I spoke to Eric Klein, Director of Analytics & Technology at C3 Presents, the company that puts on Lollapalooza and Austin City Limits (in addition to music festivals, they also do artist management, experiential marketing, concert promotions, and food and wine festivals).

Eric Klein After The Show Interview

After The Show: Can you give a breakdown of how you spend any given work week? For example, how much do you focus on trying to get ahead of the curve with predictive analysis vs. focusing on the present or recent past?

Eric: I head up our Marketing Analytics and Technology group which is a team of four full-time people and then we contract out for additional help now and then.  On the analytics side of things, we divide our time between hindsight and insight – what happened and what does it mean.  The various festival business units, such as sponsorships and marketing, have on-going “what happened” reporting needs as do our executives and partners.  So a fair amount of time is spent gathering data, analyzing it, and then summarizing it into a digestible format depending on the audience (e.g., showing a sponsor what they got in return for their investment with us).

However, the more fun part is like you said – getting ahead of the curve – or for me, the “what does it mean.”  We’re all avid live music fans here, and we all want to do everything we can to make sure fans get to both see their favorite artists and find new favorite artists – or even just find their new favorite festival experience.  So lately we’ve been spending a lot of time developing strategy and pouring through analytics determining how to better guide people through our website based on time of year and what they’re most interested in.  I come from a heavy background in predictive analytics from other industries, but I haven’t yet found a good fit for it here in the annual music event space.  Whether or not someone will buy a ticket or what bands they like isn’t easy to predict.

What are 3 of the most important skills that a music data analyst needs to succeed?

There’s a big difference between “music” and “music festival.”  So the first critical skill is knowing how to identify those differences when doing marketing analysis.  For us, we have to recognize that we run annual events – which basically means selling one thing once per year.  That is radically different from a company that is selling thousands of items every day all year.  Another top skill would be the ability to get and manage the data.  Larger organizations have separate data management roles, but for us, we have to be able to get the data before we can start analyzing it.

Another top skill is visualization and summarization…You have to know how to put the results in a digestible story – which is maybe a one line sentence, or a simple line graph – all the way to something much more sophisticated that easily shows the patterns that are most actionable…  It also includes expert use of visualization tools – anything from Excel charts to Tableau interactive dashboards to doing very detailed analysis using R.  The last I would say is knowing what is actionable.  We try not to spend time on requests that are purely curiosity or aren’t actionable.  Curiosity doesn’t sell tickets!

Besides Google Analytics, Analytics Canvas, and Tableau, what analytics tools do you like to use?

I think it’s useful if not critical to know R these days.  My background includes 20+ years of SAS, but it is very expensive and tends to just be used by the Fortune 500.  Since R is open source and free, many more companies of all sizes are using it and requiring it.  It has a pretty large learning curve (just like the rest), but there are lots of good books on it and even online courses at places like Udemy.  The thing that helps you learn them the fastest is having a good question that needs answering and the available data to answer it – and a deadline – haha.   I would also never leave out Excel.  It may not have the power that other tools have, but being able to quickly crunch some numbers and make a meaningful chart can be pretty critical to day-to-day work.

Given that quantitative data, algorithms, and web crawlers are important aspects of data analysis, what opportunities are there for analysts who are more interested in the qualitative, human element of music analytics?

I love that question!  I don’t have a good answer, but I could go on for days about the human element – and some of it even still has to do with math!  I believe a lot of the pleasure in music has to do with tension and relief…All good stories have an antagonist or some type of tension.  Even the sun and the moon’s rotation model tension and relief.  In music, we see this both in the lyrics and even in the music.  In the lyrics, the verses tend to drive the tension but then the song comes back home to the chorus which is our comfort zone or relief.  And in the actual music, even the chord progressions follow a similar model where they begin and end with the “root” chord that is our home base but take us off into all sorts of places in between causing the tension…

I love thinking about how music is just a lot of waveforms of varying frequencies and amplitudes against our ear drums – and at live events even the pressure waveforms of the subs beating against our chests.  And in our minds we were given this amazing ability to enjoy it all and to connect it all to specific memories of where we were or who we were with forever…I’ve always been fascinated at how companies like Pandora or Echo Nest map out all of the details of a song to try and come up with recommendation engines, but I’ve often wondered if they also could map out the human elements of the types of tension and types of relief…But having said all of that, I also think much of whether or not people like a song has to do with whether or not their friends like it and told them about it.  I think the true music influencers are a much smaller group of people than we’d like to admit.

What’s your view on individual privacy concerns when it comes to demographic and behavioral data? How can data analysts ethically respect consumers’ privacy while also doing their job of collecting and analyzing information?

In a time where privacy seems to be getting harder and harder to find online, I still think it’s critical.  Facebook has been recently forcing its developer partners to define exactly how demographic data will be used before approving apps – so for example, if you want to ask for someone’s birthday if they connect through Facebook, you’ll have to convince Facebook that the user’s experience will benefit from knowing that information – as opposed to just sucking all of the data they can from you to use however they want.  I think its a good move and hope other platforms follow. I have no need to know anything about an individual because that doesn’t scale.  If I knew individuals who have young kids, it would be too hard to reach each one and tell them about a family aspect of one of our events.

On the other hand, knowing something about an entire group of people can be valuable.  So for example, if I have a page on a website just about family aspects of a festival (e.g., that kids under 10 get in free), then it might be worthwhile to send everyone on our email list who viewed that page something more specific about other family oriented features of the event as opposed to sending it to everyone on the list.  It is always a balance of deciding what will creep someone out giving them the impression someone is watching them vs. offering them a valuable service.

Another way I look at it is that demographic information is currency.  If you are going to be willing to give me some of it, I had better be able to prove to you that I can keep it secure while offering you something more valuable in return.

It seems that the biggest application of your analysis is in marketing and social engagement. Have you explored curatorial applications…for example, how information gathered from streaming services or from artists’ social media profiles influences festival line-ups or helps consumers discover new music?

We consider things like that now and then, but we just haven’t seen anything that moves the needle for us.  In terms of line-ups, our talent buyers are award winning and some of the best in the world at staying tuned to who people will want to see or even who will be the next big things.  For me personally, and I assume for many fans, our festival talent buyers are today’s music taste makers much like radio DJs used to be decades ago.  I love the headliners and look forward to their performances, but the real value to me as a music fan is seeing as many of the mid-tier bands as I can (shh, don’t tell anyone that I have time during the festival to stop working and go see bands!) because I just know our talent buyers have done their homework and always nail it.

As for helping people to discover new music, I’m always eagerly looking for and open to any ideas.  We can never get too good at this.  We’ve looked into various techniques and services.  We currently try to give recommendations based on various “you might also like” algorithms, but I sort of think it goes back to that human element you asked about earlier.  If you or your readers have any ideas on how we can do this better, I’m all ears!

Thanks for sharing, Eric! To learn more about Eric Klein check out his LinkedIn.