Key takeaways from our webinar on music, data, and storytelling
Last Wednesday, September 25, we hosted our first member webinar of the year, where we walked through our popular report, The State of Data in the Music Industry.
Co-presented by W&M founder Cherie Hu and lead report author Michael Zhang, our session brought together over 130 of our members to explore current-day challenges and opportunities at the intersection of music, data, and storytelling.
Below, we recap the key themes and discussions from the webinar, along with some bonus responses to audience questions we couldn't address live. You can follow along and watch back the webinar recording in the embedded video.
Balancing storytelling with data accuracy
One of the most compelling questions from our audience came at the very end:
When interpreting data, there could be so many perspectives (especially when music is linked with personal tastes and experiences). What's more important — storytelling versus finding out the truth?
Michael and Cherie emphasized that while accurate data analysis is crucial, the music business thrives on human relationships. Data should enhance these connections, not overshadow them.
The ability to craft a compelling story that distills complexity into clear, emotionally resonant insights is what ultimately drives impact — especially when communicating with artists and executives who may not have a technical background.
Conversely, a powerful story gains credibility when rooted in rigorous, thoughtful analysis that examines the data from multiple angles, ensuring no critical perspective is overlooked.
Limitations of data-driven A&R
The world of A&R exemplifies the delicate balance between storytelling and analytics.
Data analytics tools have proliferated in scouting, with all major labels now housing in-house A&R research teams filled with trained data scientists. However, we cautioned against overreliance on these platforms without considering the human relationships that underpin artist development, especially for emerging talent.
Michael stressed that there's no one-size-fits-all approach to A&R, and that understanding why numbers move is essential. While data can highlight trends and anomalies in artist growth, it may miss the subtleties of an artist's potential. For example, the team that surrounds an artist can make or break their success, regardless of their surface-level commercial performance.
Interestingly, there's growing momentum in using predictive analytics for marketing strategies, rather than solely for A&R decisions. Cherie mentioned companies like SymphonyOS and Naviro that are investing in tools to automate audience research and simplify music campaign development.
Platform gatekeeping and data access limitations
A recurring theme was the challenge professionals face in accessing actionable data from major music platforms, particularly streaming services. Many attendees felt that strategic gatekeeping by these platforms hampers artists and their teams' ability to perform meaningful analyses and make informed business decisions.
Our audience members called out Spotify several times on this front, even though the company has the most robust music analytics offering of any streaming service. While the company offers some audience segmentation to indie artists and labels, their API access and data export capabilities are much more restrictive.
One audience member likened the available data to "smoke and mirrors," pointing out the resulting difficulty in building custom in-house data solutions. Others mentioned resorting to web scrapers due to the absence of official API solutions, highlighting the lengths teams will go to access necessary data.
Educating artists about data practices
Section 6 of our State of Data report highlighted a significant gap in data education, with less than a third of respondents receiving data literacy training from their companies.
Extending this industry education to artists themselves presents an even more nuanced challenge. Conveying the value of data to artists requires a tailored approach that respects their creative focus, while still offering meaningful insights.
Michael observed that many artists perceive algorithms — like those of TikTok or streaming platforms — as mysterious forces controlling their success. This "black box" perception can lead to industry misconceptions and wider career anxiety.
To bridge this gap, it's essential to explain the underlying rationale behind data-driven strategies to artists, without overwhelming them. For instance, instead of simply advising artists to create 20 TikTok posts before a release, one can explain how consistent posting increases the likelihood of meeting the platform's algorithmic thresholds for virality. This approach helps artists understand that they're making a strategic decision to engage with a level of variance and risk to reach potential long-term fans.
Artists should also be encouraged to interpret metrics holistically. Sudden spikes or drops in streaming numbers and social engagement can result from various factors — across ad campaigns, viral content, playlist placements, or even bot activity. By considering multiple possibilities, artists avoid attributing success or setbacks to a single cause, leading to more informed and proactive decisions.
The emergence of a community analytics paradigm
While fan clubs have been around for ages, community analytics is still relatively new in the music industry.
Community manager roles surged during the COVID pandemic, as artists recognized the strategic importance of keeping fans connected online through apps like WhatsApp, Discord, and Telegram. Since then, we’ve seen a new crop of well-funded, music-first fan community platforms emerge to serve artists across the career spectrum, including Medallion, Mellomanic (f.k.a. We Are Giant), and Fave. Anticipating their own disruption, major labels like Universal Music and Warner Music are also putting their weight behind their direct-to-fan business — investing in new in-house fan data infrastructure and, in some cases, building their own superfan apps.
Amidst this hype around superfans, our discussion touched on the need for frameworks that define what a healthy, thriving community looks like for artists today. This involves moving beyond mere artist-to-fan engagement or clicks, and focusing on horizontal, fan-to-fan interactions and group behaviors. It also means considering aspects beyond music consumption and sales, diving deeper into healthy psychology around identity, lifestyle, and belonging.
Cherie shouted out companies like Levellr that are pioneering analytics for mass-market community platforms like Discord and Telegram. Their analytics cover not just fan engagement, but also community-specific needs like moderation efficiency, e-commerce monetization, and deeper insights into fan behavior.
Of course, investing in these community tools presents its own challenges. For instance, many webinar participants mentioned WhatsApp as a significant platform for marketing and fan engagement, especially in regions like Latin America and the Caribbean. While effective for reaching fans, there are challenges and privacy concerns with exporting data from this channel, and making it interoperable with DSP customer data.
BONUS: Addressing unanswered audience questions
“What, if anything, can be done to pressure streaming platforms to offer open APIs or CSV exports of actionable data?”
Michael: Ultimately, I think it comes down to incentives. What incentivizes DSPs to offer — or more importantly, prioritize — building open APIs or CSV exports of the data you want? There's a lot of cost associated with creating and maintaining APIs, and there are a lot of decisions that necessarily go into exposing raw data in CSV form.
So what's in it for DSPs? A few pathways that seem possible, albeit unlikely, for this to occur:
- Regulation mandates some level of equal or open access for DSPs to provide around their data.
- Majors and the rest of the label/distributor/artist ecosystem all have their own levels of influence they can exert on DSPs to demand data. Different parties can prioritize requests for data, but unless the ask is universal, it's hard to see how this would result in a completely open API.
- DSPs decide to charge for data access and services on their own platform... The poor optics and backlash is probably too high of a barrier, but this certainly creates an incentive to supply data.
“Has anyone found a reputable source for sync data?”
Cherie: Publishing market data in general is hard to find, for reasons that are endemic to industry infrastructure and politics.
But this is an important question because sync is one of the hottest areas of music licensing right now, especially with the recent lawsuits that music rights holders have filed against Marriott and the NBA. There's a question of: Could technology play a role in making sync licensing deals easier to source, close, monitor, and pay out?
There are a lot of new sync marketplace startups that are launching in this vein, such as Catalog, Acrylic, and Cipher. There is also a site called TuneFind that maintains detailed databases of music syncs in movies, TV shows, and ads.
For those looking to benchmark the performance of a given sync against various kinds of channels or campaigns, TuneFind can be a great place to start — e.g. exporting a list of song IDs from their database, and then tracking their performance in some other third-party market intelligence tool such as Chartmetric or Soundcharts.
“I’d love to hear more thoughts about, besides the risk of disputes arising, why there is a lack of transparency with publishing metadata.”
Michael: A couple of disparate thoughts. There are many reasons (competitive, contractual, etc.) why people wouldn't want to show exact splits publicly. And who has a reason to host a centralized, meticulously and consistently maintained source of truth for publishing metadata, with the central aim of being completely comprehensive? It makes sense for large rights holders to want to maintain a clear source of truth for themselves, because they have customers for that information.
Also, as with pretty much all data that depends on manual entry, there's a lot of gaps and missing data and a lot of parties and formats in which they get passed around (again manually), outside of DDEX-required formats.
Cherie: Building on Michael's points — I'd add that the lack of transparency in publishing metadata is deeply rooted in the industry’s own structural complexities. Songs today often involve multiple songwriters, publishers, and rights holders spread across different territories, making it a daunting task to aggregate and maintain accurate metadata. This complexity is compounded by legacy systems and outdated infrastructures that many organizations still rely on, much of which is ill-equipped to handle the dynamic and global nature of today's music consumption.
And the strategic element is critical. Many industry entities have a strong incentive to withhold detailed metadata to maintain a competitive edge or protect proprietary information. All these factors create a fragmented ecosystem where achieving full transparency is not just challenging, but sometimes inadvertently discouraged.
Our previous reporting digs deeper into these challenges, and efforts to ease the pain of managing copyright data in the long term.
“Have we seen any evidence for a need/hunger for audience modeling? i.e. building out models that capture various audience personas that ‘react’ to demos, upcoming releases, to try to predict how a particular audience may or may not respond over time? Are people working on this?”
Michael: For sure there's high interest in audience modeling! All the talk around "superfans" is one of the prime examples of this. Plenty of people are also looking to segment audiences based on listening and engagement activity (some DSPs are starting to expose these views too), and there's also natural segmentation of audiences across different platforms, etc. to take into account as well. Growing a particular platform's following often requires specifically tailored modeling.
There are also common tactics used today to engage different segments of fans that don’t rely on data. For example, sharing additional content in a Discord or via a separate fan page — in these cases, you’re using fans’ self-selection behavior to “target” them. The same thinking can even apply to things like email lists!
Cherie: Check out Naviro, which is working on solving this exact problem of audience modeling for the music industry.
What’s next
Thanks so much to our speakers and participants who contributed to this enlightening session. Your insights and questions drive the growth of our collective understanding of how the industry works, and how it can improve moving forward.
Stay tuned for upcoming webinars and articles where we'll dive deeper into the world of AI, including high-level trends such as data marketplaces and startup funding.
Thank you for being an integral part of the Water & Music journey!
For any further questions or to share your thoughts, feel free to reach out to our inbox at members@waterandmusic.com.