00:00:05.73 [Announcer]: Welcome to the Analytics Power Hour. Analytics topics covered conversationally and sometimes with explicit language.

00:00:14.44 [Tim Wilson]: Hi, everyone. Welcome to the Analytics Power Hour. This is episode number 282. And technically what we’re going to talk about on the show is pop culture, which can feel like a pretty kind of vibey and nebulous and subjective topic. And I’m Tim Wilson, and as Michael Helbling inadvertently reminds me on a regular basis, pop culture is something where I seem to be pretty locked into like a pretty narrow window of the 1980s. I did, after all, weave in the mid-80s sitcom 227 into the introduction I did for episode 227 of this show. In my defense, that was one of Regina King’s earliest appearances, and she is still going strong. But I digress. I’m going to give a shot at facilitating a discussion centered around pop culture ago, even though maybe not my forte, because I think this show is going to uncover some useful perspectives about a range of kind of things that we want to measure and analyze that are inherently kind of vibey and nebulous. consumer sentiment, employee engagement, social influence. Anyone? Anyone? Luckily, I’m joined by a couple of much more hip cats, as we said back in the day, than I am for this discussion. Julie Hoyer, would you agree that the data shows that I am the least likely co-host to get a Chapel Roan reference?

00:01:38.31 [Julie Hoyer]: Yeah, that’s probably a good conclusion, honestly.

00:01:42.08 [Val Kroll]: At least he said her name right.

00:01:45.88 [Tim Wilson]: I do have a 20-year-old daughter.

00:01:47.27 [Julie Hoyer]: See, that helps.

00:01:49.23 [Tim Wilson]: Yeah. I get a lot of dads on those sorts of things. And Val Crowell, we heard you. What about you? Have you ever considered checking in with me to learn about what’s trending in the zeitgeist?

00:02:04.21 [Val Kroll]: You are quite plugged in, Tim. Moere than you think.

00:02:09.22 [Tim Wilson]: Perfect. Not. The funny thing is that I’ve been a regular reader of our guest’s sub-stack, which is called Can’t Get Much Higher. I’ve read that for several years now. That newsletter is a weekly data-driven analysis about the musical trends of yesterday and today. So I like to think it provides me with some analyses that help me kind of fake my pop culture knowledge. The author, Chris Dalareva, has a day job as a senior product manager of data and personalization at Audio Mac, which is a creator-friendly music streaming service. Chris is also a musician and the author of an upcoming book called Uncharted Territory, what numbers tell us about the biggest hit songs and ourselves. For the book, Chris went back and created a pretty fascinating and extensive data set by listening to every Billboard number one song from 1958 until earlier this year, which wound up being just over 1100 songs. He then conducted a bunch of analyses with that data and the results are what kind of fed into the book. And today he is our guest. Welcome to the show, Chris.

00:03:15.36 [Chris Dalla Riva]: Thank you for having me. I’m looking forward to chatting with the other hip cats here.

00:03:20.13 [Tim Wilson]: The fact that I am the oldest one on this mic by a lot. So I think hip cats might even be a little dated for me. So maybe a good place to start would be just by getting a little bit deeper on kind of the what, the why, the like, what were you thinking kind of in the how of the book. Like what prompted you to even tackle that project in the first place? And then how did you kind of land on Billboard number one songs as a place to do that?

00:03:56.41 [Chris Dalla Riva]: Yeah, good question. I don’t think if I had some grand scheme, this listening journey would have ever started or a book would have ever been written. In fact, I think to write a book, you have to be a little bit insane because it starts to feel like a fool’s errand at multiple parts during the process. But the reason this started was I was working in consulting when I got out of college, economic consulting. I was in the data world already, but I wasn’t particularly enamored with my job. I still played in bands. I had released music, but I was just looking for another musical outlet. And I just came up with this idea that I was going to listen to every number one hit. And again, I probably thought this would peter out after 50 songs or something. But I was like, oh, you know, the Billboard Hot 100 started in 1958. I’ll do one song a day. I’d play along with it on my guitar and it would just be like a nice wind down activity for my day. Eventually, I was telling a friend about this probably when I was 20 songs in and he was like, oh, I want to do that too. So every day I would send him the song, we would chat about it. And that was really it. At a certain point, we started ranking the songs, rating the songs out of 10. And it would just be like a fun, again, end of the day activity, but because I was already enmeshed in the data world through economic consulting, I started tracking our ratings in a spreadsheet. This spreadsheet grew to, yeah, naturally to include an absurd amount of information. I noticed some trends and I just started writing about it and I started sending it around to people and they were like, oh, this is pretty good. You should keep at this. I did. Eventually I realized I sort of have the makings of a book. I’d never published anything before. So ultimately I started trying to get articles published. That’s why I started this newsletter. And I eventually parlayed that into eventually getting this book published. It was a weird circuitous journey. It’s all sort of wrapped together, but it just started as a daily way to wind down, I guess.

00:06:06.26 [Tim Wilson]: And then you couldn’t wait because you didn’t have time to do that because you had to work on content, content production.

00:06:11.79 [Chris Dalla Riva]: Yeah, yeah.

00:06:13.44 [Tim Wilson]: Definitely a personal issue. Like it sounds like it started as kind of a two column spreadsheet or say a three, it was year, song, however long it was. And it grew. I’m just conceptually thinking that’s a bunch of other columns. Every time you added another field, you didn’t have to go back and re-listen to the songs. That’s this fundamental question I’ve been chomping at the bit to ask you is how the data model, I guess, if you think in classic business speak, it’s going to be let’s design our schema. And this kind of grew more organically, presumably had good intuition. But how did the width of the data set grow and how did you manage backpopulating, I guess?

00:07:02.80 [Chris Dalla Riva]: Yeah, that’s a good question because this process was super organic. And as I was going along and I would think to write about something else, I would have to go back and repopulate it. Some of those things did require re-listening to these songs. For example, in chapter five of the book, I write about the evolution of song structure. So I had to tag every single song. the structure of every single song. So the most common song structure these days in the pop world is just a verse chorus structure. So when I realized I wanted to write about that, of course, as I kept going forward, I would fill out the information, but then I was like, all right, I got to go back and fill the other stuff. And for that exercise, I had to re-listen to everything. But I would do it at the time I was living with my parents and I would take the bus every morning into New York City for my job and I would just sit on the bus and listen to these songs and tag them. But other stuff didn’t require, most stuff required backfilling. Some of it was easier than others or didn’t require listening to all the songs. Another good example is I track a lot of demographic information about who the artist is, where they’re from, gender, race, and I do the same for songwriters and producers. Looking up that information doesn’t require actually listening to the songs. There was a ton of manual work, which I think was actually very, what’s the word? Manually filling all this out actually got me so deeply in touch with this data and going back through it a bunch of times. But if I were to set out from the beginning with a hundred different columns, again, this book probably would have never happened. It was just the organic nature and the insanity of being like, all right, I’ll backfill 800 songs. What else do I have going on?

00:09:03.03 [Val Kroll]: What was one of the first themes that emerged that you were like, oh, this is a little bit of a pattern that’s like novel or I hadn’t heard talked about before during this time period or from this music. I’m curious.

00:09:20.55 [Chris Dalla Riva]: Yeah, the reason I initially was compelled to write something was I noticed there were lots of songs with very grisly, lyrical topics. You know, lots of songs about teenagers getting in car wrecks and dying. This is like the late 50s, early 60s. And I’m thinking to myself, like, that’s weird. You know, when I turn on the radio, I feel like If you were starting a focus group for how to write a pop song, you would be like, avoid all songs about teen car wrecks. And there’s, you know, 10 or 15 songs in a very short period that were all very popular. So I started looking into this and this was a legitimate, a legitimate trend so much so that it has a name. They’re called teenage tragedy songs. And again, usually it involves two high schoolers in love and one tragically dies. Often it involves a car. And I wanted to sort of pick that apart a little because it was so strange. Yeah, and that was the first weird. That sort of set me off on this journey. The rest of the book isn’t so grim, but that was the beginning of, I was like, oh, there’s, you know, maybe these songs actually are speaking to something a bit deeper about what’s going on in the world around when they come out.

00:10:32.59 [Julie Hoyer]: So the way the top hit of the Billboard Hot 100 list is, is it like a weekly thing? And did you get runs of a song being number one multiple weeks in a row? So were you like re-listening to a Beyonce song like five times in a row for certain years? You know what I mean? Like I’m just curious, like how did that feed your data set exactly?

00:10:55.47 [Chris Dalla Riva]: Yeah, I should have specified that the Billboard Hot 100 is Billboard’s pop chart. So the goal of the Hot 100 is the number one song. Billboard is trying to say this is the most popular song in the United States. in a given week. So it is weekly data. Moest, it ebbs and flows throughout the years, but most number one songs I would say are number one for like at least three weeks. So usually you get, I say on average like 20 number one hits per year. I wouldn’t, if a song was number one, like Hey Jude is number one for like nine weeks. I just listened to it one day, you know? And then the next day I go on to the next song. You weren’t a purist that way. No, no, I didn’t have to live the past in real time. One day was enough.

00:11:36.71 [Tim Wilson]: That could be one row. I presume what you had when it was number one. What was the unit of analysis, the core of the data was a song, I guess, recently?

00:11:47.95 [Chris Dalla Riva]: Yes. The core was the song. I tracked the amount of weeks each song was at number one. At various times when I was conducting analyses, I’d be like, oh, should I wait it by the number of weeks that the song is at number one? Or should I just be like, all right, if it got to the number one, that’s it. We’re just going to wait everything equally. So you could approach it differently, but the unit of the data set is at the song level.

00:12:11.82 [Tim Wilson]: So you started to dip into this and you had one of your newsletters kind of talked about some of this. You were picking You were picking the number one song, but the nature of how that is calculated and what goes into it. We think of it as being this is the number one song because it’s labeled the number one song. Everyone understands it. It feels like this very objective, concrete, hard measure But in and of itself, the number one song is kind of a proxy you said it. It’s like it is an attempt to say, what is the most popular pop song this week? But as soon as it gets converted to gathering that label number one, it’s not the number two. And it could have been a coin flip one week. And here we are 40 years, 30, 20 years later, and you’re listening to the one song and not the other one. But you had some cases where you even sort of sussed out that maybe that underlying, right? Was it the 70s or the 80s where it sort of felt like maybe somebody was putting a finger on the scale?

00:13:20.78 [Chris Dalla Riva]: Yeah. I mean, you have to realize like ultimately Billboard is a, it’s a trade publication. It’s like, if you go into, I work at the music industry, if you go into a label, I mean, you still see copies of Billboard magazine laying around in the same way that I don’t know, if you were running a trucking company, I’m sure there’s some trucking trade publication that people read. It’s just that billboards also write, yeah, billboards also writing about celebrities, right? So it does have that pop culture crossover. But I mean, I don’t know, say you’re living in 1970 and you have been tasked with defining what is the most popular song in the United States. I don’t know, I mean, how are you gonna define it? Okay, we should look at radio. We should look at what people are buying in record stores. how are you gonna get that information? At the time, Billboard would just call record stores and be like, hey, what’s selling? You know, that’s a sample of record stores does seem like a great way to do it, but of course it’s imprecise. I mean, maybe the person at the record store gives you bad information accidentally. Maybe they give you bad information on purpose for whatever reason. You know, there are many ways that it could become skewed and in the 70s, There are various stories about certain songs being put up at number one just for… You know, back there, there’s just deals that people were cutting like, okay, you give me the number one song this week. And what I write about in chapter five of the book is that you can actually see this in the data. There’s some anomalous behavior in how long songs are staying at number one, when they lose the top spot, how far are they falling down the charts? So we did a little bit of a fraud investigation there. But it’s interesting as time has gone on, of course, we listen differently these days. And again, it becomes even harder to define what is the most popular song. It’s like, okay, we’re going to look at streaming. We’re going to look at what people are buying on vinyl, what people are downloading digitally. It’s a tough task. It’s more accurate now because since in 1991, Billboard switched over to a system called SoundScan. which basically would track when a barcode was scanned at the record store. Like, okay, we’re going to log the actual purchase and we’re going to get actual playlist from the radio. And today it’s even more automated. It’s like every streaming service is sending streaming data directly to billboard. It doesn’t mean there’s no shenanigans going on, but it’s a very hard task. And, you know, I think just the exercise of thinking about it makes you realize how tricky it is to define things. What does it mean for something to be popular and how do you weight all these things differently? It’s hard.

00:16:06.82 [Tim Wilson]: But it is, you pass over. I mean, even like, Nielsen ratings kind of went through the, was better when it was like, we have a sample representative sample of set top boxes and people are watching TV this way. And then they have like really struggled to say, you know, what is the most popular show? Or I even think with like social media with trying to figure out who is like, who is an influencer? And there’s the, you know, the Kelsey brothers or Taylor Swift. That’s an easy, they’re like way up there. But when you try to get down to some sort of quantifiable measure, you’re stuck with trying to pick some sort of proxy. And if that proxy becomes widely adopted than it’s subject to be, you know, manipulated. Oh, it’s going to be the number of followers they have on these accounts. Well, then they can, people can go and buy followers. So, cool. That was just existential. Well, yeah. The Kelsey Brothers and Taylor Swift. I am like tapped in. Look at that.

00:17:13.87 [Chris Dalla Riva]: But they are ultimately proxies. You’re right. All this stuff is we’re proxying, we’re trying to get at something. Is the number one record in a given week dramatically more popular than the number two record? I don’t know, probably not. Is it unpopular? No, I also wouldn’t say that. Like it’s getting at popularity. I don’t know if you’ve reached the platonic form of measuring this, but it’s a hard problem.

00:17:41.03 [Tim Wilson]: Does it matter if one and two and three and four and five in general have some kind of similar characteristics over time, week to week, based on what you’re trying to look at? Presumably, it would come out in the wash. You’re working with some level of uncertainty, but if you’re looking at the size of the backing band or the number of writers on the song, presumably, as you’ve been trying to look at, what are these trends happening over time? You could have gone and listened to every number three song and probably come up with the exact same or very, very similar results, right?

00:18:27.53 [Chris Dalla Riva]: Yeah. I went through this at the beginning of the book where someone was like, when I was first starting to write this, someone was like, well, our number one song is really a representative sample of what’s going on in popular music. Shouldn’t you be looking at the entire top 10 or the entire top 100? That would be impossible given that I was listening to the songs and, you know, I have to… The bus ride wasn’t that long. No, well, some days… Damn tunnel. Yeah, literally. What I came upon is exactly what you were saying, Tim, is that for something to become the most popular song in the United States for even one week, it is typically representative of a larger trend that is going on. And I would see that again and again. Occasionally, there is just one random song that shoots to the top of the charts. It’s completely unconnected from everything else. But most of the time, There are musical movements and that’s how these songs become popular. So they are, if I were to have chosen the number three song, I mean, it doesn’t have as nice of a ring to it, but I think the outcome for all of the analyses in the book and my newsletter would largely look the same.

00:19:41.04 [Julie Hoyer]: I wonder if your fraud one would have looked the same or not though, because I doubt people are like, you know, if they’re trying to put their finger on the scale, are they really going to be happy with number three? They’re like, no, give me number one. You know what I mean? Like I would do wonder if some of those like drastic anomalies would exist if you had done, you know, top three instead of one.

00:20:00.97 [Chris Dalla Riva]: I should have done that as a comparison point, but I didn’t.

00:20:05.31 [Val Kroll]: a list of things to do.

00:20:08.57 [Tim Wilson]: I bet they would care more about whether they’re 10 or 11 than they would care about whether they’re 4 or they’re 5, right? I mean, I could see just what people say billboard top 10, right?

00:20:20.93 [Chris Dalla Riva]: Yeah, their top 40 has been the thing for decades where it’s like a radio format. Um, so if you were at number 40, there were lots of shows over the years that have counted down the top 40 songs every week. So if you were 41, that’s really bad for you. So that’s, there’s probably weird, really weird behavior around there too. Um, especially back when radio was very important to how music work.

00:20:44.19 [Val Kroll]: top nine at nine. I’m just like revisiting all these like memories of like the radio, like counting things down. But I remember being like on Friday nights, like I’m from Chicago. So B96, I would always record it. So I’m like, this is it. This is the top nine. Like they know they figured it out. Like we got to record it. This is, this is the playlist that everyone’s going to be listening to while they’re playing hopscotch on you. I don’t even know how old it was, but I was pretty obsessed with that. I do recall.

00:21:10.16 [Tim Wilson]: Something else that you said was that you started listening and the generating of the data was just you and your friend with a rating and a ranking. But then you already alluded to, well, there’s other stuff that you realized was other metadata about the song, but you went to other sources that wasn’t from listening to it and creating the data, which Did you remember when the first time you’re like, oh, I could add 12 columns to this data set by pulling data from somewhere? The whole supplementing, that’s intriguing as well, that your data set grew from other data sets.

00:21:57.86 [Chris Dalla Riva]: Yeah. There are a couple sources that I started tapping pretty early in the process. Of course, the big thing is like, who’s involved in making this song? Beyond the performer, the top things are songwriters and producers. Songwriters have always fascinated me. So there are some songwriter databases run by BMI and ASCAP, which are two of the biggest music publishers in the United States. And they track you know, they get songwriters paid. So I would go there to get songwriter information. Spotify also generates… a ton of information about the audio files on their platform, things like BPM, but they also gauge loudness and energy and what they call danceability and acousticness. They’ve sadly shut down a lot of these APIs because I think people were probably using them to train LLMs. Luckily, I had gotten all the information before that.

00:22:56.32 [Val Kroll]: You got some time.

00:22:59.17 [Tim Wilson]: Were there any APIs that you were like, this is like, they like really, or this data was very, very so easily accessible that was like really pleasant. And on the flip side, was there somewhere like that, you know, they have the data, but they’ve done, it’s an absolute nightmare to pull it in any sort of scale.

00:23:17.70 [Chris Dalla Riva]: Spotify, I mean, still the Spotify API is really easy to use. They’ve just changed what you can actually get access to. So that was always, Always a pleasure, Spotify. Thank you. I would have playlists of these songs on Spotify, and it’s very easy to just take a playlist of songs and ask the API to spit back certain metadata with a Python script. So that was great. I mean, lyrics are really easy to get around the web. Those were always a pleasure to grab. And I’ve done a lot of stuff over the years on my newsletter and in the book using data from Wikipedia. And Wikipedia, there’s an API and then people have built nice little wrappers for it and it works pretty well. So I never There are music data sources that you have to pay for. I never really had to rely on them for the stuff that I was doing because I feel like a lot of it was sort of wacky and based on historical data where no one’s monetizing label or songwriter credits from hits in the 1960s. If I want information on like streaming trends, like streaming trends right now, that’s stuff you really have to pay for. There’s a couple sources out there, but for the book, I didn’t really rely on stuff like that.

00:24:40.67 [Julie Hoyer]: How did you decide when you were done making this data set or is it like your baby and you’re like, I can’t stop. I’m just going to keep adding it.

00:24:49.58 [Chris Dalla Riva]: Well, I sort of joke about that at the end of the book. I’m like, well, I had to end the book, but it’s not like music doesn’t end. The day this comes out, it’s already gonna be dated. If there’s gonna be other number one songs. The dataset grew until I finished the last chapter and I was like, all right, the book is done. I’m saying it grew lengthwise in terms of songs, but also widthwise. If I was writing the last section, the last chapter and I was like, There would be this really cool analysis that I could run and I didn’t have the data. You have to be a little nuts to, I think, do something like this and not feel angry that you’re going to have to go back and fill in a ton of information. Yeah, it grew as I needed it to grow and not really much beyond that.

00:25:38.32 [Julie Hoyer]: Did you have a lot of it automated by the end? Like you would listen, you’d have maybe 20 columns you fill out. I don’t know, maybe that’s like really small and you’re like, I wish or you’re like, no, that’s crazy. 20 is too many. But did you have a lot of it automated and it wasn’t pretty easy to just keep adding to?

00:25:53.94 [Chris Dalla Riva]: Actually, very little of it was automated, but the reason it never really became automated was because I was only listening to one song a day. So I was like, you’re just listening and I’m typing information as I’m listening. I really usually only think about automating things if I’m grabbing tons and tons of data at once, but it was very tractable for me to be like, all right, today I’m going to sit down and listen to hot stuff by Donna Summer. By the time you listen through twice, you’ve probably filled out all the information you’re looking for.

00:26:29.23 [Tim Wilson]: That was actually, so was that kind of the typical? Was it sort of two listens or was it like how many times did you have to listen to feel good about?

00:26:38.02 [Chris Dalla Riva]: I would tell myself I would always listen twice, whether some of the songs Even if I wasn’t alive at the time, I mean, you’re intimately familiar with, but still I would try to listen twice with fresh ears to see if anything struck me. I mean, at this point, I’ve heard many of these songs too many times, but some of them twice is all you need and hopefully never hear again.

00:27:03.47 [Val Kroll]: Okay, that’s some of the stuff I’m super interested in. I do have a data-related question, but is there any song that you heard that you wasn’t one of your favorites or not even on your radar, but now it’s a go-to playlist item? Or in verse, you’re like, if I have to listen to the song one more time.

00:27:20.28 [Chris Dalla Riva]: There’s another nice thing I think about using number one hits as a sample is that if you were to look at some list of music of the 1960s, it would probably be packed with things that are now considered classics, but in every decade, in every era, There is good stuff released and there is horrible stuff released. The horrible stuff just gets, you know, swept away and we never listen to it again. But when you listen to just number one hits, again, it’s just what was popular in one single week. you get the good, the bad, and the ugly. And I certainly saw all that through my journey. I mean, there’s one, there’s a song in their early 70s called Want Ads by a group called Honey Cone that’s completely forgotten by however you measure it. I mean, this song has no cultural footprint and it’s absolutely awesome. I always am telling people about it. And there’s also stuff that has no cultural footprint anymore that is horrible. I mean, there are tons of novelty songs from over the years. There’s a song called Disco Duck from the 70s, which was like a joke about disco.

00:28:26.83 [Val Kroll]: It’s really been like… I’m familiar.

00:28:28.92 [Tim Wilson]: Nope, that hit me when I was at a certain age.

00:28:32.00 [Val Kroll]: Yeah, it was… It would be like my Macarena.

00:28:36.38 [Chris Dalla Riva]: The Macarena is another great example of, I mean, dance crazes have also, people have been dancing for thousands of years. So you can imagine that every generation has their handful of dance songs. Some of them seem silly now. I mean, I’m not gonna throw on the Macarena, you know, on Saturday afternoon, but it’s representative of something that was going on, something larger in the cultural atmosphere at that time.

00:29:05.51 [Val Kroll]: That’s funny. So I’m curious about, and you wrote about this in one of your posts about the one hit albums or one hit wonder albums. I’m not going to say that correctly, but in the context of this exercise, I’m curious if there was something that was a little bit more qualitative that you were rating that as you went through it, it made you think Oh, I need to reassess my definition or even when we were talking earlier about like teenage tragedies that I was just thinking, even though like on the face, like that feels, you know, like yes or no. But, you know, when the girl meets the boy from the wrong side of the tracks, like, you know, maybe as a fit in it, does it not? Like, I’m curious about how some of that stuff was scored or how you even documented for yourself what some of those definitions were to be consistent across all your listens.

00:29:54.53 [Chris Dalla Riva]: Tim mentioned this earlier is when we think about popular culture, there’s a fuzziness to it. I mean, we’re not talking about chemistry class or physical compounds. As much as I would want to be able to categorize certain things perfectly, you have to understand the fact that some of this stuff is always going to be fuzzy no matter what. Some of it won’t. I track how long the song is. I don’t think there’s gonna be ton of a debate about You know how many seconds Benny and the Jets by Elton John goes on for but I That’s something I would grapple with constantly before I was this is I think a perfect example and I write about this in the book is This topic is a little bit like inside baseball ish, but the structure of popular songs is When we, most people think of a popular song these days, I think of your typical verse chorus song. I mean, most stuff you hear on the radio now to talk about Chapel Roan is gonna be, you have your verses and you have your big repeated chorus that goes on and on throughout the song. And I wanted to see if that was the most common way popular songs were structured over the decades. The short answer was it was not. There were different song structures that were popular before that. the most common structure at the time is sometimes called like the great American song structure or A A B A in something, use something like Over the Rainbow by Judy Garland is of this structure. There’s not really a chorus in that song. Every verse opens with a refrain and then there’s a B section, which is like the someday I’ll wish upon a star. I don’t want to get too much into the weeds there, but the point is that even I wish there were just like five song structures that people perfectly adhered to, and I could just be like, all right, this is song structure three on this song. But that’s not how artists work. There would be songs like, this is very, very close to this one structure. It’s a little bit different, but I’m still just going to call it that because that’s what it’s close to. You do have to wrestle with stuff like that throughout the process. And when I write my newsletter, I often try to show the process that I’m going through because often I think the thought process is just as interesting as the final result because you have to wrestle with this fuzziness.

00:32:17.72 [Tim Wilson]: When there’s a big part of that, I think it was the one hit wonders one where there were kind of a series of pieces that all seemed like you were like, this seemed straightforward. It was straightforward and it was absolutely did not pass the sniff test. Nobody is going to believe that You know, the grateful debt is, I can’t remember what the examples were, but you were like, so I have to then change the way I do the analysis. So that was kind of a step removed from the data collection, right?

00:32:49.48 [Chris Dalla Riva]: Yeah. I believe you’re talking about, I did want about two hit wonders. A coworker of mine asked me, we were joking. at work about the band 38 special that they have two, they have these two songs that we would both always hear in the supermarket. And he was like, oh, I wonder who the greatest two hit wonder of all time is. And I was like, I think I’m the guy to find out for you. But this seems like a really simple question. All right, let’s just find every band that had two top 40 hits and nothing else. You can easily download the entire history of the Billboard Hot 100 going back 65 years. And then I’ll sort those artists by their current popularity on Spotify. also data that’s easy to get. And then the number one, two hit wonder, the most popular two hit wonder was Pink Floyd. And I was like, that cannot be right. How is that possible? This is one of the greatest selling bands of all time. But Pink Floyd didn’t really release a lot of singles in the US. They were mostly known for their albums. And Pink Floyd truly only had in the US two top 40 singles, Moeney and Another Brick in the Wall. And there’s a lot of other weird bands that end up on this list, like The Killers or The Cure, you know, bands that are tremendously successful. And I was like, that can’t be right. So then what I had to do, I was like, all right, you have to have two top 40 singles and you can never have had a top 10 album. And then suddenly you end up with a list that more passes the smell test. You get a group like AHA, who is most famous for their huge hit take on me, but they had another top 40 hit in the US. They’re more popular in Europe. Crowded House for our 2000s kids, Cascada. had two hits evacuate the dance floor and every time we touch. And when I was looking at this list, I was like, I was like, this, these feel like legitimate to hit wonders, but you have to grapple with like, You have an idea in your head about a lot of this stuff and then you follow it through and it doesn’t make any sense. And you realize that your intuitive definition is not accurate enough or does not reflect reality in the way that you think it does. And that applies in the work I do in the book and my newsletter. But I think in data analytics, generally that applies. Like you have to figure out how you want to define things.

00:35:16.31 [Tim Wilson]: But then you wind up down the path of telling your business partner that Yeah, you just add the business equivalent of what are the top two hit wonders. When you come back and say, well, that actually didn’t make sense. I have to give you enough other little context of where I had to correct for that does seem like that’s a regular challenge of the analyst.

00:35:41.41 [Julie Hoyer]: I was going to say this is such a good example of, I remember being a newer analyst and like, I think music is so personal to people and everyone can connect to it easily. So like the type of things you’re talking about Chris and running into, I think it’s more obvious to people like, oh, well, you have to consider XYZ or did you think to ask yourself this type of question? But when you’re a newer analyst and you get into data that you’re not as familiar with or don’t have as much context around it, I think you fall into the trap of taking it at face value and not asking some of those probing questions of like, how is it collected? How are they determining certain aspects of or characteristics of the data? And so it’s just interesting that music really shows the squishiness of the data around it. And when you’re trying to answer a question, how you really have to think through and refine the definitions and your assumptions. Because actually, I was going to ask too, in your, it’s funny that we all read your one hit wonder articles. I was reading the one album wonder article. And you had a quote in there saying, one of the most important things you need to learn when working with data is when the data is insufficient to solve your issue. And I was just curious, like, you highlight it for the one album wonder topic, but were there other topics you’ve run into where you’re like, yeah, the data just can’t really answer this question that I had about music or pop culture?

00:37:14.41 [Chris Dalla Riva]: I run into this. frequently where I want to write about something or someone suggests something and I just can’t get the information. The longest running topic I have related to this much to the anger of one friend that suggested it to me is he’s convinced that if you are a punk, popular punk musicians are much more likely to have had their parents get divorced than the average person in society.

00:37:43.22 [Val Kroll]: You need that trauma?

00:37:45.68 [Chris Dalla Riva]: I was, yeah, I was like, well, I mean, this again, this is theoretically shouldn’t be hard to figure out, but it’s just like, first you need a list of punk musicians, which you could get somewhere. I mean, I use Wikipedia a lot and there’s always like, there’s tons of lists on Wikipedia, like list of punk musicians from Australia or something, but then, you know, pairing that with who people’s parents are and if they, if their marriage survived, it’s, That’s almost impossible. It’s almost impossible. So yeah, every time I see it.

00:38:18.66 [Val Kroll]: Like the ancestry.

00:38:20.74 [Julie Hoyer]: Yeah, it was like genealogy. Maybe if he makes you your data set, then you could do their analysis, you know, like his labor of love, following up on how the marriages went.

00:38:31.59 [Chris Dalla Riva]: I do get upset about that topic because I thought it was, I mean, it’s a grim, but it’s sort of a, sort of a funny topic. And I sort of believe, I believe his theory too, but you would just need, not only would you need that, but I would also, I would want to have comparisons for other genres than comparisons to like the general population. It’s never gonna happen, but you have to know when to say, I can’t write about this because I don’t have the information. Like data can’t address, that’s something that is measurable, but you can’t get the information. Then there’s other stuff. It’s like, it’s probably not measurable in the right way.

00:39:09.74 [Tim Wilson]: a Python code. That is another one that kind of parallels to the more qualified leads instead of hit singles would be that your interest, the topic of interest is around punk musicians and their parents, but you can’t really draw conclusions without seeing if that is great. Somebody thinks, well, if you had all of that data and then you just, you’re like, yeah, but even all that data, that wouldn’t tell you if that is how that compares. So all of a sudden, the data set that’s needed has to, expand to some degree.

00:39:57.85 [Chris Dalla Riva]: This is a problem I have frequently because I do some stuff where there’s a preexisting data set and I crunch the numbers. But I tend to, a lot of the stuff I do, I have to build a data set, which leads to, I think some interesting writing, but it’s a pain to do. So that case would, it’s never going to happen, but we can do it.

00:40:19.84 [Julie Hoyer]: Do you ever time box yourself on analyses or like,

00:40:24.71 [Tim Wilson]: The publisher probably did for the…

00:40:29.62 [Julie Hoyer]: You definitely did but like in general when you get like a I guess two two thoughts here one How do you source some of your? Like hypotheses going into an analysis and then once you get into it Are you like okay? I’m gonna give myself X amount of time or X amount of effort Or you just kind of go until you hit an obstacle that’s a lot you can’t overcome.

00:40:49.59 [Chris Dalla Riva]: Yeah, like I’m so I publish every Thursday and really pretty religiously and though maybe the only thing I’ve learned on the internet over the last decade across social media platforms is that you should just put stuff out, even if it’s not perfect or finished in the way that you want, because you’d never know what people are gonna connect with. And I’ve had stuff where I’m like, ah, this is okay. It’s not perfect. And then it ends up doing really well. And then you have the thing you’ve worked on for two and a half months and you put out, no one cares. So I usually, I just usually spend a couple of days on this stuff, unless it’s something like, I know this is gonna take more time. I’m gonna publish it. in a couple of weeks, but it’s usually in the matter of a week I’m going to sit down. I have a list of topics that whenever I think of something, I’ll be like, all right, let’s try to do this. If it’s not going anywhere, move on to something else. But I know that I have to put something out on a Thursday, so I’m very schedule-oriented. So, I’m always working under the clock. Funny if the book was less like that because I couldn’t get a publisher for this book for so many years that by the time I found a publisher, nearly all of it was written. I had to edit it, but I wasn’t on a time clock as much as then with the weekly newsletter, which I’m in complete control of.

00:42:08.32 [Val Kroll]: That’s nice. So then your list of topics, are those like the hypotheses like Julie was saying? And how do you come up with that list? Is it like comments or questions from prior weeks in publishing or how do you come up with that?

00:42:21.63 [Chris Dalla Riva]: Yeah. I mean, at this point, people do ask me a lot of questions that if I think I can get information on, I’m willing to take a hack at. But other stuff is just, I don’t, I spend a lot of time thinking about music. I work, my day job is in music. I play in bands over the years. You know, I’ve worked on this book. I read about music. So stuff just comes up over time and I just write it down. And then when I sit down one week, I just see which one seems interesting.

00:42:54.36 [Val Kroll]: What inspires you? So do you tell your bandmates, like, I have all this data, I know all the ingredients, the secret sauce for the perfect song. We just have to follow the AB classic pattern. Got to have some heartbreak in there.

00:43:08.21 [Chris Dalla Riva]: Sadly, no, I actually joke about this at the end of the book because this is a frequent question. It’s like, have you discovered the secret? And I’m like, well, if I discovered the secret, I would have wrote the song instead of the book because probably more lucrative and would take much less time. So yeah, I mean, there are, The people I play music with joke that if they make an observation about music to me, they may see it in a newsletter. But no, there’s no real secret sauce. I mean, the only thing I could tell you is that things that are pop, when something’s very popular, usually there are many other things that are popular that are similar-ish to it. So if you want to write a really popular song, you should be working within the constraints or the structures that are popular at a given moment.

00:43:58.48 [Tim Wilson]: Given on that front, though, did you pick up on any sort of dimensions, very clear inflection points that something shifted? We intuitively think 60s pop sounds different from 70s pop sounds different from early 80s to late 80s, but is it always kind of a gradual continuum or were there ones where you said, no, like the invention of this technique or something shifted everything?

00:44:30.06 [Chris Dalla Riva]: One of the key themes throughout the book is that musical innovation. This is my theory, not my personal theory, but what I write about a lot. Musical innovation is often downstream of technological change. So things are gradual, but there are certainly moments where something is invented or something changes and it fundamentally shifts the way things sound. You’re talking about the 80s. And it’s actually, it is interesting because even people who only listen to music very passively can usually listen to something and have a general idea when it’s from. The creation of drum machines in the late 1970s, which became ubiquitous throughout the 1980s, is many, many popular songs are created with drum machines. That’s sort of an inflection point where the sound of popular music changes. In the early 90s, hip hop becomes very popular. And that’s a fundamental shift you can see in the data on multiple fronts. Because for the first half of the 20th century, your most popular songs were usually focused on melody and harmony, you know, something you could sing along to in the shower. And then by the late 90s, early 2000s, Popular music becomes much more focused on lyricism and rhythm, not to say that there’s no melodic qualities of hip-hop, but you’re not gonna hum, lose yourself by Eminem in the shower. I mean, just, it’s not. Speak for yourself, Chris. That’s not a show. But you can see stuff like this, and I write about it, it’s like, It’s words per minute in songs. I mean, it dramatically increases in the 90s because hip-hop is a much wordier genre than the things that come before it. And one of my other favorite examples I always talk about where in the first half of the 20th century, and I measure this, there’s a larger portion of songs without vocals, so instrumental portions of songs. After 1960, if there’s a front person in a group, it’s almost always the singer. We basically take this for granted that the singer is the star, but that wasn’t the case for decades upon decades. If you look at all these old big bands, the band leaders usually like the tuba player or the clarinet player. Part of the reason you see this shift is just because microphone technology improves. Previously, If you were, again, imagine a world without microphones. If there’s an orchestra playing, how are you gonna hear a voice over it? You’re gonna have to sing at the top of your lungs, and even then, you probably need a choir of people to be heard. But in like the 30s and the 40s, microphones get a lot better, and suddenly you can be Frank Sinatra, and you can sing with this very soft croon, and it can be heard over everything. with clarity, that technology fundamentally changes, you know, who’s the star in groups and which vocal styles can even be performed. Again, if you can’t pick up a quiet voice, then you don’t get Bing Crosby or Frank Sinatra. It’s just a different singing style. So, longest short is, yes, there are these inflection points, and I don’t think it’s deterministic, but technology really does shape how we hear songs and what songs are made, and you see that in musical data.

00:48:00.05 [Julie Hoyer]: I have to admit, when you said musical technology, because I’m obviously not a musician, my brain went to the whammy bar. That’s like all I could think about. I was like, what’s musical technology? That is a musical technology. So yeah, that’s really it.

00:48:15.85 [Chris Dalla Riva]: No, that is totally. That’s even like the electric, the creation of the electric guitar, like you don’t get rock and roll if someone, someone, Leo Fender or whoever doesn’t, figure out how to plug in a guitar and distort it and whatever.

00:48:30.81 [Julie Hoyer]: Yeah. Tim, before you cut us off.

00:48:33.58 [Tim Wilson]: No, no, I got to shout out the Roland TR-808 drum machine, which the 99% Invisible did an entire podcast episode about, and also not really a musician, so I didn’t. But when you said the drum machine technology, like they walk through an exhaustive explanation similar to what Chris, I just, there’s some listener who’s listening to other podcasts and saying, but what about the 99 PI episode?

00:48:58.91 [Chris Dalla Riva]: I have heard that episode. That’s a great, I mean, that whole podcast is great. Yeah.

00:49:05.52 [Julie Hoyer]: That’s awesome. Cause as I was going to say, I want to ask one more question. I’m curious. I’ll allow it. Do you? Why? Thank you. I want to know, because again, when I was reading through some of your newsletters, you mentioned that you go through this process. Sometimes you have to iterate, iterate on definitions. And I love that you show your readers that process as you go through. But it made me think, are there any analyses, as time has gone on, that you’ve gone back and wanted to redo? Or are there any that you did in the past that you’re like, oh, that was such a bear, I hated it, or that was my favorite? I kind of wanted to get a hit list of your Top 10 analyses are top five.

00:49:45.73 [Val Kroll]: You know, your top charts, your top five charts, whatever category you want to pick.

00:49:51.60 [Chris Dalla Riva]: Yeah, there are a couple that come to mind. Every Christmas season, I try to do an addition on predicting new Christmas classics, usually from some data approach. And I try to… I’ve tried to use Spotify playlist data. I’ve tried to use looking up songs. What are people looking up on Wikipedia? Like I said, I love Wikipedia.

00:50:13.25 [Julie Hoyer]: I never would have thought to use Wikipedia as a data source, by the way. But you talking about it, I’m like, oh, wow, it sounds like a plethora of information.

00:50:22.59 [Chris Dalla Riva]: It is. I use it in the book and I use it in the newsletter from time to time. But yeah, I go back to that Christmas one a lot, not because I think that the earlier ones were bad, but I don’t think it’s constantly changing and the ways I’ve tried are not exhausted. I’m sure there are other ways to pick up on which Christmas songs are becoming more popular. There’s others sometimes.

00:50:47.78 [Julie Hoyer]: Have you predicted some correctly?

00:50:50.64 [Chris Dalla Riva]: I think the two… I don’t think this is a surprise. Moedern classics or on their way to classic dumb are Santa Tell Me by Ariana Grande and Underneath the Tree by Kelly Clarkson Seem to from the last decade seem to be getting more play There are some things I write where I start taking an analytical approach. I’m like this isn’t working and I just end up, I don’t always write, sometimes I just write about like the music industry or some trend, not really from a quantitative perspective. One that comes to mind was I wanted to write about the evolution of music and supermarkets. Just how has that changed over the years? What you hear when you’re in the produce section? And I couldn’t get historical data on it. I was talking to some people who like companies that they pipe in the music to stores and I ended up not getting it. And then wrote a sort of a history of music and supermarkets, which was sort of interesting. If I could get that data. I could get that data. Yeah. So there’s a lot of stuff like that where… I can’t get the data, and I’m at least like, well, there’s something interesting here that I can write about and not quantify in the way that I was hoping to. But yeah, there’s sort of a laundry list of topics like that.

00:52:20.33 [Tim Wilson]: Well, man, I feel like we could ask a million more questions. Plus, it’s just fun to have somebody who’s got a near encyclopedic reference point for examples of X, Y, and Z on the music front. But now we’re gonna actually take a quick break with our friend Michael Kaminski from ReCast, the MMM and GeoLift platform, helping teams forecast accurately and make better decisions. Michael’s sharing bite-sized marketing science lessons over the coming months to help you measure smarter. So Michael, over to you.

00:52:56.78 [Michael Kaminsky (Recast)]: In our last measurement byte, we talked about the fundamental problem of causal inference and how our inability to control for unobserved differences across individuals makes observational causal inference either extremely difficult or just impossible. The beauty of randomization is that, in an experimental setting, if we’re able to truly randomize across individuals, we don’t have to do any statistical adjustments at all. That randomization takes care of adjusting for observed confounders is of enormous utility, but that it takes care of adjusting for unobserved confounders is truly miraculous. And the reason why is beautifully simple. With large enough sample sizes, if we can truly randomize treatment, all of those individual differences will simply average out between our treatment and control group. and we can empirically show this either with a mathematical proof or via simulation and code. The magic is that since this works for everything we can observe, we know it will also work for everything we can’t observe. This is the magic of randomization, and that’s why a randomized experimental design is always the first stop of every researcher. It just works. It not only handles the hard work of statistically adjusting for observed confounders, but without any extra effort, it works the miracle of adjusting for your unobserved confounders as well.

00:54:07.97 [Tim Wilson]: All right, so if you enjoyed that mini lesson, Michael and the team at Recast have put together a library of marketing science content specifically for analytics power hour listeners. For everything from building MMMs in-house to communicating uncertainty to your board, head over to www.getrecast.com slash aph. That’s www.getrecast.com slash aph. So now the last thing we do on the show is go around the horn and share a last call, something that you’ve read or seen or heard or thought about that might be of interest to our users. And Chris, you’re kind of just a fountain of last calls on this whole episode. So, but do you have a last call, something you haven’t shared that you’d like to share?

00:54:54.79 [Chris Dalla Riva]: Yeah, there’s this very, very popular music YouTuber named Rick Beato. Also, this has nothing to do with analytics, but for some reason I think people may enjoy this. Rick Beato is very, I think he’s probably the most subscribed to person on YouTube that just talks about music. It is a lot of great stuff, but he’s been in a war with universal music recently because they’ve been filing copyright claims on his videos because he usually has like 10 seconds snippets of songs and he’s been posting about. them trying to take down these videos. And there hasn’t been a resolution yet, but there’s been a couple videos and then other people in the music YouTube space have been posting, backing them up saying this should be fair use. So I’ve wrote about this a little, but I’ve been watching with bated breath to try to figure out what’s going on here. Like he’s definitely providing traffic to music in the universal catalog. So it doesn’t seem like something they should really care about. given that the ad revenue they would generate from his videos would be a rounding error on their balance sheet. So I’m just curious to see where this goes. But I just checked, right before this, I just checked. I don’t know when this is going to come out, but there was no update up to this moment. But when this comes out, maybe you could see if Rick Beato has resolved his issues.

00:56:19.63 [Tim Wilson]: Well, and you did, we’ll link to the newsletter that you did about that so people can refer. That was kind of a fascinating one about what the options are when you push. music. The reason you can find that stuff on YouTube is because the person who uploads it can say, I want to leave it up, but any ad revenue goes to the rights holder for the music. Is that right?

00:56:45.91 [Chris Dalla Riva]: Yeah, yeah. That’s generally right. If I use 30 seconds of a Chapel Roan song, you know, her people can claim the revenue from the video and they can say, all right, you can leave it up, but we’re going to collect the money. But usually 10 seconds and it’s criticism, you know, that should be in my rudimentary understanding of fair use. It should be fair use. But here we are.

00:57:08.77 [Tim Wilson]: Here we are. Nice. Val, what about you? What’s your last call?

00:57:14.02 [Val Kroll]: All right, so this one is actually related to the topic. One thing that I always find so interesting when it comes to music and thinking about quantification is genre. And however many genres you think there are of music, there’s like a hundred more than whatever that number is. And there’s this website called Every Noise at Once. Have you heard of this one, Chris?

00:57:38.20 [Chris Dalla Riva]: Oh, I am intimately familiar with every noise.

00:57:45.27 [Val Kroll]: So this site has, it’s a visual kind of display of every genre that exists. And you can click it and it will play like examples of those songs or you can like search for an artist to figure out exactly where they fall. in the spirit of like quantification of genres is kind of a fun thing to play around to figure out like what is the difference between techno rave hard industrial techno and belgium techno well turns out a lot a lot is different so anyways it’s kind of just a fun thing if you’re looking to burn 10 minutes or however deep in your hole you get maybe an hour in my case so

00:58:22.08 [Tim Wilson]: Just thinking if you homeschooled Abby, how that would somehow be part of the core curriculum for when she turns eight. You bet. You bet. I love that. Nice. What about you, Julie? What’s your last call?

00:58:36.17 [Julie Hoyer]: Well, I wish I was. on trend and had something about music. It’d be much more fun. But actually, this one does loop back to you, Chris. I was reading your newsletter, and you had made a shout out for BI bytes, and you had linked to an article by them about time series data. And so I was reading through that, and I actually thought it was a really good article. It’s called Same Data, Different Questions, Transforming, Time Series Data for Better Insights. And I love that it was because we talk about time series data a lot, especially in the industry that we work in. It’s just something you run up against. And so I love anything that helps break down better ways to think about looking at time series data or how to analyze it. And so I love that it was a list of four to five different visuals that are really helpful when you’re trying to break down and understand like, is there any significant change in this time series data? So I thought that was awesome. So thanks for leading me to that.

00:59:34.27 [Chris Dalla Riva]: Oh, no problem.

00:59:35.07 [Tim Wilson]: Yeah, Tim, what about you? Yeah, I missed that. So now I’ve got something to go through.

00:59:42.04 [Julie Hoyer]: I love that I found something Tim didn’t find. So I will. That’s rare.

00:59:46.84 [Tim Wilson]: I didn’t know if one of us was going to do the Kassi Kassarkov article that we were having an exchange about, but we’ll save that. Okay. Well, I will march even farther away from music with a past guest current. superfan of Ben Stansel, but in one of his recent newsletters he linked to something he wrote back when he was still at mode back in 2020 called a dispassionate examination of the empirical evidence regarding positional punctuation in SQL. So this was basically for many listeners or sequel writers. So the old debate around the leading or the trailing comma as to what makes more sense. And he basically concluded that, yes, objectively, putting a leading comma at the start of the line is like the better way to go. But the trailing comma is just kind of, that’s kind of still what people do and it doesn’t look weird. So it’s funny because he’s a hilarious writer, but I did want to go ahead and quote the way that he wrapped it up, said, so the next time you find yourself questioning where commas go in SQL, SQL query select statements, the answer is simple. While leaders lead with leading commas and trailing commas are leading signs of failing lines and the tail lines, no matter the database breed, we’re not agreed that it’s best to concede to lead because the more we scale our query needing, the more we follow the trail to trailing from leading. because it’s people who do the reading. And that was written pre-LLMs hitting the mainstream, but it’s… Nicely done. Yeah. So, all right. And with that, thank you, Chris, again, for coming on. I had high expectations for how much fun this would be, and it exceeded those dramatically. So, thanks for coming on.

01:01:45.56 [Chris Dalla Riva]: Oh, it was a pleasure even though I do use trailing commas in my SQL queries.

01:01:57.05 [Tim Wilson]: As do I. I think, yeah, it’s a fun, I mean, it’s actually kind of an empirical analysis. So he does, he pulls from different places. So it actually, maybe there’s a little bit of a link to a similarly trying to answer a question that maybe doesn’t have deep consequence, but he dug into it anyway. So thanks for coming on. That was a great discussion. We’ll have links to your newsletter and the book. And yeah, while I was just thinking about that, I was like, am I going to link to every single song or album that was referenced? And I will tell you, no, probably not.

01:02:35.98 [Val Kroll]: It’s not as thorough as Chris’s data set.

01:02:40.30 [Tim Wilson]: Oh, that. Have you considered or have you or are you going to ever like publish or share the data set itself?

01:02:47.47 [Chris Dalla Riva]: Oh, it’s out there. I would love if you shared. It’s just in a Google sheet. Anyone can use it.

01:02:56.05 [Tim Wilson]: Okay. We will definitely add that for now. I wish I’d taken a look at that before.

01:03:02.74 [Val Kroll]: Yeah.

01:03:04.42 [Tim Wilson]: So many more questions. Okay. So for that next book, we’ll have you back. So everyone out there, thanks for listening. As always, we would appreciate, love to have a review or a rating or both on the platform that you listen on. Should that be a possibility? We do kind of ebbs and flows on our end, the volume of requests for podcast stickers that we get. But if you go to our website, there’s a simple link in the global nav and we’d be happy to send you a sticker or three or four for free, just fill in a little form. I am hosting this, so I’ll go ahead and say, if you haven’t bought a copy of Analytics the Right Way, a business leaders guide to putting data to productive use by yours truly and Joe Sutherland, the holidays are coming and nothing is the perfect stocking stuff or like a book on analytics. We’d love to hear from you. Feel free to reach out to us on LinkedIn. You can reach out through the podcast page. You can reach out to any of us as individuals. You can reach out on the measure Slack. You can just send us an email at contact at analyticshour.io. So regardless of what what era, what decade you are stuck in listening to and thinking about, and now you’re gonna go back and re-listen to and analyze and whatever songs you’re not, you’re listening to and getting distracted by when you should be looking at the data. For Julie, for Val, I’m Tim. Keep analyzing.

01:04:38.41 [Announcer]: Thanks for listening. Let’s keep the conversation going with your comments, suggestions, and questions on Twitter at @analyticshour on the web at analyticshour.io, our LinkedIn group, and the Measure Chat Slack group. Music for the podcast by Josh Crowhurst.

01:04:56.37 [Charles Barkley]: Those smart guys wanted to fit in, so they made up a term called analytics. Analytics don’t work. Do the analytics say go for it, no matter who’s going for it? So if you and I were on the field, the analytics say go for it. It’s the stupidest, laziest, lamest thing I’ve ever heard for reasoning in competition.

01:05:15.83 [Tim Wilson]: You should tell that story to, although he probably has everybody tells him their like quirky billboard stories now that.

01:05:23.98 [Val Kroll]: I was telling, do people tell you all the time?

01:05:28.12 [Chris Dalla Riva]: No, no, I’m curious what this is gonna be.

01:05:30.33 [Val Kroll]: You’re like, it’s one of two directions. No, I’ll try to be brief. When I was in sixth grade, I went to CD Warehouse with my allowance and bought the Billboard top 10 CDs from 1960 to 1969. It just became obsessed. And so on Fridays, for music class, everyone could bring a CD. We wrote the names of the board, teacher would draw a number, you got to play your song. and the week that the new Spice Girls album came out, I brought Billboard 1965 and played Eve of Destruction, which is like a very heavy topic protest song that’s about four and a half minutes long. And it wasn’t the favorite of the people in the class. Surprisingly, to my surprise, this is not what everyone was dying to hear.

01:06:15.90 [Val Kroll]: But I was obsessed with it.

01:06:19.71 [Tim Wilson]: That’s awesome. And he said the bullying was already happening. Oh, yeah. Or that’s what. Yeah. That’s good. And that really helped.

01:06:28.44 [Val Kroll]: I couldn’t figure out what what I was what I was doing wrong, you know.

01:06:34.45 [Tim Wilson]: I mean, if you made that choice, it feels like maybe maybe there are other choices being made.

01:06:42.03 [Val Kroll]: should have been like, okay, all right, maybe right. Oh boy.

01:06:51.63 [Tim Wilson]: All right. Okay. With that, I really wanted to have that recorded. I’ve heard that now three times.

01:07:00.80 [Val Kroll]: Never on it’s still funny. Yeah, it’s good. Maybe I’ll give you a couple bars at the end in my best Barry voice, or what’s his name, Barry?

01:07:10.19 [Tim Wilson]: That’s gonna be your last call. There’s this hot song.

01:07:13.15 [Val Kroll]: Yeah, it’s a TikTok sound, but yeah.

01:07:28.03 [Tim Wilson]: Rocked flag and disco duck.

Podcast: Download | Embed

Subscribe: RSS