In August 2022, a team of Water & Music members began aggregating data from music marketing campaigns, in an attempt to create a first-of-its-kind public database of ROI benchmarks.
While we estimated a two-month timeline to collect data and summarize our findings, we’ve since encountered numerous, seemingly insurmountable obstacles in seeing this project to completion. Research hurdles are hardly unique. But we believe the delays around this project have become as much a part of the story as the intended output — and reveal endemic problems with how the music industry measures growth and success.
In this piece, we disclose the setbacks and missteps we faced in this research process, plus some initial conclusions on where the project could potentially head in the future. We also want to make a clear call to action on how the music community can help. As we explore new collaborative research models, we believe sharing our journey can open the door to new contributors within the community. Water & Music is unique for its decentralized structure powered by network effects, where the efficacy and robustness of our research scale in proportion to the number of people participating in it. This holds especially true for a database-centered project like this one, where the benefit of 10 rows of data pales in comparison to the fidelity of 100.
Thus, the goal of this article is to inspire greater support and invite new subject matter experts to help us build a marketing ROI guide that can serve as a foundational resource for artists and their teams.
Our initial aim was to create a database of the most commonly utilized marketing channels in music promotion – focusing on what musicians should expect to pay and what ROI they should look for in return, with the ultimate goal of greater transparency.
For the purposes of this project, we define ROI as the bottom-line financial return on one’s spend. For example, if an artist spends $1,000 on Instagram Ads, an expectation of a 1.3x ROI would mean that the artist would expect to receive $1,000 x 1.3 = $1,300 in incremental income in return, attributable to that ad spend.
We planned to compile music campaign data in bulk — first by seeding data directly from a curated list of marketing agencies and freelancers, which we hoped would inspire individual artists and management teams to contribute their individual campaign data in turn. This two-pronged plan would result in an open and shared, yet anonymized, data repository from which to glean greater insight into what to expect when developing a music marketing budget.
While it’s impossible to give an exact ROI measurement considering revenue streams amongst many other variables will differ from artist to artist, we assumed that agencies would have originated their own internal, standardized benchmarks for approximating connections between bottom-line revenue with top-line campaign metrics, for the purpose of business planning at both the artist and agency level. As an example, assumptions and benchmarks would look like the following:
- Revenue streams – The average artist has 4 main revenue streams: streaming, live performances, merchandise, and fan clubs
- Annual revenue from a fan – A marginal “true fan” (defined by someone who would participate in the above revenue streams) is worth $100 on a yearly basis
- Churn rate of a fan – Each true fan has a churn rate of 2 years, thus every true fan has a customer lifetime value of $200
- Follower to fan conversion – 1,000 followers on Instagram is equal to 10 true fans
- Dollar to follower conversion – Paying $1,000 for IG ads optimized for followers typically results in ~1,000 followers, for a true fan acquisition cost of $100
- Final ROI calculation – If our customer acquisition cost is $100 and our customer lifetime value is $200 each, then our bottom-line ROI is 200% or 2x
Initially, we intended to segment the database by artist and budget size, to account for significant variables affecting ROI. We also planned to divide the data across multiple tables to differentiate between ROI across different channels, including but not limited to audience growth, live performances, and radio & PR, since they each accomplish different goals and take on a life of their own.
In all, our original database resembled the screenshot below:
After several months of research, we’ve arrived at a new perspective such that we plan on restructuring the database to address the obstacles mentioned above. We’ll use data estimates in place of campaign-specific figures, incorporate different artist goals that more accurately represent an artist’s career, and style the data as a decision tree as opposed to a traditional spreadsheet.
But before unveiling the reformatted model, it’s important to understand what roadblocks arose along the way and how those precipitated our data model adjustments.
Roadblock 1: Does the data exist to create marketing ROI benchmarks?
To begin building the database, we interviewed various music marketing experts, notifying them that we would be sending a Google form to collect our desired data. We expected marketing agencies and freelancers to track campaign results and log them in SQL databases. Every column in the screenshot above would be captured through a SQL query, meaning the contribution of every interviewee would be ten or so queries. If we were interviewing a freelance music marketer, we expected a more eyeballed/estimated ROI based on data kept in a Google Sheet or CRM.
However, from conversations with Water & Music members who were domain experts, we quickly realized that campaign data is much less available than expected. Thus it became a part of our interview process to learn more about interviewees’ data practices, if they existed at all.
We found that data tracking and analysis in music marketing is less organized compared to other industries, and that data processing is uncommon altogether. Of those interviewed, zero participants employed a SQL database or CRM to track campaign results, and only a handful tracked data via Google or Excel spreadsheets. Only one company we interviewed, Gupta Media (the largest company we spoke to), had a proprietary database setup. The rest of our experts only looked at data to measure results via the respective ad platforms (e.g., Meta Ads, Google Ads, etc.).
To be clear, we do not mean to criticize the people we spoke to, nor negatively evaluate the effectiveness of their marketing ability. Rather, we simply intend to give color to one of the principal findings of the project — namely that the lack of data hygiene is structural, and a product of the culture and economics of the music industry.
Discovering a sample-wide lack of data was an early threat to the project’s direction. While we could ask our interviewees for ROI estimates, if our benchmarks were merely a product of gut feelings rather than hard, campaign-backed data, would our database actually provide value?
Given the limited information available for artists and managers regarding what to expect when spending a marketing budget, we decided that even a rough estimate would provide some baseline in an industry where artists are so often sold the dream with nothing to show for it. Further, providing some level of expectation can go a long way in helping artists navigate an opaque landscape and combat bad actors.
Moving forward from this initial obstacle required us to unwind some early assumptions regarding data hygiene in music marketing campaigns and accurately reframe the value of our findings.
Roadblock 2: Is the music industry too unique to create marketing ROI benchmarks?
The lack of concrete data led to an unforeseen point of friction with our marketing experts, who consistently stressed the difficulty in estimating ROI due to artists being too different from one another to give an accurate approximation. Some even went as far as to say that trying to establish practical benchmarks is a completely futile task altogether. The Water & Music community shared similar sentiments internally, with one member arguing that as opposed to other industries where ROI benchmarks may be viable, quantifying ROI is impossible because “music is a completely different beast.”
As the project lead and originator of the project’s concept, this felt wrong to me. There are benchmarked CPM rates across many different industries that one can easily search on Google, and I was highly skeptical that music would be dissimilar. Everyone likes to think their industry is special, but is music truly different? And if so, is it distinct in such a way that makes ROI unquantifiable? Despite best efforts to convince people otherwise and due to the sheer number of voices expressing similar viewpoints, I endeavored to find some prevailing logic one way or another.
After interviewing multiple marketing experts across music and various other fields, I was proven wrong: Music truly is its own animal.
While there are a few interesting tidbits that make music unique from a marketing perspective that will be expounded upon in a follow-up article, the most essential differentiator is that music is more omnipresent than other products. As a result, the number of touchpoints that lead someone to become a fan is endless.
Think about where you can experience a new artist or their music: Your room from your computer, the gym over the loudspeaker, an Uber from the driver playing their playlist, the list goes on. Compare this to a classic Consumer Packaged Goods (CPG) product like Tide Pods, which is, conversely, very restricted in terms of its range of potential exposure. The fewer opportunities a product has of being seen or experienced, the simpler its marketing strategy, and the more reliant the product’s marketing team is on traditional means of promotion (advertising, influencer, etc.).
Due to the diversity in potential paths to discovering an artist, each audience’s behavior is inherently novel. Unlike other businesses, there aren’t guardrails to herd people towards similar behavior, which lead to a lack of repeatable customer patterns that develop to determine a predictable return on effort spent. Thus the question had to be asked, is assessing expected ROI in music marketing even achievable – and if not, is this project worth continuing?
After careful consideration, we concluded that although ROI is not a meaningful pursuit if trying to attribute a price and a value of a “true fan,” there is still value in continuing the project if we restrict its scope. Instead of focusing on bottom-line ROI, we decided to shift our attention to the very top of the funnel, i.e. to channels that drive wider awareness of an artist or brand. Where it might be fruitless to try to decipher a financial return, there’s still merit in identifying how many top-level streams an artist should receive in exchange for spending, for example, $1,000 dollars.
Thus, we decided to redefine our goal and establish a new mission statement that is less quantitative and more qualitative: “This project exists to start the conversation around limited data in music marketing and provide a framework for artists and their teams, exclusively from a top-of-funnel perspective, so that there is more transparency around what to expect from their marketing budgets and why.”
Roadblock 3: Are there too many variables to create marketing ROI benchmarks?
With our new directive, we conducted more interviews, before bumping up against additional feedback that our channel segmentation was not nuanced enough to account for the variability between artists. In response, we pivoted from formatting our ROI estimates as rows in a spreadsheet to a more multi-dimensional decision tree (approximated in the example below). After an internal brainstorm and more feedback from continued interviewing, we concluded that we would segment as follows:
- Goal – Content-based revenue, Ticket sales, Direct-to-Consumer revenue, Social media growth
- Platform – Spotify, YouTube, Tik Tok, Instagram, Soundcloud, Twitch, PR
- Marketing Channel – Ads, Influencer/playlisting, Collaboration
- Ads Platform Objective – Acquisition, Engagement, Traffic
- Artist Size – Small, Medium, Large
- Territory – US/Canada, Europe, Asia, Latin America, Other
- Genre – Hip-Hop, Pop, Rock, Country, Electronic
However, as our segmentation became more sophisticated, we quickly realized that our participants would have to exert effort proportional to the added complexity when submitting their estimates. Rephrased as an example, if we have a three-layer decision tree that’s comparing results across four (4) marketing channels on two (2) social media platforms in three (3) territories, that’s 24 ROI estimates we’d need to collect from each participant (4*2*3 = 24). If we then wanted to account for artists in five (5) different genres as well, we’d have 120 estimates we’d need to collect from each respondent (4*2*3*5 = 120). In other words, each variable (e.g., goal, artist size, territory, etc.) exponentially increases the number of ROI estimates to collect in order to accurately represent the differences between artists.
If we take the seven-layer segmentation we’d arrived at above, each participant would have to submit ~20,000 different permutations of ROI estimates, obviously completely infeasible to ask of our participants. To find a happy medium between catering to the nuances in artist makeup and not asking too much of our participants, we settled on three segments with a reasonable 27 permutations: Artist goal, platform, and marketing channel. The resulting decision tree now looked like this:
The final roadblock: Will there be enough industry participation to create marketing ROI benchmarks?
Now that we’ve addressed the obstacles above, we finally have a framework to request and present music marketing ROI meaningfully. Only one problem remains: The number of sources providing us with our ROI estimates.
So far, we’ve interviewed 15 experts. While this is a great start, more than 15 estimates are necessary to create a robust industry average and add confidence to the directional correctness of the ROI figures.
A high response volume is essential to sanitize the inherent messiness of the differences between marketing channels. For example, influencer marketing doesn’t offer a clean output. Often described as a lottery, if some campaigns “hit” and go viral, it’s challenging to average out successes with failures; estimates from memory add another layer of inaccuracy. Standardizing this data would detract from such nuance and result in a loss of informational integrity. Therefore, the only solution is to leave the responses qualitative, and have patterns emerge from a high volume of quantitative input.
Trying to deliver actionable insights on artists’ marketing ROI is a mountain of a mission. With all the unforeseen roadblocks, lack of data, and near-infinite number of variables, it’s no wonder an undertaking like this has yet to be published. And we’re under no pretenses that our final presentation will be well-tailored to each kind of artist. As the marketing landscape evolves, we know it will take a village to build and maintain this resource. No one can be an expert in every component of a field as intricate and complex as music marketing.
If you have read this far and have music marketing experience, we would love your help in taking this project across the finish line. We are planning to have this database live as an iterative resource that updates in parallel as the music marketing meta changes as well. Thus, it’s not too late to offer us your support regardless of when you may be reading this.
To contribute, please introduce yourself in the “Artist Marketing ROI” thread under the “collab-databases” channel in Discord, or email me at firstname.lastname@example.org with a sentence or two on your background and depending on your preference and availability, we will either set up an informational interview or send you a Google Form for you to submit ROI estimates.
Shoutout to the people who have helped get this as far along as it’s gone, and a special thank you to those who have been generous enough to interview with us so far.
Artist Marketing Project Research Team Members (listed alphabetically):
- Andrew Apanov
- Brodie Conley
- Bryan Kim
- Chrissy Greco
- Danny Scheiner
- Diana Gremore
- Dorothée Parent-Roy
- Jason Feinberg
- Joseph Hennessey
- Kristin Juel
- Stephanie Guerrero
- Tyler Budd
- Yung Spielburg