Why a COVID data strategy for live music doesn’t exist (yet)
This report started as a statistical analysis looking at correlations between COVID data and live event activity in the U.S. — but quickly expanded into a larger, qualitative study of the live music sector’s fraught relationship with COVID data and regulations, and why a COVID data strategy has been largely absent from live music companies to date. In the hopes of making our findings more accessible, we’ve structured this report as a dialogue between the two of us, broken into two parts: An analysis of live event activity and COVID data in partnership with Bandsintown, followed by takeaways from our qualitative interviews with professionals across promoters, agencies, venues and more. We invite any feedback you have on our research and will update our analysis with more findings as the pandemic develops.
Brooke: We’re over a year-and-a-half in, but the global COVID-19 pandemic continues to be the most important issue affecting the economy, especially the leisure/hospitality industries and particularly the live music sector.
In mid-April 2021, I proposed to Water & Music that we study the relationship between COVID data and the return of live events across the U.S. The main motivating factors for this study were significant shifts in consumer behavior, health data and government funding for live shows at the time. Vaccination card selfies were clogging Instagram feeds, a third wave of COVID stimulus checks was driving up national consumer sentiment, and research was revealing strong positive correlations between vaccination rates and interest in domestic travel and in-person concerts. Meanwhile, billions in grant applications from the Save Our Stages Act were being submitted and funded, enabling indie venues to stay afloat at least temporarily, and COVID cases and fatalities were finally, mercifully, on a steep decline.
In turn, these developments spurred a flurry of economic activity in live music throughout the spring and early summer, including but not limited to the relaunch of major festivals like Lollapalooza, official date announcements for Outside Lands and Coachella and the return of regular programming at major venues like Red Rocks.
Because so much is at stake financially and ethically in the return of live music, we wondered to what extent agents, promoters and venues were making informed booking decisions based on COVID cases and vaccination data, in the same way that consumers were monitoring this data to inform their own spending behavior. Will there continue to be waves of COVID cases that shut down cities into the foreseeable future? Are live music professionals watching these potential waves and avoiding areas that are high-risk? And more scientifically, can we use vaccination rates or COVID cases to predict booking patterns, and will this data correlate with successful shows — success being defined as 1) the show happening in the first place, and 2) the show selling well?
Cherie: As someone who loves both data analysis and live music, I was really excited when Brooke proposed this project. I thought this would be a great opportunity to map booking/vaccination correlations in a way that would be not just scientifically interesting, but also actionable for live music professionals.
We envisioned a dynamic dashboard that would visualize and track county-, city- and state-level outliers in either direction — i.e. counties that had an unusually high amount of booking and/or consumer purchasing activity despite low vaccination rates, and vice versa. Such data could potentially provide helpful signals to agents and promoters as they reschedule or route new tours through the end of 2021 and beyond, especially with respect to highlighting second- or third-tier cities that otherwise wouldn’t be on their radar. If we found any interesting correlations in the U.S., we would then expand our analysis globally.
The initial results of our analysis turned out… differently, to say the least. But first, let’s share more about our initial methodology.
Part I. The statistical relationship between COVID data and live events
Brooke: First, we got to work finding a comprehensive, digestible source of live music event data that we could start comparing with publicly available data from the CDC on COVID cases, hospitalizations and vaccinations in the U.S.
Reliable live music data is tough to find for several reasons. Information on announced shows and ticket sales is reported manually by thousands of various stakeholders with varying degrees of timeliness and accuracy. A lot of this info is locked up in internal, proprietary spreadsheets and databases, and gaining access to this information often comes with a big cost that is not feasible for a one-off report. Unlike, say, MRC Data or the Billboard charts for recorded music, there is also no central, widely accepted authority of live music consumption data, not even for the sector’s own internal use. (The closest might be Pollstar, but their data service is expensive and all sales listed on their platform are self-reported by the promoter or venue, increasing the risk of incomplete or inaccurate information.)
Luckily, after a few conversations, we found a great data partner with Bandsintown, who in early June 2021 graciously agreed to provide some of the most comprehensive live music event data that exists in the U.S. Still, due to the proprietary and valuable nature of this data, we were not given raw numbers, but rather growth of various measures over time.
In particular, Bandsintown provided three key data measures for both industry activity and consumer sentiment, looking at percentage growth in each measure over different periods of time (month-over-month, year-over-year and vs. two years ago):
- Growth in live event announcements
- Growth in live events happening (actually playing out)
- Growth in customers clicking to purchase tickets to live music events
While Bandsintown has official pages and partnerships for practically all of today’s major artists, much of their event data is also user-submitted — which meant several hours of data-cleaning work to make sure that county/city/state names were standardized, so that we could start doing meaningful comparisons and correlations with the CDC’s datasets on a county level. This involved index matching cities with standardized geographical codes for merging and mapping the data, removing any observations with missing values from the CDC data (thankfully their data has been very robust) and doing some basic exploratory data analysis (EDA) to narrow down the most relevant variables that may tell our story. For example, we looked at the percent growth of event announcements and ticket clicks as they correlated on a county-level with COVID cases per 100K population, deaths/100K, vaccinations as a percent of the population and other relevant socioeconomic factors.
Once I narrowed down some key variables from the initial EDA, I ran some linear regressions to see if we could find anything predictive from the COVID data around where events were announced, consumers were purchasing and events were happening. Again, our hypothesis at the time was that more live event activity would be happening in areas where vaccination uptake was higher, and where COVID case and death rates were lower.
Cherie: When we first ran an analysis of Bandsintown data in June and early July 2021, we found almost no statistically significant correlations or predictive factors with COVID health data on growth in live event announcements, happenings or ticket clicks over time. The main reason for this is probably that vaccinations stalled in the U.S. right around the time that we started diving into our data.
The Biden administration had set and easily met the goal of administering 200 million doses by the President’s first 100 days in office (April 28, 2021); evidence of hours-long waits for vaccines at sites around the country made it seem that Biden’s goal of 70% population inoculation (i.e. at least one vaccine shot per person) by July 4 was just as easily attainable and inevitable. But as shown in CDC’s own data below, vaccination rates began to plateau across all age groups beginning around May, in part due to the stubborn politicization of vaccines as a whole. This plateau suggested that vaccination rates would no longer act as a relevant signal for booking live events. (The U.S. did reach its 70% inoculation rate on August 2, a month behind schedule.)
Brooke: Of course, a lot has changed since our initial analysis in June — with the Delta, Epsilon and Lambda variants of COVID now dominating the news, driving national case counts up, shifting local government regulations around vaccinations and forcing many artists to reevaluate their upcoming tours.
With these changes in mind, we decided to do another data analysis with more recent Bandsintown data up to July 29, 2021, as well as pulling in more comprehensive COVID community transmission data from the CDC, which includes copious information about the community on the county-level, included not only COVID case rates, fatality rates, and vaccination rates by age, but also hospitalization and ventilation rates, COVID transmission rates and its forecasted trajectory, and social vulnerable factors as a percentage of the county-level population.We are still seeing mostly low significance across the board — but a few things do stand out and are worth monitoring as the Delta variant progresses.
Looking at the last 28 days vs. the same period last year, we are seeing a weak but statistically-significant increase in relative ticket click activity in counties where COVID cases are forecasted to decrease. Inversely, in counties where the trajectory of COVID cases is uncertain or forecasted to increase, there is a negative coefficient with ticket click growth.
This is not entirely surprising, and could indicate a correlation between local COVID data and consumer sentiment around attending events. In the box plot below, we see a median percentage of event growth (indicated by the black line in the middle of the boxes) that is higher for areas forecasted as “Likely Decreasing” in COVID transmission than those that are “Likely Increasing” or “Uncertain” — although it should be noted that the height of the red box plot indicates that there may be outliers that are bringing up this median.
At the county level, we do see a tiny positive statistically significant relationship between the growth in events playing out over time and the percentage of the local population that is fully vaccinated. Again, this is what we would have expected to see from the beginning — but correlation does not equal causation, and this relationship could be simply due to the fact that larger cities with more events also tend to be vaccinated at a higher rate.
BUT, on the county level, we’re also seeing a slight (almost flat, but still statistically significant) positive correlation between the percentage growth in events happening in the last 28 days and the number of COVID cases per 100,000 people over the last 7 days. While this may be an apples-to-oranges comparison given the different time periods measured, we could potentially infer here that COVID cases are rising in cities where more events are taking place — which, again, is not surprising and poses a significant health and safety concern for the general population.
We’ll continue to monitor these relationships over the next month as we collect new data from Bandsintown, especially considering the attention that has been focused on the recent Lollapalooza Festival (which happened from July 29 to August 1, and therefore is not included in our data).
Part II. The COVID data strategy that wasn’t
Cherie: After several weeks of data analysis, our next step was to gut-check our findings with sources in the live music sector. We wondered whether seasoned live music professionals were picking up on the same signals as we were, in terms of the small handful of significant correlations we found among event announcements, consumer sentiment and COVID data. We held several qualitative interviews with reps from some of the top talent agencies, promoters and safety management consultancies; most of our sources requested to speak on background or off-the-record, given the financial and political sensitivity of the topic.
What we found was a bit shocking: By doing our own, ad-hoc correlation analysis, we were already doing more work than some of the industry’s top agencies and promoters when it came to monitoring COVID data. Put another way: Up until the last few weeks, COVID data has simply not been a factor in the live event booking process.
Of course, most of the professionals we spoke with have been aware of COVID case and vaccination trends — but have yet to operationalize that “awareness” into a dynamic, data-driven system to help drive booking decisions. There are several reasons why this is the case, which can be broken down into two categories: Workflow and money.
Workflow: Long timelines and lack of safety standards
Cherie: It’s not like the live music business doesn’t use any data at all to inform their booking decisions. As music data consultant Julie Knibbe previously wrote for Water & Music, agents, promoters and venues rely on several sources of data — including historical ticket sales, regional streaming data and Google search trends — to mitigate risk and optimize investments around their shows.
Knibbe did mention health data as a new frontier for understanding who shows up to a given concert the day of — allowing venues and promoters to paint a more holistic, real-time picture of the fan along variables like when they arrived, how much they spent, where they spent most of their time and which sponsors they engaged with the most, in addition to their COVID vaccination status. But what about health data as an input in the booking process itself, before a show is even scheduled? Do live music professionals have systems in place for monitoring COVID cases, vaccination rates, local reopening or mask-mandate policies on a regular basis?
Brooke: The issue here is that agents are often booking shows several months, if not a year or more, in advance. This is especially the case for larger artists, or in situations like now where agents are competing for rescheduled tour dates six to 12 months out. In contrast, COVID case data and related policies are changing on a daily basis, which makes it logistically impossible for agents to rely too much on those data sources in their long-term booking decisions — otherwise, nothing would get booked. One source from a major promoter told us that around 50% of their upcoming shows in 2021 are rescheduled from early to mid-2020, and that it would have been futile to try to reschedule these shows based on future COVID trends that were difficult to predict.
Even if our agent and promoter sources could access real-time COVID vaccination and case data and a regularly updated list of local government regulations, it was unclear from our interviews how or when exactly they would use that data, given long booking timelines. To reiterate, even getting centralized access to concert announcements and ticket sales data across competing promoters and agencies is difficult in the first place. Bandsintown helped us get access to quite detailed data on this front — but the process was very ad hoc, took a few weeks to finalize and ultimately did not include the raw data that live music professionals would likely want access to.
In fact, the recent rise of the Delta variant might be the first time that many live music professionals are actually taking COVID data into account seriously, although on a short-term and reactive rather than long-term and proactive basis. And even then, despite several instances of COVID breakouts being linked to live events, the reactions and messaging from the live music sector have been inconsistent. Some artists like Garth Brooks have postponed their fall stadium shows, others like Japanese Breakfast have required masks and proof of vaccination or negative COVID tests, others like Limp Bizkit have canceled their events entirely and still others like the organizers of Lollapalooza have gone on without a blink.
Cherie: This points to another reason why it’s difficult to incorporate COVID data into booking decisions: Because COVID has become so political, local regulations and safety protocols are all over the place right now. Some cities like New York now mandate proof of at least one vaccine dose for admission to indoor dining and entertainment, while at least 20 states — including popular destinations for touring like Florida, Georgia, Tennessee and Texas — have outlawed vaccine mandates entirely, even for private businesses. This inconsistency makes it near impossible for promoters to implement standardized safety policies across all of the venues and tours, because in some cases those policies might be deemed illegal.
We see this tension playing out in real time with promoters like Live Nation. In early July 2021, CEO Michael Rapino went on the record saying that Live Nation’s U.S. business is “fully reopened,” and that 70% of the promoter’s business in the coming months will be in the U.S. and U.K. At the same time, during Live Nation’s latest call with investors, Rapino addressed the tension between what the company and its artists want, and what is possible with local governments. While the CEO promised to ensure that all of Live Nation’s employees eventually got vaccinated, and to accommodate “artists that are going to want fully vaccinated and tested shows,” he also admitted that the company’s biggest challenge has been “abiding by different local COVID laws — mask, no mask, now test, no test. I think that’s been our only real challenge from an HR and communication [standpoint].”
For now, Live Nation is taking the approach of deferring to artists and local governments to determine safety standards at shows. The promoter just announced this past weekend that they would allow artists to require attendees and staff to show proof of vaccination or a negative COVID test on their tours — barring any local regulations that would prohibit them from doing so.
We’ll dive into this in our concluding section of this report, but this points to a core ethical question at the heart of this analysis: Who is ultimately responsible for standardizing and implementing COVID safety protocols for live shows?
Money: “We just want work”
Brooke: The second major reason why live music professionals haven’t incorporated COVID data into their booking workflow is more financial and psychological. Simply put, many of these professionals are reluctant to slow down the momentum of financial support and economic activity that the sector has seen so far in 2021.
There are multiple lenses through which to view this issue. One is through macroeconomic data, especially through market activity around publicly-traded companies like Live Nation.
As COVID cases and fatalities dropped throughout early 2021, Live Nation’s stock price grew from a low of $66/share on January 29 to a high of $93/share on March 1. That same peak was reached again in late June, alongside the rise in consumer sentiment and booking activity that we discussed earlier. But the stock price then fell sharply again in recent weeks, corresponding with news of the rising prevalence of (and CDC guidance around) the COVID Delta variant. Such pressure from Wall Street is likely why the words “COVID,” “pandemic” and “safety” are nowhere to be found in the promoter’s Q2 2021 earnings release (although they do briefly address safety measures in their 10-Q); in this case, the COVID communication strategy seems to be no COVID at all.
A second lens is through government grants to both consumers and the live music industry. On the consumer side, COVID stimulus checks led to a sharp increase in disposable personal income in the first quarter of 2021, driving increases in consumer sentiment and spending that directly benefited the live sector. One agent source we spoke to said that the live sector was expecting that “the dimmer switch” would be slowly raised on booking activity throughout 2021, when in reality it was more like “flood lights” in late March.
On the industry side, as of late July 2021, the Shuttered Venue Operators Grant program has awarded $7.6 billion in grants to around 10,000 live entertainment venues, museums, theatrical producers and talent representatives as part of the Save Our Stages Act. This has partially contributed to the recent growth in employment in the leisure and hospitality industries, with 343,000 jobs added in the U.S. in the month of June, according to the Bureau of Labor Statistics. 74,000 of those jobs came from the arts, entertainment and recreation sectors, and we’ve seen this play out in the return of furloughed employees to live music companies across the board.
Cherie: Last but not least, there’s the lens of artist (and agent) incomes from touring. Comfort levels with touring during the pandemic will obviously vary from artist to artist. But one recurring theme from our interviews is that many agents and their clients just “want work,” and will do whatever it takes to get that work, now that it’s actually legal.
I do think it’s fair for performing artists and the live sector to want to go back to working regularly, given that they’ve been unemployed for the good part of a year-and-a-half. But financially, this also paints a picture of desperation to me — desperation not just to be able to pay one’s bills, but also to keep up the PR momentum of getting “back to normal,” however true that sense of normality may be.
The worst case scenario is that this desperation clouds people’s statistical understanding of the COVID data that will impact the future of their industry. There’s definitely a tension between the peppiness of the government-backed “return” of live music in major metro areas — as exhibited by events like Lollapalooza and the Clive Davis-curated We Love NYC: The Homecoming Concert — and the reality of the emerging COVID variants, which some models predict will not peak until this October.
Part III. So what?
Brooke: So, after crunching a bunch of data and talking with several live music professionals, we found what could be interpreted as simply more bleak news to add to the already grim picture for the future of live music — namely, that a COVID data strategy for live music currently doesn’t exist at a time when the sector arguably needs it most.
But we didn’t want this to be the only conclusion. For now, we’re coming away from our analysis with more questions than answers, but we think some of these questions could generate potential actions that live music companies can take to incorporate health data more concretely into their business decisions, in a way that ensures maximum safety for everyone involved.
Cherie: The first major question to emerge from our analysis is what a more solid COVID data strategy for the live sector could look like, taking all of the challenges we discussed earlier around booking workflows and local regulations into account.
Arguably, COVID data should be an integral part of any brick-and-mortar business’ data strategy in 2021, and especially for the live sector given the larger risks of people getting COVID at in-person shows. This is already par for the course at the government level, where predictive modeling and pattern recognition around real-time data are table stakes to ensure the health and safety of local, national and global populations.
How could this mindset apply realistically to live music companies? Even if it’s impossible to incorporate future or hypothetical COVID data into tour routing decisions early on, our initial analysis of Bandsintown data presents a clear case for a data strategy that can measure changes in consumer sentiment (e.g. citizen acceptance of local government safety mandates and variations in vaccination rates) and state/city/company regulations in the weeks leading up to a tour launch or announcement.
This approach could be especially helpful for artists who are booking shorter tours from scratch this fall 2021 through winter 2022, as the impact of health and sentiment data trickles down to all areas of the touring business, across pricing, logistics, marketing, PR, business development and brand sponsorships. For instance, if consumers in a particular market have higher vaccination rates and/or the local government cares more about safety at shows, how would that impact the budget for a show in that market, including ordering rapid COVID tests, PPE and other requirements? Would the increased costs be passed on to the consumer in the ticket price? If consumer sentiment is low, what kinds of marketing and outreach initiatives might the promoter, venue or artist have to invest in to make sure fans feel safe and encouraged and industry stakeholders can still break even?
Brooke: This brings us to our second takeaway that is more of a question, namely how to address the lack of standards in who should be taking responsibility to ensure the safety of live events.
As we addressed earlier, the implementation, communication and even sheer legality of safety protocols for live events remains inconsistent across the U.S. — partially due to the federal government’s approach of holding off from official COVID vaccine mandates, instead deferring to city/state officials to lead local regulations. This situation has essentially forced promoters to do the same, deferring to each individual artist, venue or market to determine what safety protocols make the most sense for them based on local mandates (or lack thereof) as well as venues’ own budgetary constraints.
Even still, one source from a safety management company told us that they hoped there would be a “stronger coalition of senior leadership” in the live sector in terms of setting, implementing and communicating base-level safety protocols, especially from larger promoters like Live Nation and AEG. Such a coalition would be especially helpful from a budgeting and grant perspective, in that venues would have a clearer idea of the resources they would then need to cover staff and safety equipment and keep their businesses operating. (Ironically, all the major promoters and agencies did form a task force in March 2020 to address the mass cancellation of tours at the time, but it’s unclear whether that task force is still in effect in the current context of shows slowly returning.)
Cherie: To be fair, as we’ve addressed throughout this report and are seeing through recent announcements from the likes of Live Nation, we seem to be at a tipping point where live music professionals are no longer merely “aware” of COVID as a risk factor, but are now monitoring it more closely and in real time, because it is a direct threat to the long-awaited return of the live music business as a whole. Even if the standardization of safety practices across the live sector might be legally or politically impossible, there’s still a clear case for individual live event companies to adopt a more robust internal COVID data strategy to drive their decisions across logistics, marketing, booking and communications.
Brooke: We’ll continue to follow live event trends in partnership with Bandsintown, monitoring correlations among ticket clicks, event announcements/happenings and local COVID data. This story is developing rapidly and we anticipate a lot of these trends to change in the next few weeks alone, in line with rising case counts, a potential decline in consumer sentiment and potential outbreaks from large events like Lollapalooza. City-level vaccine mandates might also lead to more significant changes in vaccination data in the near future.
All in all, we’re starting to see some signs that COVID data correlates significantly with both consumer sentiment and consumer risk when it comes to live shows, and we hope our analysis encourages live music companies to take this data more seriously in their day-to-day operations — because, again, so much is at stake financially and ethically. We look forward to issuing more updates on this research in the coming months, and encourage you to reach out to us with any feedback or suggestions for additional avenues to explore to get the best understanding of the real, if precarious, return of live shows.