Why Spotify Analytics Aren’t Everything (yet)

“100 People Listening Right Now.” Those haptic vibrations every time your new track gets a stream—a constant influx of aesthetically pleasing graphs, charts & other visualization. I’ve even heard an artist compare the addiction of checking their Spotify analytics to crack.

But after having Spotify for Artists at our fingertips for the last five years, I think we need to start discussing what value, if any, is hidden in these dashboards for an artist and their teams.

I spend most of my time creating & interacting with similar web dashboards for my day job at a large CPG company. Oversaturation of information is an issue that’s become increasingly apparent during the growth of data visualization thru platforms such as Tableau or Microsoft’s PowerBI. In the name of freedom, these services allow companies like Spotify or Apple Music to provide the end-user with a wealth of data and statistics but provide few insights or lessons on interpreting or leveraging this information.

Before we dive into Spotify for Artists specifically, we need to discuss the four overarching types of data analytics that can be employed:

• Descriptive: tells you what happened in the past

• Diagnostic: helps you understand why something happened in the past

• Predictive: predicts what will likely occur in the future

• Prescriptive: recommends actions you can take to influence the future outcome

The spectrum that differentiates descriptive & diagnostic versus predictive & prescriptive looks at the past versus the future. The difference between descriptive & predictive versus diagnostic & prescriptive is telling you static information versus telling you the why or how something occurred. 

Depending on your goal of working with your data, the two most impactful will likely be Diagnostic & Prescriptive data analytics. Diagnostic is helpful when there is an event, maybe an album release, an extensive marketing campaign, or a tour, where you are trying to understand the impact on the output or your streams. Prescriptive is the largest value add because it helps you make decisions that will impact your future. Perhaps you can use the data at your fingertips to forecast ticket sales for each market or understand how much merchandise you need to purchase for each show, including the correct ratio of shirt sizes. Or maybe Prescriptive Analytics helps you properly budget your release campaign, splitting money between an in-person pop-up, Facebook ads & an influencer campaign on TikTok.

The danger with these analytics types understands that correlation DOES NOT equal causation. A tired & real lesson you might learn in math or science class in school, but probably the key to understanding these insights as Spotify and other platforms continue to evolve.

Descriptive analytics might allow you to draw your correlations, but they will never determine the causation statistically. Diagnostic & Prescriptive Analytics are far more technical measures, typically leveraging statistical knowledge ranging from simple (linear regression) to complex (neural networks, machine learning). These provide you the correlation between two variables to tweak the inputs to create the causation you prefer.

Analytics that provide insights, rather than force you to draw your own, unsubstantiated causation, are where value is genuinely added with data.

Spotify, at the moment, only provides Descriptive Analytics. It provides a simple postmortem that allows you to attribute your streams to the first degree. You can see whether those streams were driven by playlists, your page, or a user playlist. You can see what day you had the most streams. You can even see a loose demographic breakdown of who is streaming your music. But Spotify refrains from telling you what any of this means, why it’s happening or how to make any of it happen again.

That’s not to say there’s no value provided by Spotify for Artists. It’s still substantially better than anything a DSP or platform has provided artists in the past, and it’s done an excellent job of engaging artists and their teams. It’s also done an excellent job of highlighting Spotify playlists’ power, especially the algorithmic playlists for smaller artists, a product that Spotify is ready to start monetizing. 

But it’s also led many astray, making some artist teams focus on increasing those numbers without really understanding what’s driving them. If you have a release that sees massive success, with most streams coming from Spotify Algorithmic playlists & your page - how do you go about recreating that success. Some artists create similar songs. Other artists try to make songs that sound more like the recommended artists on their page. Some artists specifically collaborate with those recommended artists. Others try to create songs that sound like they’re made for a playlist in-mind. The result of all of these attempts is a self-fulfilling prophecy. It’s a classic case of correlating the song’s sound, or whatever factor seems vital in your head, with success. When, in fact, it could be something entirely separate.

Whether or not these attempts succeed, the artist has trapped themselves in a perpetual cycle. They are trying to recreate quantitative success without the ability to deduce where the success is coming from.

Perhaps that release gained traction because an influencer posted it on TikTok, or maybe a specific geographic region supported the song on Instagram or Twitter. It redirected new fans to your page directly. Perhaps you only got on the algorithmic playlist because no other artists with a similar sound release music on the same day, or maybe it was a complete chance that a Spotify curator noticed your track.

By Spotify only showing an artist the stream drivers at the most basic level, the first degree, it becomes impossible to use this data to make an informed decision. Instead, some teams use this data to justify their every whim when there is possibly no relationship between the two variables in your head.

How many of your peers were overconfident about the outcome of the presidential election several weeks ago? No matter your political affiliation, this is a clear case of misunderstanding the insights being provided.

Descriptive Analytics are accurate, but they might not be telling you what you think they are. If your audience is 70% male, that doesn’t mean focus only on male fans. It could also be interpreted that you need to appeal more to female fans, as it is an untapped market. If you are only showing massive growth in one region in the US, it doesn’t mean only route tours through that area. It could also mean finding an opening slot on tour in another part of the country to continue growing your fan base.

So what should you do when making decisions in the future? 

I’m sure Spotify is working to hard to build out more complex insights, but until then, take a step back from the data for a few minutes. Think about how you would have tackled the same question before you got inundated with data & numbers. If you saw success on an early track or a release before 2016, consider the fact that when you created that song, you didn’t have access to any analytics or data, that success came entirely from whatever you did well. If you see success trying to follow the numbers, be careful that you don’t find yourself trying to sell others or yourself something you don’t believe because you could misconstrue the numbers in a particular light. 

Data is power! But you need to interpret, understand, and plot actions based on that data to keep the lights on.