in his genre by Spotify popularity and monthly listeners but he’s trailing behind on CPP, Spotify followers, and Deezer. Overall, John Prine is doing alright among his genre-peers, but he’s far from leading the pack of singer-songwriters.</p><p>Intra-genre comparison gives me a completely different perspective compared to that of Method One described above. This comparison can help me evaluate how competitive my target artist is within a particular listener base, or it may help me find similar artists to try to work with on tours or collaborations. Emmylou Harris would be a good artist for me to keep an eye on, as both Harris and Prine seem to have similar performance and audiences.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Hk_ho1_r5V_CjoB1" /><figcaption>Doggo has a question. (Photo by <a href="https://unsplash.com/@camylla93?utm_source=medium&amp;utm_medium=referral">Camylla Battani</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><h3>What are the Benefits of Machine Learning?</h3><p>Couldn’t we just filter artists based on genre? Doesn’t that basically do the same thing? Fair question.</p><p>Cluster analysis does something slightly different than filtering and has two major benefits. First, it finds natural associations instead of exact matches. Second, it minimizes personal bias.</p><h4>AI Benefit #1: Natural Associations Emerge from Multi-Dimensional Space</h4><p>If you filter, you have to create categories manually and filter based on <strong>an exact match</strong>. So suppose we create some categories like:</p><p>Pop, Rock, Hip-Hop, Indie, Country, Metal, EDM, and Classical</p><p>Using that list we could filter, say, Pop artists to be grouped together. Seems like it makes sense. If you give me Katy Perry, I can give you back Taylor Swift and Ed Sheeran. Cool.</p><p>But you don’t have to go far to run into problems.</p><p>What do we do with genres like K-Pop? Or Heavy Metal? You could lump these examples with a “parent” genre (e.g., Pop and Metal respectively).</p><p>But going deeper, you run into trickier problems. Where do we put Indie Electro-Pop? Is it Indie or is it Pop? The answer is not immediately clear.</p><blockquote>While identifying a single parent genre for Indie Electro-Pop isn’t immediately obvious, it is fairly obvious that Indie Electro-Pop is “closer” to Indie, Electronica, and Pop music than it is to Heavy Metal or Grunge.</blockquote><p>The promise of machine learning is that we can run an algorithm that “learns” how similar or dissimilar genres are based on how frequently they co-occur in artists profiles. Artists that are identified as playing Pop, Indie, and Indie Electro-Pop will increase the association between those genres. If you’re not up on machine learning concepts and terminology, then all you need to know is that we’re just observing how similar genres are to each other out there “in the wild.” If you’re up-to-date on ML, you’ll realize that we’re talking about plotting genres in a multi-dimensional space and measuring how far apart they are.</p><p>Because “plotting genres in multi-dimensional space” can be hard to conceptualize, here is a simple example.</p><p><strong>An Example of AI Clustering: Kane Brown</strong></p><p>Let me show you an example of how this looks in action, inspired by the genre tags associated with artist Kane Brown. I see that Kane Brown is identified as playing “Contemporary Country”<strong> </strong>and<strong> </strong>“Country Road”.</p><p>If we wanted to compare Kane Brown to “similar” artists, who would we pick? Kane Brown has blurred genres beyond just “country” music, so filtering on country isn’t the best bet. We’ve talked <a href="https://podcast.chartmetric.com/episodes/2019-04-05-your-daily-data-dump-from-chartmetric">elsewhere</a> about how Becky G introduced him to the Latin music world through a collaboration–so how would you categorize this track?</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FVAuz6mlqxNY%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DVAuz6mlqxNY&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FVAuz6mlqxNY%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/ce1ebb71e1c2605d9f8b8250f3794309/href">https://medium.com/media/ce1ebb71e1c2605d9f8b8250f3794309/href</a></iframe><p>Categorizing that one song is difficult, let alone Kane Brown’s entire repertoire. Does this mean Country Road is Latin infused? Or is his collaboration with Marshmello a better example of Country Road? Instead of relying on my own knowledge (which is limited) to try to collapse Country Road into a parent genre, we can use machine learning to track how often that genre tag co-occurs with other genre tags.</p><p>For simplicity, I’ll only take a very small sample of artists (12) and put their genre tags together into a list. Each row represents one of my 12 artists’ genre tags:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*evFFdvfKpSq9YKBBDT4daA.png" /><figcaption>A sample of artist genre tags.</figcaption></figure><p>Just with this list, we can start to get an idea of how Country Road is situated around other genres. We can already infer that:</p><ul><li>Country Road and Country Dawn almost <strong>always</strong> show up together</li><li>Artists that play Country Road also <strong>often</strong> play Contemporary Country</li><li>Contemporary Country overlaps with Modern Country and Pop <strong>just as much</strong> as it overlaps with Country Road</li><li>Pop and Christmas Pop <strong>never overlap</strong> with Country Dawn</li></ul><p>We can see those patterns with our own eyes, since this list is only 12 entries long. But it would be better to get a computer to do it instead, because then we could scale the process up to millions of entries. If only some very savvy people had spent a lot of time pioneering a range of techniques to turn observations about similarity into quantifiable results… Oh wait! They have! Hey thanks, Data Scientists! (Although, its probably more accurate to say, “Thanks Statisticians.”)</p><p>We can use cosine similarity with this list of artist genres to calculate a single numerical value that captures the degree to which each term is associated with each other. You can <a href="https://en.wikipedia.org/wiki/Cosine_similarity">read a lot about cosine similarity elsewhere</a>. It’s a fascinating and handy tool, but I won’t clutter this article by going into it here. Let’s look at the top results, and I’ll help explain how to interpret them:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/494/1*O9YLGpR9yXDfp-kzj8W5oQ.png" /><figcaption>The top similar-genres to “Country Road” calculated using cosine similarity and the above 12-item list.</figcaption></figure><p>You can connect these numbers directly to our insights above:</p><ul><li>Because County Road and County Dawn almost always show up together, Country Dawn gets the highest similarity value (0.95)</li><li>Country, Contemporary Country, and Modern Country Rock all commonly overlap with Country Road, so they have relatively high values (0.82–0.83). But because those tags <strong>also</strong> show up <strong>without</strong> Country Road sometimes, their association values are lowered.</li><li>Pop never shows up with Country Road so its score for “similarity” is very small (0.22).</li></ul><p>Well look at that, these numbers match up with the exact same insights that we created by manually looking at the list. That’s handy! If we were to expand our scope to include even more genre tags, we would really start to get a profile of Country Road and understand where it fits with the overall music.</p><blockquote>This is just a small sample to illustrate how measuring similarity works. We can easily apply the same process to all 1.5 million artists in our database while looking at 1657 genres.</blockquote><h4>AI Benefit #2: Removal of Personal Bias</h4><p>One of the biggest perks about machine learning is that personal bias is greatly minimized. In the above example, I didn’t decide whether Country Road is legitimate Country. I don’t get to act as a gatekeeper about what “is” and what “is not” inside a parent genre.</p><p>Instead, the landscape of musicians decides organically. Proximity is measured as a result of artists’ identification with similar genre-tags. If a lot of people who play “Pop” also play “Indie Electro-Pop,” then we draw a stronger association between Indie Electro-Pop and Pop.</p><p>Now you’ve seen some benefits to unsupervised machine learning for clustering, but how it is actually done? Read on, dear reader. Read on.</p><h3>Nuts and Bolts to Genre Clustering</h3><p>So when we began our mission was: “Group genres by co-occurrences. The more genres show up together, the more coherent a group they make.”</p><p>Sounds easy!</p><p>And… it is easy… if you know some Natural Language Processing (NLP). Be aware though, that the NLP decisions you make will impact how the genres are grouped. How? I’m glad you asked.</p><p>Of course, you know that computers don’t actually read text, they just work with numbers. In order to compare two instances of text (for us that is two artists’ lists of genre tags), we have to translate the text into a format that a computer can understand. Computers have no concept of what “pop,” “jazz,” or “strut” are. That also means:</p><blockquote>A computer does not inherently know that there is some association between “pop” and “indie pop.” We have to tell it how to parse these phrases.</blockquote><p>There are a lot of machine learning buzzwords related to the process of parsing text into computer readable formats, and maybe you’ve heard of some of them: bag of words, tokenization, one-hot encoding, count vectorizing, dictionary vectorizing, and TF-IDF, to name a few.</p><p>Each is important (and in fact we used each at different points), but I’m going to give you a high-level overview of how they affect our process differently.</p><p>Take two hypothetical lists of artist tags. We’ll reuse some of the example tags we saw above.</p><pre>Artist 1 - {&#39;contemporary country&#39;, &#39;country road&#39;}<br>Artist 2 - {&#39;country&#39;, &#39;pop&#39;}</pre><p>The simplest way to parse these genres is to break out every single word by itself. Then, we keep a master list of all the words and count up the number of times each word from the master list shows up in each entry.</p><p>The consequence of this method is that it obliterates word-order and context. Here, we treat all words exactly the same, essentially just throwing them into a single bucket or… a bag. (Hint: this is the “bag of words” method!)</p><p>Our lists have a combined total of 4 unique words ( contemporary, country, road, and pop). So applying the bag of words method, we get this result:</p><pre>Artist 1 - {&#39;contemporary&#39;: 1, &#39;country&#39;: 2, &#39;road&#39;: 1, &#39;pop&#39;: 0}<br>Artist 2 - {&#39;contemporary&#39;: 0, &#39;country&#39;: 1, &#39;road&#39;: 0, &#39;pop&#39;: 1}</pre><p>From this, you can easily see how the artists diverge in almost every case, but they do overlap on the word country. Using each artists’ degree of overlap, we can start to measure how similar or dissimilar each artist is from one another. This method is a staple of NLP because it is remarkably effective and easy to implement.</p><p>Still, there are situations where this method would fall short. What about the following list?</p><pre>Artist 1 - [&#39;pop&#39;, &#39;christmas pop&#39;]<br>Artist 2 - [&#39;comedy&#39;, &#39;humor&#39;, &#39;pop parody&#39;, &#39;christmas parody&#39;]</pre><p>These artists would show up as having some overlap because they have counts for both pop and christmas even though the contexts are very different. The second artist here is doing a completely different style of performance. In fact, they might be a stand-up comic! You can see that the additional context we get from identifying that parody follows pop is really important. Now you see why word-order, the very thing we threw out in bag-of-words, is important.</p><p>There are two methods available to help us capture word-order information:</p><ol><li><strong>N-grams</strong> : We take all the combinations of words in our bag of words and count the combinations. We would start with pop, christmas, comedy, humor, parody but also add pop christmas, christmas pop, comedy humor, humor pop, pop parody, parody christmas, christmas parody. This will capture word order without a problem. Artist 1 will have no instance of &#39;pop parody&#39; or &#39;christmas parody&#39;, which will help us draw dissimilarity between these artists. We’ll still see them overlap on pop and christmas, so this solution still isn’t perfect. This also has the potential to spiral out on memory usage because the number of combinations increases exponentially as your number of words increases.</li><li><strong>Dictionary Vectorizing</strong> : Alternatively, we can require that every genre tag be an unique entity and only count perfect matches. By doing this, we turn the list above into Artist 1 — {‘pop’:1, ‘christmas pop’:1} and Artist 2 — {‘comedy’: 1, ‘humor’: 1, ‘pop parody’:1, ‘christmas parody’:1}. This solves the problem that we ran into with n-grams because these artists now do not overlap at all! That seems great right? We have all of our genres distinct and we can tell when someone is a comedian versus a serious musician! However, this is a double-edged sword. Now our algorithm can’t tell that an artist who does folk rock is similar to someone who does rock. Because every genre is unique and has to match perfectly, our computer sees folk rock as no more similar to rock than it is to classical performance or screamo metal. Well, that’s not ideal. Similarly, hip hop and hip-hop count as distinct genres, too. That’s bad — alternative spellings lead to all sorts of problems.</li></ol><p>Even with all of these methods available to us, there is clearly no silver bullet to all of our problems. Each decision clarifies in some instances and hurts in others. Still, there are some techniques that are generally a good idea. For example, replacing all instances of hip hop and hiphop with a uniform spelling of hip-hop will help us get more consistent results.</p><p>NLP would be a lot easier if there were a consistent playbook that guaranteed good results, but alas you have to experiment with what works best for your use case. I’ll cut to the chase and show you what seemed to work best for us.</p><h4>Steps to Create Chartmetric Genre Clusters for Artists</h4><p><strong>Step 1 -</strong> <strong>Genre Cluster Analysis Using Dictionary Vectorizing:</strong> This catches the distinctions between categories like pop , pop parody, and K-pop. It helps us make fine-grained distinctions between groups. For the artists that this stage worked for, the clusters work pretty well. (TF-IDF helped here too, but let’s not open that can of worms in this article.) Using this method, we get finalized clusters for about 80% of artists.</p><p><strong>Step 2 -</strong> <strong>Identify Ineffective Clusters:</strong> The first step did not work for about 20% of artists, but why? Looking at the data, we had two problematic groups. One cluster was radically larger than every other cluster. We looked into it, and it turned out this was a “catch all” cluster. Every artist in that group was similar insofar as they were uncommonly unique. The second cluster comprised every artist that was missing genre tags–the “null” cluster.</p><p><strong>Step 3 - Address the “Catch-All” Cluster with a New Bag of Words Analysis:</strong> Artists in this cluster had super specific genre tags. For example, it is hard to group artists with tags like Finnish emo electro-indie death metal when you are matching genres exactly (which we were doing in the dictionary vectorizer in Step 1). In the first step, an artist in this category would be tagged like this: {Finnish emo electro-indie death metal: 1} Instead, we run a new cluster analysis using the bag of words method. The above artist now has {Finnish: 1, emo: 1, electro: 1, indie: 1, death: 1, metal: 1} . With that kind of profile, we can start to tell that this artist overlaps with other emo , metal, and indie artists. For the artists in the “catch-all” cluster only, we update their initial cluster id with their new cluster id.</p><p><strong>Step 4 - Address “Null” Cluster with New Data Sources:</strong> A lot of artists did not have genre tags directly associated with their artist profile. That makes it pretty impossible to match them! Fortunately, we also have genre tags associated with specific songs/tracks. We went to each artist’s catalogue, picked their top songs , gathered the genre tags associated with those tracks, and lifted those genre tags to be associated with the artist. This method is less effective than the initial cluster, so again we only update the cluster identifications for artists that were in the original “null” group with the results from this analysis. (We still have a number of artists that are missing genre data and are actively working to incorporate different data sources to fill in our blind spots.)</p><p>That’s it! All the hard work is done.</p><p>We just have to bring all the correct cluster identifications together into one spot so we can quickly refer to and assess each cluster.</p><p>Our machine learning model determined the groups, so we have to do a quick assessment of each cluster to get an idea of what brought them together. We went cluster-by-cluster and counted up all the genre tags within each cluster. Now whenever we select a cluster, we also return the top 5 most common genre tags in that cluster. That gives us an idea of the types of artists that were brought together in that cluster.</p><p>Here’s an example of an artist, their cluster, and some of their new neighbors:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*piIX9_fxRLYZCvL8ZOmcOA.png" /><figcaption>An example Chartmetric search for Paramore, showing a sample of genre-neighbors.</figcaption></figure><p>You can check this new feature out on the Chartmetric website right now. Simply click on any metric for any artist to bring up a universal comparison on that stat. Then click the green-toggle labeled “Within Genre Cluster” in the upper right corner to compare within your artist’s genre cluster.</p><p><em>Shout out to </em><a href="https://www.linkedin.com/in/komala-prabhu/"><em>Komala Prabhu Tirupachur</em></a><em> and </em><a href="https://www.linkedin.com/in/gregory-yannett-85b55a182/"><em>Greg Yannett</em></a><em> for their critical roles in helping this feature go live.</em></p><p><em>Do you dig data as much as we do? Let </em><a href="https://blog.chartmetric.io/@jasonjoven"><em>Jason Joven</em></a><em> keep you in the loop on the music industry’s latest data gems by tuning into our podcast: </em><a href="https://chartmetric.transistor.fm/episodes"><em>Your Daily Data Dump from Chartmetric</em></a><em>.</em></p><p><em>Or if you can’t wait to play with genre clusters, then head over to the Chartmetric website and grab your </em><strong><em>free account</em></strong><em> </em><a href="https://app.chartmetric.com/signup"><em>here</em></a><em>. Also, we’d love to hear from you at hi@chartmetric.io.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4bd5a0e536cc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/chartmetric-blog/using-ai-to-make-meaningful-comparisons-between-musicians-4bd5a0e536cc">Using AI to Make Meaningful Comparisons Between Musicians</a> was originally published in <a href="https://medium.com/chartmetric-blog">Chartmetric</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Chartmetric Originals: Cross-Platform Performance (CPP) Scores and Ranks]]></title>
            <link>https://medium.com/chartmetric-blog/chartmetric-originals-cross-platform-performance-cpp-scores-and-ranks-38df7d3a8566?source=rss-7a06e0067f24------2</link>
            <guid isPermaLink="false">https://medium.com/p/38df7d3a8566</guid>
            <category><![CDATA[insights]]></category>
            <category><![CDATA[music]]></category>
            <category><![CDATA[artist]]></category>
            <category><![CDATA[rankings]]></category>
            <category><![CDATA[product-updates]]></category>
            <dc:creator><![CDATA[Josh Hayes]]></dc:creator>
            <pubDate>Wed, 05 Jun 2019 00:05:07 GMT</pubDate>
            <atom:updated>2019-06-27T17:01:38.131Z</atom:updated>
            <content:encoded><![CDATA[<h3>Chartmetric Originals: Cross-Platform Performance (CPP) Ranks and Scores</h3><h3>A detailed breakdown of how to instantly compare artists’ general performance by utilizing Chartmetric’s data warehouse</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*uhLCRjsz77CwfFRI" /><figcaption>Photo by <a href="https://unsplash.com/@lindsayhenwood?utm_source=medium&amp;utm_medium=referral">Lindsay Henwood</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>Update as of 2019/06/22: Cross-Platform Performance now also includes Apple Music as a data source. As a result, all weights and calculations have been tweaked for better balance.</strong></p><p>The biggest challenge to understanding artist performance is contextualizing different metrics from an incredible number of potential data sources. Many people start their quest to analyze artist performance by gathering as much data as they can. However, they soon find themselves surrounded and overwhelmed by all the data available to them.</p><p>Our goal at Chartmetric is to push boundaries by strategically utilizing data to empower the music industry, so we set out to find a solution to this problem.</p><blockquote>The biggest challenge to understanding artist performance is contextualizing different metrics from an incredible number of potential data sources.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*lIjUt7-8tRDjVD60" /><figcaption>“Of course I know the way. It’s clear as day… right here… in all these maps…” (Photo by <a href="https://unsplash.com/@andrewtneel?utm_source=medium&amp;utm_medium=referral">Andrew Neel</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><p>Let’s say we want to assess just one artist’s performance. Consider the dizzying number of data sources:</p><ul><li><strong>Music Platforms</strong>: <em>Spotify, Apple Music, Youtube, Deezer, iTunes, Soundcloud, LastFM, Songkick, etc.</em></li><li><strong>Social Media Platforms</strong>:<em> Facebook, Instagram, Twitter, Wikipedia, and more.</em></li><li><strong>Success Measures</strong>: <em>Streaming activity, Shazam searches, chart position, active monthly listeners, followers, subscribers, likes, retweets, Wikipedia searches, and fans to name a few.</em></li><li><strong>Secondary Measures</strong>: G<em>rowth rates, averages, aggregate statistics, </em><a href="https://blog.chartmetric.io/measuring-attention-on-spotify-followers-vs-listeners-da0383d90796"><em>Listener-to-Follower Ratios</em></a><em>, </em><a href="https://blog.chartmetric.io/weighting-music-charts-grammy-2019-effects-or-not-8bb791dfba7f"><em>weighted Chart summaries</em></a><em> — the sky is the limit.</em></li></ul><p>All of these metrics are important, and savvy individuals will look at each measure directly when the situation calls for it. For example, did that banner ad for your YouTube channel actually increase traffic? But very often, people just want a general assessment of how artists are doing–and they need it quickly. Unfortunately, quick assessments are often vulnerable to relying on, or being easily swayed by, personal biases.</p><blockquote>Contextualizing all the potential data-driven insights consistently, systematically, quickly, and without personal bias is very difficult.</blockquote><p>To meet this challenge, our team has been consulting with users from all over the music industry and working diligently to create a quantifiable, logic-driven, and objective measure for cross-platform performance.</p><h3>Introducing: CPP Ranks and Scores</h3><h4>The Goal</h4><p>Chartmetric users have a lot of different analytical needs, and our goal is always to <strong>create tools that empower you to do work in your own way rather than dictate or limit your perspective</strong>.</p><blockquote>You know your job better than we do. We know data. Let us empower you to do the amazing work you do.</blockquote><p>With that in mind, we had the following goals while creating a cross-platform performance measure:</p><ul><li><strong>Universality</strong>: The measure needs to apply just as well to an artist that is just starting out in the first year of their career as well as for top-artists aiming to grow or extend their established success.</li><li><strong>Flexibility</strong>: The measure needs to offer potential insights to analytical projects based in publishing, marketing, sales, A&amp;R, and to creative artists themselves.</li><li><strong>Market Responsiveness</strong>: The music world is changing radically. The significance of absolute measures change <em>daily</em>. That 10K increase in your monthly streams could still mean you’re losing market share if the digital market grew radically (which by the way, it is). Scores need to reflect the landscape of the music market today, not yesterday.</li><li><strong>Transparency</strong>: If users don’t know how a tool works, then even in the best case they can’t see how useful it is. In the worst case, they won’t trust it. Users need to understand how scores and ranks are calculated. They need to understand why an artist’s values move, whenever they do.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NUi3wPBlZimzeT7r" /><figcaption>This picture is just a metaphor: “We want to be the foundation for you to do beautiful things.” But, if I ended up with abs like that, well then that’s just a win-win. (Photo by <a href="https://unsplash.com/@theformfitness?utm_source=medium&amp;utm_medium=referral">Form</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><h4>Inspiring Examples:</h4><p>To understand CPP ranks and scores, let’s look at some examples that illustrate what our users told us they care about. Let’s say we want to compare the career-trajectory of three hypothetical artists that are trying to grow their audiences and revenue. Each artist just experienced a major event in their career:</p><ol><li>One artist has had mainstream popularity for over a decade, but has been relatively inactive for a few years. They released a new single and have promo-ed it by performing on a mainstream TV-show.</li><li>The second artist already has two fairly successful albums and now their team is trying a new staggered-release strategy for their next album. They’ve teased a few tracks and announcements on Instagram and Facebook that just culminated in the full album release on Spotify.</li><li>The last artist has a small but dedicated following on Youtube. They just did a Soundcloud collaboration with a Twitch streamer to try to break into the gaming-music scene in order to grow their audience.</li></ol><p>Regardless of their job title, many people in the music industry have a similar question about the above scenarios:</p><p><strong>“How successful were these artists’ in gaining overall popularity, listener engagement, and growing their audience?”</strong></p><p>So how do we compare?</p><p>Of course there are many ways to slice and dice artist performance, but there is near universal consensus on some general, undeniably good indicators for artists’ success. Using those indicators to identify a “best case” scenario, we can then compare all artists on a scale of how close they got to the “best” artist according to today’s stats. We concluded that a hypothetical best case scenario for performer success is exemplified by:</p><ol><li><strong>Broadest exposure to audiences</strong>: More streams! More views! More retweets! The more audience contact points the better. Each day there is a finite number of eyes and ears out there, and the more of them trained on an artist the better.</li><li><strong>Healthy fan engagement</strong>: Just getting streams is not all that matters. You want to absorb some of those new listeners into a reliable fanbase. Loyal followers can be vital to longterm success.</li><li><strong>Success across multiple platforms</strong>: Different platforms cater to different audiences, but viral sensations still commonly jump from one platform to another. Having a strong presence on each platform ensures you reach as many people as possible and can also capture viral momentum rather than inhibit it.</li><li><strong>Having longevity with audiences, rather than just being a flash in the pan: </strong>Longevity with audiences can have major influences on revenue and/or return on investment. Sadly, some people’s 15-minutes of fame seem to only last 1-minute. While we want to catch viral sensations as they’re happening, the best artists don’t just have a lucky hit and then fade into obscurity. They leverage that hit’s momentum to keep producing, keep engaging, and keep growing.</li></ol><p>With those principles in mind, we turned to the data to match those concepts up with concrete numbers.</p><h3><strong>Calculating Cross-Platform Performance (CPP) Scores</strong></h3><p>Looking at <a href="#e718">all possible measures for success</a>, we realized that the most common measures used to assess artists fall into two separate categories:</p><ol><li><strong>Stage:</strong> This is the amount of “reach” or “exposure” that an artist has over audiences. The bigger the stage, the more people <strong>actively</strong> listening, watching, or consuming what the artist is creating.</li><li><strong>Followers:</strong> This is the size of an artist’s “fanbase” or an artist’s “stickiness” with audiences. Followers have opted into tracking an artist and therefore are more likely to re-engage with the artist’s products in the future. Followers are <strong>not actively</strong> engaging with an artist all the time, but artists have an easier job of connecting with followers than non-followers.</li></ol><p>Our users told us that both Stage and Followers are important, but not equally. A vibrant fanbase may be more likely to continually consume an artist’s products, but frankly, active audience size is most directly tied to revenue. The key to assessing an artists’ potential is to balance the importance of each.</p><h3>TL;DR here’s how we calculate the Cross-Platform Performance (CPP) Score:</h3><ol><li>Every day, we calculate a single <strong>Stage</strong> <strong>score</strong> to encapsulate artist performance across numerous platforms for all of our 1.6 million artists, on a scale from the individual with the biggest stage down (1.0) to the smallest stage (0.0). Then we repeat the process to get each artist’s <strong>Followers score</strong> on a similar scale (0.0–1.0).</li><li>Now that we know artists’ relative performance in both categories, we take a weighted average of the two that heavily prioritizes <strong>Stage </strong>over<strong> Followers</strong> to find their <strong>initial Cross-Platform Performance (CPP) score</strong>.</li><li>Next, we apply a number of small <strong>bonuses</strong> and <strong>penalties</strong> based on simple, but important, characteristics. For example, it is undeniably better to have a presence on two music platforms rather than just one (regardless of your levels of success on each platform).</li><li>The result? The <strong>final Cross-Platform Performance (CPP) Score </strong>between 0 and 1. We order each artist by their CPP score and then we order them to get the top 100,000 artists by <strong>Rank</strong> for each day.</li></ol><p>Now we’ll deep dive into each stage of the process. <a href="#c907">Or you can jump down to the bottom of the article if you’re ready to see real examples of what you can do with CPP scores and ranks.</a></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F340303978%3Fapp_id%3D122963&amp;dntp=1&amp;url=https%3A%2F%2Fvimeo.com%2F340303978&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F788597071_1280.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1920" height="1080" frameborder="0" scrolling="no"><a href="https://medium.com/media/2a808cad894ca02f9e0537bbbbcb134d/href">https://medium.com/media/2a808cad894ca02f9e0537bbbbcb134d/href</a></iframe><h4>CPP Primary-Component 1: Stage</h4><p>Stage is the bedrock of assessing artist cross-platform performance in the digital age. It measures how many eyes and ears an artist is reaching at this moment and most accurately responds to the dynamic tastes of the general public.</p><blockquote>“Stage” captures the extent to which audiences are actively consuming an artist’s work, which connects directly to revenue generation.</blockquote><p>Most music industry analysts use consumption patterns to inform their business decisions with the most recent, relevant, and insightful information about the market right now. Stage lets you know who is hot right now.</p><p><strong>Stage is a weighted average of an artists’ reach measured across the following platforms</strong>:</p><ul><li><strong>Spotify</strong> (total monthly listeners over the past 28 days)</li><li><strong>YouTube</strong> (total views over the past 7 days)</li><li><strong>Apple Music </strong>(number of playlist inclusions over the past 7 days)</li><li><strong>Wikipedia</strong> (average lookups over the past 3 days)</li><li><strong>Twitter</strong> (average retweets over the past 3 days)</li></ul><p>There are a few important things to note here:</p><ul><li><strong>Different measures for different platforms.</strong> Whether we use total monthly, total weekly, or multiple-day average values is based upon critical assessment of our data sources, gathering schedule limitations, as well as each feature’s behavior, characteristics, and distributions.</li></ul><blockquote>At every decision point, we opted towards making a measure that was more robust and defensible, while still sensitive to major shifts in the music market.</blockquote><ul><li><strong>Normalized values within each platform.</strong> Today, Spotify monthly listeners range from 0 to 149.6 million while average Wikipedia lookups range from 0 to 292,000. However, being the #1 most popular artist on any platform is going to garner you a lot of industry attention, regardless of the size of the platform. To balance these tensions, we look at the normalized distribution of values for each platform. Essentially we ask “Where do you fall in the pack of Spotify monthly listeners?” and “Where do you fall in the pack of Youtube views?” The answer to both those questions can be easily compared using percentiles.</li><li><strong>Rolling average defends against “flash in the pan” situations while still capturing major shifts of the public eye. </strong>By using at least a 3-day average, we still capture the virality of one-off events, but artists that spike one day and drop the next day will not rise as much as artists that continue to carry public interest. For example, a legacy artist recently passed away and consequently saw a massive (multiple thousand times) increase in Wikipedia lookups–for one day only. We found that the artist’s streams and followers did increase, but not as strongly as one might expect in the middle of their Wikipedia spike. The rolling average more accurately identifies artists’ shifts without over-estimating changes to their Stage.</li><li><strong>We weigh the importance of music platforms multiple times more than non-music platforms.</strong> Spotify, Youtube, and Apple Music account for the majority of the stage score, while Wikipedia and Twitter account for just a small portion of the score.</li></ul><blockquote>We look at how well an artist is performing on each platform (relative to all other artists on that platform). Then we average each of their platform-performances together, while prioritizing music platforms over non-music platforms. That is the Stage score.</blockquote><h4>CPP Primary-Component 2: Followers</h4><p>There is some general grumbling that goes on about how surprisingly unsupportive “followers” can be. Not every “follower” is a die-hard fan. But some are, and having a vibrant fanbase has been a vital element for growing many artists’ careers. If an artist has a viral hit, but can’t convert any listeners to their fanbase, then that artist will have a very hard time reaching the same market-share of listeners for their next track. Every add to a follower-base is another listener that is easier to reach the next time around. You’d never turn down more followers, right?</p><p>Correspondingly, the secondary consideration for the CPP Score is followers. <strong>Followers</strong> <strong>is a weighted average of an artists’ fanbase measured across the following platforms</strong>:</p><ul><li><strong>Spotify</strong> (followers)</li><li><strong>YouTube</strong> (subscribers)</li><li><strong>Instagram</strong> (followers)</li><li><strong>Twitter</strong> (followers)</li><li><strong>Deezer</strong> (fans)</li><li><strong>Soundcloud</strong> (followers)</li></ul><blockquote>The Follower score reflects an artists’ total fanbase across multiple, both music and non-music, platforms.</blockquote><h3>Turning Stage/Followers into CPP Scores and Ranks</h3><p>At this point we take a simple weighted average of the Stage and Follower scores to get an initial assessment of the relative performance of every artist across 8platforms. This value is each artist’s “Initial Score.”</p><p>Next, we combined industry insights from our users all over the music world to identify a number of simple features that encourage or inhibit artist career success. These alterations aren’t meant to radically shift artists’ scores, but to help draw distinctions between artists that are similar to each other.</p><p>To match the exponential-logarithmic distribution of most metrics in the music industry (including our initial CPP score), the magnitude of each bonus and penalty is proportional to an artists initial score. Functionally speaking, if one artist has 200K followers and a second artist has 20M followers, and neither artist has much of an Instagram presence, the second artist is failing to utilize more potential revenue generation than the first artist so we penalize the second artist’s CPP score more.</p><h4>Bonuses and Penalties</h4><ul><li><strong>Example Bonus — Cross-Platform Presence</strong>: Cross-platform presence is increasingly important for top-tier artists. If an artist has a viral break-through on Youtube, inevitably interest spreads and users will attempt to look up the artist on other platforms like Apple Music or Spotify. If users cannot find an artist, they quickly lose interest and move on. Therefore, artists get a small boost for having a presence on each additional platform.</li><li><strong>Example Penalty — Non-Musician</strong>: We noticed that some profiles rose to the top that were not who we wanted to showcase in our measure for <strong>musicians</strong>. For example, the initial run of Stage scores saw profiles like Pewdiepie, European Soccer Clubs, and The Bible performing quite well. Each sees huge success on Youtube, Instagram, as well as having tons of Wikipedia lookups; and as long as they have a Spotify profile, then they flow into our database. To address this, we impose a penalty on entries if their Spotify popularity is low.</li><li><strong>The significance of Bonuses and Penalties are relative to our initial assessment of artist performance.</strong> Profiles are given an initial score, then bonuses and penalties are weighted by that initial score, and lastly the initial score, bonuses, and penalties are combined to determine a final score and ranking. What does that mean practically? A musician that is a local act and overall ranked fairly low (37,000th) may still be hit by the low Spotify-popularity penalty but will only lose a small amount of score, because they already started far from the top. If a soccer player released one poorly-received song but also has a massive online presence that got them ranked relatively high (1000th), then the same low Spotify-popularity penalty will be weighted much larger and will shift the profile down much more.</li></ul><p><strong>After the application of bonuses and penalties, we are done. Now, we can compare artists scores directly or order artists into ranks.</strong></p><p>Phew! You made it!</p><p>Take a breath. Enjoy the view. Now that you’ve made it through all that, you can see just how valuable the Cross-Platform Performance score and ranks are.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*XPDAfS10wniK_y9A" /><figcaption>With a macro-view, you can take in a lot of information quickly and easily. (Photo by <a href="https://unsplash.com/@dnevozhai?utm_source=medium&amp;utm_medium=referral">Denys Nevozhai</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a>)</figcaption></figure><h3>Test-Driving the Cross-Platform Performance metrics:</h3><p>Let’s do two quick examples to showcase how powerful CPP scores and ranks are to quickly assess artists.</p><h4><strong>Example 1: Artist Performance Over Time (Ranks)</strong></h4><p>You can use CPP ranks just like you would use any track or artist charts. Ranks give you a simple, direct, way to rank who is performing best. Let’s try an relevant and recent example.</p><p>The past few months, the music world has been watching Billie Eilish with rapt interest. While she’s been making great music for years, Billie Eilish really exploded onto the scene with the release of her March 29th, 2019 album “<em>WHEN WE ALL FALL ASLEEP, WHERE DO WE GO?</em>”</p><p>Shortly thereafter, Billie Eilish usurped Ariana Grande for the coveted top position by “Popularity” on the Spotify platform.</p><p>Can we use the new CPP ranks to visualize how Billie Eilish climbed into the top spots of the music scene, not just on Spotify but across all major platforms? We sure can.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*q0ulO3tbPNU6_cP--lZMrg.png" /><figcaption>Each of these artists is extremely successful, yet faces different career challenges right now.</figcaption></figure><p>Notice how fast Billie Eilish climbed in just a few weeks. Quite an impressive feat. I’ve discussed elsewhere that jumping ten spots from <a href="https://blog.chartmetric.io/the-gender-play-gap-7f5e3a1b0066">#12 to #2 is multiple times harder than jumping from #100 to #90</a>. Yet, while Billie Eilish beats out Ariana Grande on Spotify’s popularity measure, Ariana Grande still is out-performing Billie Eilish across the board, and the gap between the two is now increasing.</p><p>A few other interesting things to note: compare the steady momentum of Marshmello, the relatively steady growth of Halsey, and the peak and descent of Anitta. Each of those artists will likely need a different strategy to leverage their creative talents, fan bases, release schedules, and social media campaigns to help themselves grow and vie for more market share.</p><p>Lastly, I was exploring with different artists on the graph to get context and that is how I easily discovered the extent of BTS’s recent gains. Looks like dropping a new album followed by going on SNL was a powerful combination for the Korean boyband.</p><h4><strong>Example 2: Fastest Growing Artists (Measured by Scores)</strong></h4><p>Let’s track the top 200 artists over the past few days (2019–05–24 to 2019–05–28). Then let’s take the difference between their latest and earliest scores to see who grew the most.</p><p>That will allow us to rank artists by who had the most growth:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/204/1*scTi9j1D_Q1_DGdihr70ew.png" /><figcaption>Top 10 artists that grew the most from 2019-05–24 to 2019–05–28.</figcaption></figure><p>First, it is very helpful to so easily identify which artists are improving their performance across multiple-platforms at once. Now we know a “short-list” of artists that we may want to dig deeper into, or artists to keep an eye on in the coming days.</p><p>All the artists in that list are moving up, but to give us a little more intuition into how their performance compares we can plot our top-two gainers (Nio Garcia and Mc Davi) against some other mainstream artists (Ariana Grande and Bradley Cooper).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7Thjsh8L1CppAS3-i_VcHQ.png" /></figure><p>With CPP scores we can see that Mc Davi has essentially caught up to Bradley Cooper and that Nio Garcia started about the same but has been blasting ahead. Both artists have a long ways to go still to catch up to Ariana Grande’s lead. With this graph, you can see that CPP scores are a very simple way to compare your own performance to artists that you identify as your competitors.</p><h3>Now you’ve seen CPP Ranks and Scores in Action:</h3><p>The more data tools in your toolbox, the more empowered you are to choose the best tool for the job.</p><p>CPP scores and ranks both capture a similar concept, Cross-Platform Performance, with different levels of granularity. If you want a general comparison, ranks are a good way to size up multiple artists. If you want to track your market share over time, scores will be the way to go.</p><p>We look forward to seeing how our clients utilize the new Cross-Platform Performance metrics to empower their work.</p><p>In closing, here are some final notes about special cases and model specifics, for the truly algorithm-savvy among you:</p><ul><li><strong>Missing Links</strong>: If an artist’s external links are inaccurate or we can’t see an artist’s presence on a platform, that tends to hurt them. And… logically speaking, it should right? If an artist has 50k Soundcloud fans (and let’s say the Chartmetric database has no indication they have a Soundcloud presence at all), then the artist should have a higher score than we are giving them. To guarantee that the artists you are assessing have the most accurate values, make sure those artists have up-to-date links for all their off-site connections. We are constantly working on verifying, updating, and cleaning our links but with 1.5 million artists it is a monumental task. We automate it where we can, manually update where we can’t, and encourage crowd-sourcing whenever you stumble upon an error.</li><li><strong>Missing Data:</strong> We wake up everyday working to make sure our data is the best in the music industry. But still, on occasion missing data happens. Because it can be difficult to distinguish a data input error from a valid change in artist performance, we have to count missing entries as marks against artists. If an artist is not on Youtube at all, then their CPP should obviously be lower, right? If we receive a missing value for an artist who is on a platform, we could erroneously assume they are not on that platform. We cross-validate missing entries to past records and then utilize rolling averages in order to give the CPP numerous lines of defense to missing values.</li><li><strong>Future Developments</strong>: We hope to implement further bonuses and penalties going forward. For example, giving boosts to artists that show hockey-stick growth or penalizing artists with stage-sizes that are stagnating. We want to only add more aspects to the CPP Score if they are defensible, widely embraced, and tested. If you have ideas or feedback, we would love to hear from you.</li></ul><p><em>Do you dig data as much as we do? Let </em><a href="https://blog.chartmetric.io/@jasonjoven"><em>Jason Joven</em></a><em> keep you in the loop on the music industry’s latest data gems by tuning into our podcast: </em><a href="https://chartmetric.transistor.fm/episodes"><em>Your Daily Data Dump from Chartmetric</em></a><em>.</em></p><p><em>Or if you can’t wait to work with Cross-Platform Performance measures yourself, then head over to the Chartmetric website and grab your </em><strong><em>free account</em></strong><em> </em><a href="https://app.chartmetric.com/signup"><em>here</em></a><em>.</em></p><p><em>Lastly, we’d love to hear from you at hi@chartmetric.io.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=38df7d3a8566" width="1" height="1" alt=""><hr><p><a href="https://medium.com/chartmetric-blog/chartmetric-originals-cross-platform-performance-cpp-scores-and-ranks-38df7d3a8566">Chartmetric Originals: Cross-Platform Performance (CPP) Scores and Ranks</a> was originally published in <a href="https://medium.com/chartmetric-blog">Chartmetric</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Gender Play Gap]]></title>
            <link>https://medium.com/chartmetric-blog/the-gender-play-gap-7f5e3a1b0066?source=rss-7a06e0067f24------2</link>
            <guid isPermaLink="false">https://medium.com/p/7f5e3a1b0066</guid>
            <category><![CDATA[music]]></category>
            <category><![CDATA[gender]]></category>
            <category><![CDATA[insights]]></category>
            <category><![CDATA[music-business]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Josh Hayes]]></dc:creator>
            <pubDate>Sat, 13 Apr 2019 22:54:46 GMT</pubDate>
            <atom:updated>2019-11-17T01:15:48.364Z</atom:updated>
            <content:encoded><![CDATA[<h4>Breaking down Men vs Women’s Performance on Music Charts</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iOp5x1ZhYF8IZUgB" /><figcaption>Photo by <a href="https://unsplash.com/@alicemoore?utm_source=medium&amp;utm_medium=referral">Alice Moore</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Billboard’s March edition proclaims “Women are Dominating The Billboard Hot 100.” Women have indeed claimed <a href="https://www.billboard.com/articles/columns/chart-beat/8504199/ariana-grande-lady-gaga-halsey-women-no-1-hot-100-2019">more no.1 spots on Billboard’s 2019 charts already than they did for all of 2018</a>.</p><p>Two years ago, <a href="https://www.billboard.com/articles/columns/chart-beat/7616906/leading-ladies-women-success-hot-100-chart">Trevor Anderson also tracked how female-acts have increasingly claimed the No. 1 slot of Billboard’s Hot 100 over decades</a>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K8LcoQVpcumD7rskei1eJg.jpeg" /><figcaption>BIllboard, “Leading Ladies: A Look at Women’s №1 Success in the Hot 100’s History”</figcaption></figure><p>When it comes to streaming, <strong>Spotify</strong> and <strong>Apple Music </strong>have built their own top charts to rival the longstanding <strong>Billboard Hot 100 </strong>chart. Because Apple and Spotify both cater to different markets, it is worth investigating how fans’ tastes differ across those platforms.</p><p>Casually flipping through the past few years’ chart rankings, you’ll immediately notice that the majority slots go to male acts. That seems counter to the recent buzz. So what’s up?</p><ul><li>How big is the gender divide for top performers?</li><li>How different are streaming charts from the more universal Billboard chart?</li><li>Have audience’s listening preferences changed over time?</li></ul><p>We can find our answers by looking to the stars! Well, if by “stars” you mean top performers on the charts.</p><h4>First, a few considerations:</h4><ol><li>Because lots of individuals can be credited on a song (multiple artists, or multiple members in a band) we’ll count just the primary song credit and for bands we’ll look at just the lead singer.</li><li>Spotify chart means <a href="https://spotifycharts.com/regional">Top 200 chart</a>. Spotify tracks the top 200 tracks but for consistency we’ll just look at the top 100. For this analysis, we took the data since July 2017.</li><li>Apple Music chart refers to <a href="https://appleinsider.com/articles/18/09/07/apple-rolls-out-116-daily-top-100-charts-to-apple-music">Apple Music Daily Top 100 chart</a>. The Daily Top 100 chart was released in August of 2018. We started tracking it in September.</li></ol><h4>A Gender Divide of the Charts:</h4><p><strong>How many women make it onto the charts at all?</strong> To answer this quesiton, we compiled a list of over <strong>104,000 data points over the last two years</strong> across Billboard, Spotify, and Apple Music.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PI1qHycJNCdjB9bYeSVmZw.png" /><figcaption>These pie slices are like when my Grandma is serving everyone and I ask for “just a small slice”… but each serving ends up pretty much the same. (Data covers 2017-07–30 to 2019-04–06)</figcaption></figure><p>The gender disparity is striking. Just like all the parties I threw in high school, the boys drastically outnumber the girls. Except that unlike my high school parties, all the people represented in these graphs are cool.</p><p>Let’s see how these ratios have changed over time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uXksWkShiFkYW30ULtKKoA.png" /><figcaption>Apple Music has some of the widest ranging values, despite having so little history.</figcaption></figure><p>Well that is interesting! A number of things to note so far:</p><ul><li><strong>Gender disparities are large across the board. </strong>On all three platforms, there is a significant gender divide. Each chart has been cruising close to the 80–20 split of men to women.</li><li><strong>Streaming is making Billboard rethink what it means to be Hot. </strong>Before June of 2018, the connection between the Spotify and Billboard charts was looser but from June on the two start to align a bit more. After Apple Music joins the party and we are moving into 2019, all three charts are starting to look very similar. Those patterns coincide with huge <a href="https://variety.com/2019/biz/news/u-s-music-industry-posts-third-straight-year-of-double-digit-growth-as-streaming-soars-30-1203152036/">growth in streaming</a> and sequential changes to the way that <a href="https://www.billboard.com/articles/news/8427967/billboard-changes-streaming-weighting-hot-100-billboard-200">Billboard accounts for streaming</a>.</li><li><strong>Female artists are growing their stream-share rapidly.</strong> Over the last 6-months, the weekly average of Spotify female performers on the charts has achieved roughly <strong>110%</strong> growth. On Apple Music we can see a whopping <strong>145%</strong> growth. That growth appears more significant and stable than the episodic gains that female acts had in the past. There are a lot of peaks and valleys across the charts for both genders, but based on the latest trends, my guess is that female performers will at least keep their new foothold even if their growth slows. It’s also possible they keep climbing.</li></ul><h3>Digging Deeper with Chartmetric Scores (CMS)</h3><p>Comparing raw counts of performers is a good start. However, if an artist shifts from No. 99 to No. 1, they still count the same. If an artist moves from No. 100 to No. 101 they no longer count. Moreover, not all chart-positions are created equal. It is harder to jump from No. 3 to No. 2, than it is to shift from No. 100 to No. 99. To see just how competitive the upper ranks are, look at a sample of song streams from a global chart:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/746/1*09l7pQpWBtn84WWPfn9UHg.png" /><figcaption>With this, we can make accurate estimations of a song’s total reach, even if that data isn’t available.</figcaption></figure><p>Notice how historically a No. 2 track will have many more streams than a No. 10 track which in turn will have more streams than a No. 50 track. Yet tracks around No. 70 — 100 will tend to have pretty similar stream counts.</p><p>To account for the exponentially higher success reflected by the higher ranks, we identified a formula that estimates a song’s overall traction (the red-line). Using that formula, we calculate a single score that allows users to easily compare the overall traction of an artist or a song. We call that measure the <strong>“Chartmetric Score” (CMS). </strong>(If you’d like a more in-depth discussion of CMS, you can read more <a href="https://blog.chartmetric.io/weighting-music-charts-grammy-2019-effects-or-not-8bb791dfba7f">here</a>.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WpFVpFivyNnqWF8klBcaDw.png" /><figcaption>Chartmetric Score formula</figcaption></figure><h4><strong>Chartmetric Score? Why is that interesting?</strong></h4><p>It is interesting because CMS allows you to track artists’ <strong>relative performance on the charts</strong> rather than just their <strong>inclusion on the chart</strong>. It’s like how FICO scores try to help you make sense of risk by plotting it in a single number on a linear line. Then, it is much easier to conceptualize what is happening.</p><blockquote><strong>At its simplest, higher CMS means an artist is doing great and lower CMS means an artist is doing less well.</strong></blockquote><p>With CMS, we can track how successful the average male and female artist have been, across our streaming platforms:</p><h4>Spotify Top 100 Global</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Yajp3ATDloKeFm3NDLVz1w.png" /><figcaption>Women’s scores oscillate widely around men’s scores, but have started trending upwards.</figcaption></figure><p>There are huge jumps on this graph for women (dotted line).</p><blockquote>When a female mega-star drops an album (like Taylor Swift or Ariana Grande) you can see women leap ahead. But when the spotlight isn’t on a star, women’s relative success drops really far.</blockquote><p>The good news for female artists is that women have frequent bouts of out-performing men. Moreover, in the past year or so Spotify’s CMS-values show a small but steady upward trend for female performers’ success and thus market share.</p><h4>Apple Music Daily Top 100 Global</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3_0pSnuJPreiyDNiumrWmw.png" /><figcaption>Notice how, after February, women’s trends start influencing men’s much more.</figcaption></figure><p>This is fascinating. That first half of the graph matches the patterns we’ve seen previously. Women move above and below, relative to men. Yet men’s averages hold fairly steady because they hold so much of the chart real estate.</p><p><strong>But around February 2019, things change.</strong> Some of that major jump for women is due to <strong>Ariana Grande’s <em>Thank U, Next</em> release</strong>. More than that though, Ariana Grande was at the spearhead of a hail of great female-led tracks. The entire top 10 pretty much went to Ariana Grande (Brava!), but other female artists burst through the door with her and landed all over the charts (like <strong>Halsey, Billie Eilish, Cardi B, Lady Gaga, City Girls, Ava Max, and Ella Mai</strong>). Women holding so many top chart positions then drastically impacts the average for men.</p><p>That’s something new and different.</p><p>Moreover, around February, <a href="https://chartmetric.io/charts/apple_music_daily?country=GLOBAL&amp;date=2019-02-15">you can observe that</a> the Apple charts were much more evenly distributed by male and female artists across the entire chart. It wasn’t just one or two female rockstars amid a sea of dudes.</p><blockquote>That degree of equality on the charts is rarely seen.</blockquote><h3><strong>What We Learned From Spotify and Apple Music Charts</strong></h3><ul><li><strong>When women succeed, they succeed big.</strong> Across both platforms, when women enter the top 100 they tend to surge to the top of the charts. Either it is harder for female acts to linger further down the list, or perhaps there are just more male acts that linger in the top 100 but don’t reach the peak of the pile. Perhaps our next analysis should be on the lifespan of songs by male/female artists. Either way, when female artists hit they hit home-runs.</li><li><strong>Success is more concentrated for female performers.</strong> Male artists tend to have more conservative and stable reach, but women are less likely to make it into the charts overall. The averages for men are drawn down partly because male-led tracks routinely populate the lower portions of charts. For every male-led track with runaway success in the top-10, there is often a handful of male-acts trailing along in the lower ranks also.</li><li><strong>Hot and cold.</strong> It is not uncommon for women to have phases where they outperform men in regards to streaming reach and, thus, profitability. Neither gender has been simply winning out-right.</li><li><strong>Status quo no more. </strong>The old patterns don’t look like they are going to keep holding. Female performers’ success on Spotify has been trending up for the past year. On Apple, women just recently upset the gender balance all together.</li></ul><h3>Comparing Across Countries and Genres</h3><p>If we want to get an idea for how influential gender is across music markets and genres, we can turn to <a href="https://developer.apple.com/documentation/applemusicapi/get_catalog_top_charts_genres">Apple Music’s genre chart</a>. We have been tracking this chart since Sep. 15th, 2017.</p><h4>Again, a few considerations…</h4><ul><li>We have data on more than 100 countries, but we will only look at the top five music markets: United States, Japan, Germany, Britain, and France</li><li>There are 19 genres available, but that’s a lot to take in at once, so we picked 5 of the largest and most polarizing: All Genres, Pop, Hip-Hop/Rap, Country Music, and Latino.</li></ul><p>I could play with this data forever. With just 5 countries and 5 genres, the following graphs contain info on over 900,000 chart positions.</p><p>First, let’s take a look at gender ratios across top music markets:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iB5H6tRUKiHWA0cpqo_PLA.png" /><figcaption>Apple Music country and genre rankings (All Genres) from 2017-07-17 to 2019-04–11.</figcaption></figure><p>You can instantly notice that women hit the charts more often outside the US market — compare 24.9%(U.K.) with 14.2%(U.S.). By the way, Japanese listeners are really digging female artists — 素晴らしい (<a href="https://www.google.com/search?client=safari&amp;rls=en&amp;ei=3VqyXJ2ICca5-wTj8rnYBQ&amp;q=%E7%B4%A0%E6%99%B4%E3%82%89%E3%81%97%E3%81%84+translate&amp;oq=%E7%B4%A0%E6%99%B4%E3%82%89%E3%81%97%E3%81%84+t&amp;gs_l=psy-ab.1.0.0i67j0i4l7.723.3615..4949...1.0..0.422.942.0j1j0j1j1......0....1..gws-wiz.......0i71j0.k6-ZLnHhK5E">translate</a>)!</p><p>Now, let’s look at this gender comparison genre by genre, beginning with Pop.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZZZL5FNHMge6SsCAEsOfGQ.png" /><figcaption>Apple Music country and genre rankings (Pop) from 2017-07–17–2017 to 2019-04–11.</figcaption></figure><p>With power performers like <strong>Ariana Grande, Rihanna, Camila Cabello, and BLACKPINK</strong> we could have expected the ratios would be more balanced here. Women are clearly the Queens of Pop. It is the genre where they shine the most.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8ynwj2MX3EKvIKcuOE33NA.png" /><figcaption>Apple Music country and genre rankings (Country) from 2017-07–17 to 2019-04–11.</figcaption></figure><p>American Country music is received very differently in its domestic market compared to foreign ones. Women are killing it on the Country music charts outside the US. <strong>Johnny Cash</strong> still tends to have the most tracks on the French and German charts, but otherwise its mostly ladies leading the pack: <strong>Shania Twain, the Dixie Chicks, Taylor Swift</strong>, and <strong>Dolly Parton</strong> are adored. Newer male country acts like <strong>Eric Church</strong> and <strong>Zac Brown Band</strong> struggle to land.</p><blockquote>Perusing the Japanese “country music” charts is almost synonymous with just looking at <strong>Taylor Swift</strong>’s discography.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hq5dvD-ApaEDitGzyAsLMA.png" /><figcaption>Apple Music country and genre rankings (Hip-hop/Rap) from 2017-07–17 to 2019-04–11.</figcaption></figure><p>Hip-Hop/Rap is the hardest genre for women to make headway in. American and French listeners seem very resistant to embracing female artists. Other markets are more responsive… just, not that responsive. For the past few weeks, <strong>Cardi B</strong> and <strong>City Girls</strong> have been doing the heavy lifting for defending women’s foothold in these genres.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*11KrdtvkhPtffy-adEBg2Q.png" /><figcaption>Apple Music country and genre rankings (Latino) from 2017-07–17–2017 to 2019-04–11.</figcaption></figure><p>Recently, these charts have been filled with numerous tracks from <strong>Bad Bunny, Ozuna</strong>, and <strong>Anuel Aa</strong>. It’s hard for any other artist, male or female, to break the hold those megastars have on the Latin charts. Also interestingly, many songs in this genre are collaborations and feature prominent women artists (Cardi B, Selena Gomez, Demi Lovato), but the first billing still tends to go to male artists.</p><h3>Final Thoughts</h3><p>How do you think these trends will change going forward?</p><p>Perhaps the growth of streaming, and women’s tendency to drop mind-blowing tracks has started opening new doors. Will women continue to get more of the coveted chart positions? Or after the recent reign of Ariana Grande and Billie Eilish, are we due to see fans pendulum-swing back to more male artists?</p><p>Apple Music and Spotify trends give us some insights, but of course music streaming is a wide world with many unique platforms. Still, the big changes taking place suggest that streaming-growth is disrupting old listener patterns. You should keep your eye on how audiences are engaging with top performers because “the times, they are a changin’.”</p><p><em>Hungry for more? Let </em><a href="https://blog.chartmetric.io/@jasonjoven"><em>Jason Joven</em></a><em> keep you in the loop on the music industry’s latest data gems by tuning into our podcast: </em><a href="https://chartmetric.transistor.fm/episodes"><em>Your Daily Data Dump from Chartmetric</em></a><em>.</em></p><p><em>Or if rolling up your sleeves and getting messy with data is more your style then head over to the Chartmetric website and grab your </em><strong><em>free account</em></strong><em> </em><a href="https://chartmetric.io/signup"><em>here</em></a><em>.</em></p><p><em>Or you can shoot us a message at hi@chartmetric.io.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7f5e3a1b0066" width="1" height="1" alt=""><hr><p><a href="https://medium.com/chartmetric-blog/the-gender-play-gap-7f5e3a1b0066">The Gender Play Gap</a> was originally published in <a href="https://medium.com/chartmetric-blog">Chartmetric</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>