0:00:06.0 Announcer: Welcome to The Analytics Power Hour. Analytics topics covered conversationally and sometimes with explicit language.

0:00:15.4 Michael Helbling: Hi everybody, welcome. It’s the Analytics Power Hour. This is Episode 272. It was pretty early on in my career when I learned that the way we tracked revenue and marketing differed from from other departments. In fact, there were three different acceptable definitions depending on who you were talking to. In data and analytics, we don’t really talk about that too often, but it really starts to matter as those metrics become more complex and are imbued with meaning intended or not. And today we’re diving into a world where sometimes the numbers don’t feel like they add up. They go on these complex journeys involving weighted averages and obscure formulas and maybe a little sprinkle of dark magic. Because sometimes the most important numbers in your business are the ones that nobody actually understands, maybe including the people who created them. Okay, so we’re gonna talk about it. And of course I’m joined by my co host today, Moe Kiss, director of data science at Canva.

0:01:17.9 Moe Kiss: Hi there. How are you doing?

0:01:19.7 Michael Helbling: Hey. Good. Sorry, I usually ask you how you’re doing and I kind of left it, so that’s fine.

0:01:25.8 Tim Wilson: Gotta keep you on your toes.

0:01:26.9 Michael Helbling: And of course speaking of things that no one understands, Tim… No, I’m just kidding Tim. My other co host, Tim Wilson, co founder and head of solutions at facts & feelings. Hey Tim.

0:01:40.3 Tim Wilson: How you going?

0:01:41.2 Michael Helbling: Thank you. I’m going quite well. And I’m Michael Helbling, I own and run Stacked Analytics. So we wanted to bring in a guest who could help us navigate this discussion. I think we found the perfect one. Professor Dan McCarthy is a leading expert at the intersection of marketing, statistics and finance. He serves as an associate professor of marketing at the University of Maryland’s Robert H. Smith School of Business, where he’s recognized for pioneering the field of customer based corporate valuation. His research has been published in numerous journals, as well as in publications like the Wall Street Journal, Harvard Business Review, and the Financial Times. Outside of academia, he co-founded Zodiac, a predictive analytics firm acquired by Nike in 2018, and and later co-founded Theta, a company specializing in customer based valuation models. He holds a PhD in Statistics from the Wharton School of the University of Pennsylvania. And today he is our guest. Welcome to the show, Professor McCarthy.

0:02:35.8 Dan McCarthy: Yep. Thanks so much for having me.

0:02:38.0 Michael Helbling: Do people call you Professor McCarthy or Dan or what do you go by?

0:02:42.7 Dan McCarthy: They usually do. And I usually promptly say, “Just call me Dan.”

0:02:45.9 Michael Helbling: Oh, okay, well, I’ll call you Dan then. That’s perfect.

0:02:49.9 Tim Wilson: Do you call. Do you call Dr. Peter Fader just Pete? Is he… Like, is that…

0:02:54.7 Dan McCarthy: That’s what I do, yeah.

0:02:56.1 Tim Wilson: Okay.

0:02:56.7 Dan McCarthy: There’s a lot of Professor Faders, too. Yeah.

0:03:00.6 Michael Helbling: Well. And yeah, and we had the fortune to have Dr. Fader on our show a few years back as well, which was awesome. All right. So I think diving into a conversation like this one of the things that we recently went. Went through was some material I think you presented probably to one of your classes, perhaps about constructing customer acquisition costs or CAC. And that was sort of a conversation we sort of had a little bit about, okay, yeah look at this and see how he’s breaking it down. But you spend a lot of time kind of like pulling metrics apart that are kind of big for companies. Like, first off, maybe what got you headed that direction in terms of research and work and then maybe share anything you want to share about sort of like that direction.

0:03:49.0 Dan McCarthy: Yeah. I’ll be frank. I think a lot of that really stemmed from Zodiac and Theta. I think that in academia, there’s not much appreciation for getting the number exactly right. The bookkeeping exercise is not… It’s not that sexy. It’s not that exciting, but it’s incredibly important. I mean, if you’re a firm, it really pays to know, is my CAC $40 or is it $100? And depending on what you choose to include or exclude, you really can see these really big swings in the numbers. And because it’s such an important metric, there’s a lot of companies, public and private, that will basically do whatever they can to make the numbers look as good as possible. So for CAC, it would be to make the numbers look as small as possible. And so part of it is really to understand when you start tearing apart pre-IPO prospectuses, what are you going to see and what is real and what is bullshit. Pardon my French.

0:04:57.4 Tim Wilson: Totally allowed.

0:04:58.0 Dan McCarthy: You know, it’s just… Yeah, it’s amazing how certain companies will try to kind of dip their hand into the candy jar and just go a little too far. There’s… I think, the realm of reasonableness where they’re doing things. Maybe I wouldn’t fully agree with everything that they do, but it’s kind of within the gray zone. And then there’s some where they’ve clearly gone over the line. And there’s been a number of examples, I think…

0:05:24.2 Tim Wilson: Charlie. Charlie Javice?

0:05:26.7 Dan McCarthy: Oh, well, that I would say yeah, that certainly… That’s outright fraud.

0:05:31.2 Michael Helbling: Yeah.

0:05:32.2 Tim Wilson: Fraud. Okay.

0:05:32.3 Michael Helbling: That’s way out there.

0:05:33.9 Dan McCarthy: Yeah. Yeah. Because they had defined, in that case, I think they had said to JP Morgan, we have this many users and this is how we define a user. And I think that to me is part of the key. If you don’t provide a definition, then you’re not on the hook. And if it’s a non-GAAP term, then you’re not really required by the SEC to provide a definition. There’s no universal standard definition for terms even as common as the churn rate. And so Netflix, back in the day, I forget what year this was, but it was back when they were a DVD by mail company. They had been hit by a class action lawsuit over how they had defined churn. And the investors had claimed that they defined it in a way that understated the churn rate. And I think it was that they divided… They were basically using like the number of… In the denominator, they were using the number of subscribers at the beginning of the period instead of taking an average of the three months. But the reason that the judge threw out the case was because they very transparently provided their definition of churn rate.

0:06:43.7 Dan McCarthy: And so if you really wanted to, you could kind of go back in and you can make other assumptions. So they said come on. And there’s no universally… You know, there’s no term… There’s no GAAP definition for the churn rate. Now that’s a very benign example. But in the case of Charlie Javice, they said we have this many users and they all like created an account and did stuff and like they were all fake.

0:07:10.9 Michael Helbling: Yeah. It just wasn’t true.

0:07:13.1 Tim Wilson: But that’s… I mean that’s actually kind of an interesting, unintentional, interesting distinction between being well and well-intentioned. But it still gets murky or simplified and maybe should not have been simplified versus… I mean I would think most people, they fall into, they’re simplifying it. They may have some subconscious or subtle… They may have some motivations to pick a definition that makes their CAC a little bit lower. But it’s not overt, out and out fraud. Right? So we’re kind of playing more in the realm of these things feel kind of benign and oh, I’m just picking whichever one makes sense when in reality it matters. It actually does matter. Is that fair? That make any sense?

0:08:04.9 Dan McCarthy: I think so, yeah. I think there’s different intentions for an issue and honestly for the churn rate, I’m not even sure it was a mistake. That’s why I call it an issue because like God did not come down from the heavens and say churn rate needs to be defined where we divide it by three. In truth, even that’s probably not technically correct. And I actually wrote a whole academic paper about how the churn rate itself is not a very useful measure. It really is not.

0:08:33.5 Moe Kiss: Hmm.

0:08:34.7 Dan McCarthy: And so…

0:08:35.6 Tim Wilson: Can you say more?

0:08:36.2 Dan McCarthy: You know, what happens is it tends to go down over time, is what usually happens. So if you think of a SaaS firm, usually you have a number of people you acquire them all and then you have a lot of people who try to churn early. And then you have some people who don’t churn early and usually they stay a lot longer. And so what ends up happening is it’s as if the churn rate within a cohort starts off really high and then it goes lower, effectively. And so you could think of the overall churn rate as being this kind of weighted average of all of the churn rates of your cohorts. And so as you get older, more and more of your base are these highly tenured customers that don’t really churn very much. And so you kind of expect that the churn rate is going to tend to move down over time from when the company was first launched to like when it’s like a mid sized company. And so it’s not… But it’s not a sign that anything has actually gotten better. It could be that every single cohort is identical to the previous cohort, but the churn rate will move down just because you have this early shakeout that tends to happen when you acquire a group of customers. So you know, as soon as you start kind of going down that road, it’s like, well, what are we really going for here?

0:09:48.7 Michael Helbling: I mean, it could almost be a negative signal. If new customer acquisition is lagging, then churn would look better than it should because you’re not even getting that new customer churn.

0:09:58.7 Dan McCarthy: That’s one of the corollaries that actually sometimes when churn, when the churn rate moves down, down, it’s a negative signal because it could mean that acquisitions is falling because that means even more of your base is going to be the older customers because you’re not bringing in the newer customers as much.

0:10:18.4 Michael Helbling: Note to self, when a software as a service company starts touting their churn rate top of the page, dig deeper.

0:10:26.5 Tim Wilson: Imagine this, your data instantly wherever you need it. Like having cold brew coffee on tap. Ah yes, data on tap. Giving you that little pick me up that’s the dream, but in reality, it’s more like data on hold. That is, until now. Fivetran already the kings of automated data integration, they just acquired Census. Yeah, that Census, the reverse ETL pros, which means now it’s all connected. Source systems. Check. Data lake. Check. Apps like CRM and marketing tools. Check. It’s like your fata finally turned 16 and got its driver’s license, and stopped asking you for rides everywhere. With FiveTran and Census, your data just works everywhere with more insights, less waiting, zero headaches. Fivetran the only pipeline you’ll actually want to maintain. Go check out their site with interactive demos and more at fivetran.com/aph. Once again, that is F-I-V-E-T-R-A-N.com slash A-P-H.

0:11:37.2 Tim Wilson: Is that always kind of a challenge when you’ve got a numerator denominator like CAC as well? That, that’s always kind of a blended metric where it can move… Two things can move it if it’s a ratio. Right? Like that does that always… Maybe I’m making a over generalization, but I think even in your… Like your presentation on CAC saying, look, you’ve got a numerator and a denominator, you can make this number go down by increasing the denominator or reducing the numerator.

0:12:10.4 Dan McCarthy: Yeah, make the numerator smaller, make the denominator bigger. But yeah, with the… So it’s almost like there’s kind of a different set of issues with CAC. So with all the churn dynamics, that’s more that there are kind of customer life cycle effects. The things are happening as a function of customer tenure. And then we’re looking at this aggregate metric that kind of blurs across all of the cohorts within a certain calendar time period. And that it just can create some dynamics that it’s not that the churn rate would be bad, it’s just that you would need to really look to the cohorts to really understand what’s going on. And to me, what would be bad would be that the cohorts are degrading, like they’re getting worse. Like the cohorts I’m acquiring more recently are worse than the cohorts I acquired before. For CAC, yeah, I think that you’re right. That you kind of have some notion of how much did you spend to bring them in and how many did you bring in. And those are like the two fundamental pieces. And when you think about it like that, it becomes pretty clear how companies will try and make CAC smaller, is they’ll start removing a whole bunch of costs from the numerator and they’ll start trying to find ways to count more things as acquisitions in denominator. And so there’s just tons of examples of both of those. You don’t want to go down a rabbit hole. But there’s a couple examples that are just so good for the latter point that…

0:13:34.2 Tim Wilson: Well do share.

0:13:35.6 Dan McCarthy: So Warby Parker, they prominently put their CAC in their pre-DPO prospectus and they have charts in their filing that show the CAC over time. So on the one hand I was very happy to see that, but on the other hand, they put a footnote that said we define CAC to be acquisition related sales and marketing expense divided by the total number of active customers. And the active customers, they could have been acquired recently, they could have been acquired five years ago. They’re putting everyone in the…

0:14:07.4 Moe Kiss: And as long as your user base is getting bigger, it’s going to look better. Oh wow.

0:14:12.7 Dan McCarthy: Yeah. Well, even if it’s, even if it’s getting worse, it has to be… The active customer base over the past year has to be bigger than the number of customers that they acquired. Because it’ll always be include them, but then it’s gonna include all these other people potentially. Yeah, so that’ll drag the ratio down. And so yeah, I kind of scratched my head like, “So this is customer acquisition costs.” So why, you know. And it was featured in a few different places. And yeah, there was no real pushback about it because you know, what can you really say? Like that would be another measure, but it would not be customer acquisition cost. It would be like marketing related active customer costs or something. So that was one and then the other was Toast, the POS company. They would define their CAC where the denominator not only includes acquired customers, which in their case are stores. Like the more stores that they’re in, like the store is the unit for them. But they also included any upsold customer. So you know, they had the Toast basic and they upgraded to the Toast Premium or something new customer. And I was like, well…

0:15:28.2 Moe Kiss: Oh.

0:15:29.8 Dan McCarthy: That’s something but I’m not sure that’s an acquisition. So…

0:15:35.7 Michael Helbling: That’s interesting.

0:15:36.2 Dan McCarthy: Yeah, so yeah, it’s kind of conflating a notion of like acquisition with a notion of development. I would consider development to be pretty fundamentally different.

0:15:46.2 Moe Kiss: Out of curiosity, I mean, I feel like I spend my life talking about the intricacies of metrics like this and the pros and cons and the different ways to calculate it. And I don’t know if this is controversial, but is the problem the way you choose to define it or is it a bigger problem when you change those definitions? So we calculated CAC one way. Now we’re incentivized size to make this number bigger, this number smaller. So we’re going to change it Is like… What’s the danger zone for you? Is it the definition to start with or the changing of it down the course?

0:16:24.9 Dan McCarthy: Usually for me, it’s the definition to start with. Like, I’ve got a pretty wide latitude where I… Yeah, I would say this is within the… It’s in the fairway. You know, I can kind of quibble here and there, but it’s largely reasonable or at least it’s kind of in line with norms. But when certain companies deviate really far from the norm and it’s just totally whack, then that’s hard for me to get behind. And it can create issues because then other companies will see hey, this company is reporting this really low number. Our number is bigger, but it’s bigger because we’re defining it correctly. I think it creates a lemon effect that a lot of companies feel like they need to use the worst definition because otherwise they feel like they’re going to be reporting numbers that might be perceived as being poor. I think the other thing is, usually, at least from what I’ve seen, most companies when they start disclosing a metric in their filing, usually they don’t change the definition. Usually the thing that they’ll do in the public filings is they’ll stop disclosing.

0:17:24.4 Dan McCarthy: And so they’ll say, we used to disclose the churn rate. Now it’s not relevant for our business. We’re doing X, Y and Z, whatever, so we’re just going to stop sharing that anymore. And Netflix is an example. Actually, just last week was the first quarter in their history where they stopped disclosing the total number of subscribers that they have. Yeah, it’s like, ah, there’s like a piece of me that dies when I see that, you know. And then they give the reasons. You know, just to say that you’re going to not disclose it, it’s important. But they start saying all these weird reasons why it doesn’t matter for them anymore. Like we’re moving into advertising. And you’re like, well, if you’re moving into advertising, it’s even more important now because you need to know how many people you have and how much they’re watching, right? Because that’s what determines the amount of time that people could be exposed to ads. But…

0:18:20.0 Tim Wilson: But there is kind of an allergy to a corrective. Like I most like… I feel like there’s sort of two things that can happen. And using Toast, I don’t think this was the case. But if, like if Toast was selling and they only had one offering when they were new and then all of a sudden they said, oh, we have a… We have another offering. And then somebody somewhat innocently was just because the way the data they’ve just been counting the number of rows in this table and then the way that the data got captured. Because nobody was really thinking about it along the way, said, oh, we just keep counting the number of rows in this table. And then by the time somebody says, wait a minute, maybe this doesn’t make sense. Then they’re having to have a discussion of saying, wait a minute. If we correct that definition… And this is, this is all, I’m sure not the case. But I think of times like when Facebook kept screwing up some of their metrics for like what a video view is. I mean I feel like in digital marketing that happens all the time. Like what is an impression or a viewable impression or a video view is a great example because it’s…

0:19:31.3 Moe Kiss: Video views kill me.

0:19:33.5 Tim Wilson: Yeah. Is it, they start the video, they watched at least 10 seconds. It was auto playing, is included, isn’t included. And it does seem like it’s this combination of forces. One, like what’s the incentive to make the metric look as good as possible. But also it’s not really thought about. Somebody just came up… Some mid level analyst was like, well I’ve been told to pull this number and I found the way to pull it and wasn’t thinking through connecting back to wait a minute, what is this really trying to measure? And I’ve got multiple parties who I’m trying to use one metric to serve multiple people. Back to the revenue example, Michael. You can have multiple definitions of revenue and they’re all correct because GAAP revenue is for one reason. You know, raw bookings are for another reason. But people want revenue to be revenue. And then you wind up in these arguments about like, but what’s the right definition? And that’s a case where it’s like, well they’re all right, they’re serving different purposes. Sorry, I just, I think I just headed down like six parallel paths. So if any of those made any sense, feel free to run with it.

0:20:47.2 Dan McCarthy: Well, I think the key is the if… It is helpful to have kind of proper definitions. And I think it’s okay to have what I would call like a taxonomy. And so revenue. I’m not sure I would say revenue should be the same as bookings, but I like the fact that we have the term bookings, you know. So I think as long as it’s clear, because otherwise then when someone says we did this much in revenue, then you really don’t know what that even means, you know.

0:21:18.3 Tim Wilson: Oh, used to drive me nuts because people… But I think there are a lot of people who think revenue is revenue. And so when they say revenue, they don’t know what a booking is. So somebody simplifies it and says revenue and then bookings versus billings. Like I remember like shouting, well, metaphorically shouting from the rooftops, because I was like, we had two different people referring to revenue and reports out to the company and one was referring to bookings, one was referring to billings. And I was like, these… You’re confusing everyone because these two back to back sets of slides don’t make logical sense. They’re being reported by different departments. Better off to say I’m reporting bookings and this is what a booking is. But wait a minute. Sorry, I think I just went down a little PTSD path there.

0:22:04.1 Moe Kiss: Can I go down a totally divergent path then?

0:22:07.0 Tim Wilson: Please do.

0:22:08.1 Moe Kiss: I actually do love to talk about CAC and I also love to talk about LTV to CAC, purely. Because I feel like it’s a metric so many people just like roll off their tongue and like investors and boards and everyone, like it’s just something conceptually, people seem to mostly get. And I think the thing that’s top of mind for me… So like one thing that really stood out from some of your work was talking about the like lead lag between spend and marketing. And I guess I’d just love to hear from you how do companies ultimately do this really well, noting that there are so many players in the market who make kind of poorly incentivized decisions.

0:22:53.4 Dan McCarthy: Yeah, it’s a really hard question because when you start disclosing these measures for external stakeholders, I think the goal is often a little different than if it’s purely to help optimize your business. Because usually if you’re trying to optimize your business, you probably want more granular measures. They might be division specific. You can get a lot more in the weeds. I think when you’re putting a figure in public filings, then it really pays to have something that’s a little simpler or it almost kind of has to be. And so I think that I’ll have this when I give my CAC lecture. I’ll have kind of like the level one definition, level two, level three, level four. But even the four definitions I provide, they are still more like what would be on executive dashboard for the CEO as opposed to that operator that works in the marketing department. But if you think about what a public company would disclose, usually they would disclose what I would call the Level 1, the most basic definition. And that’s usually something like sales and marketing expense divided by customers acquired both in the same period. And it’s not the best definition for many reasons.

0:24:08.3 Dan McCarthy: And that’s why we kind of peel back the onion. But from an external reporting, SEC disclosure standpoint, it’s so simple. And it’s also probably the most common. It’s most commonly disclosed. So you also then have the ability to have some comparability across firms. Whereas if you start kind of messing around with leads and lags in the public filings, it becomes a little bit… It makes them very hard to compare to one another. And it’s kind of yet one more like degree of freedom that you’re providing the companies. And I just worry a little bit that they’re gonna again, dip their hand into the cookie jar and for growing companies, just start going with these crazy long lags. So, yeah, yeah, because that will make the denominator a lot bigger. So…

0:24:53.8 Moe Kiss: Do you think there’s a challenge, though, when we’re talking about CAC specifically? Like… Because there’s different levels of business maturity. And I suppose the thing that I’m kind of trying to grapple with is I feel like the data folk often in the team who are doing the work they understand this complexity. Like, we were having a discussion about LTV the other day, and it’s like, we can get into the weeds and be like, this is the way we would think about it here. And this is the consequence of projecting out further, and all these different caveats. But for the business, and particularly for the leadership team I do think they think of that level one definition and probably don’t fully appreciate the value of getting to the level four all of the intricacies. Like how much do you think you can drive that conversation versus how much should you just try and tailor it to the different stakeholders and their requirements? I guess.

0:25:49.8 Dan McCarthy: Yeah, it’s a really good question. I do think if… I make… I draw a pretty fine line between internal and external. If you’re the CEO of the firm, you should really care about like the best definition. That’s like the most… Gonna drive the best possible outcome for the firm. You know, I mean. So yeah, I think hopefully it might be a little more complicated, but hopefully everyone can kind of get around that as being like I’m okay to kind of take the cough syrup because I know that it’s going to lead to my firm making more money. Now I do think executives, they kind of serve two masters. There’s obviously long term shareholder value creation which that’s where the pedal hits the metal. But there’s also hitting guidance. And I’ve just heard so many examples of these really, really big ecommerce firms where it’s more for metrics like the number of active customers. They know they have a certain customer count number that they’ve guided the street to. We’re getting close to the end of the quarter. What do we have to do to be able to hit that number? And they start doing some things that they probably would not have done had they been purely focused on long term value.

0:27:06.4 Dan McCarthy: I think it’s just an example where hopefully in an ideal world they should be focusing mostly on the best possible definition, even if it’s a bit more complicated. But then you kind of have to acknowledge that in the back of their mind they’re also going to be thinking about any measures that they might be disclosing to the street. And it might cause them to focus on both. But at least it should be both. It shouldn’t be only like the simpler definition that they provide to the street.

0:27:37.9 Tim Wilson: Well, so what about like a pre IPO scenario where maybe you’ve got something like monthly active users. Do you… Where you’re like, do we include the free users versus the other ones versus people who are actually paying? Is that one where you can kind of count on the sophistication of the investors? I mean I don’t know if you seen that and I’m not… Now realizing this sounds like I might be talking about another co host company. And I have no idea how that company worries about that sort of thing. But like what about… There’s the street and the street’s reaction. There’s kind of the pre… Like at what point when there are external parties that you’re trying to serve, are there cases where you’re like this is a small enough group that we can have a more nuanced definition and they will both… You know, they’ll skewer us if we have the simplistic definition because they’ll realize where the numbers feel a little fudged, but they’ll also be able to get on board with us and figure out what the nuances are. Does that make sense?

0:28:49.6 Dan McCarthy: Yeah.

0:28:50.3 Tim Wilson: I guess it’s another frame of the, how much do you… Trying to weigh this simplified versus what’s better when tied to a better measure of value or progress? How much do you have to… That’s like another audience. There’s the executives who have to be able to internalize it and speak to it. There’s… If you’re public, there’s the street and the analyst ability. If you’re not public, but you want to be public or you want to be acquired or something else, they’re the people you’re selling to and you’re trying to maximize the value, you realize that way. Are people generally sophisticated enough to get on board for the more nuanced definitions?

0:29:35.2 Dan McCarthy: I think private investors, you probably can be more. They’ll be more sophisticated. But I think also being on the other side of the aisle, as a person who started a few companies. Yeah. I think you’re also keenly aware that anything you say can and will be used against you. So you start disclosing a metric, and I think as we were talking about a moment ago, then people will wonder, if you stop disclosing, why did you stop disclosing? There must be a rat. And so if you get really sophisticated and granular and you start giving them all that information at the beginning if that starts to kind of plateau or there’s some noise in the data you’re going to have to probably continue to disclose that. And I don’t know, I think a lot of people, they try to kind of strike a good balance where they don’t… They don’t want to be too influenced and have to be… Have to do things that they wouldn’t want to do and also spend an inordinate amount of time explaining little wiggles and jiggles they might not actually be enduring. So, yeah, they would be sophisticated, but maybe too sophisticated.

0:30:52.4 Moe Kiss: It sounds like you spend… Like you do read a lot of kind of public filing documents or reports from different companies. When you go through them what are you looking for? It seems like you have a really keen understanding of oh, they’ve defined this here, and this metric is no longer being shared. Like, what are the things that you’re watching out for when you’re reading those reports?

0:31:14.2 Dan McCarthy: Yeah, usually it’s for the… I’m like a bird watcher for these metrics. And there’s certain areas, there’s the figures and then there’s the tabular data. And usually a lot of the good stuff is in the figures where they show these charts. Like Spotify in one of their investor day presentations, they’ll show trial to conversion curves by cohort. And they had all these different lines for the different cohorts. And it was really kind of cool. And I think it was just an example of the sort of richness that you can get in some filings that you would typically not see, like on a regular, ongoing basis. But once you’ve seen enough of the filings like that, then you kind of know what to look for. Like, there’s a certain type of customer disclosure that you often see. And so I’m just kind of typically just focused very much on like, where are the customer numbers? I just go… Beeline towards the customer numbers.

0:32:09.5 Michael Helbling: Yeah, and actually that’s a good segue to maybe talk about sort of like why you look for those and what that means to you as you kind of analyze that. Because that’s sort of the field that you kind of kind of pioneered a little bit.

0:32:22.9 Dan McCarthy: Why did I start looking? Honestly when you have a hammer, you start looking for nails. And for me, I think for a lot of companies, they’ll put these figures in there, they’ll put these disclosures in their filings, and it’s more like window dressing. But I look at it and I see a way I can calibrate my model. It’s a really valuable input data point for some model for acquisition or repeat purchase or spend over time. I really need as much of that data as I can get to be able to have a more… A more accurate model, a model that I can trust more. So yeah. So for me, leaving out data points would be really, really bad. And I think if you are someone who is kind of doing a similar type of exercise, it brings like a whole new level of reality and you start seeing all these issues that you wouldn’t have otherwise seen. Because if you have metrics that are not internally consistent with other metrics in the filing and you start trying to calibrate a model on it, you realize these two data points cannot happen at the same time.

0:33:40.4 Dan McCarthy: Like, either the one has to be true or the other. And so there’s likely some sort of definitional conflict between them. But if you weren’t… If you weren’t actually building a model, you probably wouldn’t necessarily even see something like that. So yeah, I think to me that’s what got us started. The original work in my paper was on Blue Apron and Sirius XM, two publicly traded companies. And then after that, the first company that I really kind of gained some notoriety for was Blue Apron. And in the case of Blue Apron, the reason I could kind of carry out that analysis was because in their pre-IPO prospectus they put a whole bunch of exactly the sort of data that I would have needed to have run my model and because they did that, I could do what I do.

0:34:26.6 Moe Kiss: And what kind of data was that?

0:34:28.4 Dan McCarthy: So first they put in like cumulative customers acquired from one date to another date. They basically gave us that data point, which is extremely important. Having some customer acquisition data is the only way we’re going to get a CAC data point. So yeah, companies that only disclose active customers, for example, really hard to get a CAC estimate. They had active customers over time, total orders over time. And then they also disclosed it was like a customer lifetime. Over the first six months of the customer’s life, they did this much in revenue. Cumulatively, over the first 12 months it was this much. And I think they provided that for a couple of cohorts. So that information is really helpful to be able to calibrate these sorts of models because it gives you like a longer lifetime view of how these customers monetize. And usually that’s typically very hard to get unless you have some kind of cohort level data.

0:35:26.3 Michael Helbling: I want to jump to something else that just sort of is, I’m curious about and maybe you have a perspective on which is we’ve been talking about some of these metrics and how they’re not necessarily governed by GAAP and maybe the SEC doesn’t really probe them too deeply. Is there interest by like FASB, the Financial Accounting Standards Board to like define these? Like, are we outrunning the standards currently? Is that kind of what’s happening here?

0:35:50.8 Dan McCarthy: Yeah, we’ve spoken with FASB actually and we had submitted a letter we were trying to get a case open about it and you know, they were very interested in it. And we’ve actually, we’ve also spoken with some of the people at the SEC, just some of the economists there. And I think the big issue that the people at FASB had is that it’s very hard to have like a one-size-fits-all rule for any measure sure. You know, I think that the churn rate for a B2C business versus a SaaS business there just can be these different norms. And if you have a conglomerate… There’s all these weird edge cases, you start going to like, what is a customer? You know?

0:36:36.3 Michael Helbling: Yeah.

0:36:37.7 Dan McCarthy: And if you operate in multiple different business lines, then you can consider that customer to be… Like the same person could be a customer at multiple of the business units, or it could be at the full at the overall company level. And then there’s concerns about not… Different companies can have different levels of trackability of the customer. So if you have a business like grocery stores, some purchases are made in cash. Some grocery stores have loyalty programs, others don’t. If you don’t have a loyalty program, it’s probably gonna be harder for you to track how many customers you have. So it’s like you start getting really bogged down. So I think basically what that can motivate is not having broad, sweeping statements that apply to every single firm in every single industry, but maybe like a narrower definition. Like, this applies to SaaS firms. We kind of treat the SaaS firms in a more homogeneous way. So I think that’s kind of where we ended up leaving it. I think that my hope is that we have kind of informal standards that we can get to where within SaaS, there’s the net revenue retention rate, there’s certain metrics that you typically will see in SaaS, but you don’t really see them a whole lot elsewhere.

0:37:51.5 Dan McCarthy: If we can get everyone to disclose that sort of data point and use the same definition for it, I would consider that to be a win. So I think informal standards is probably the best we’re gonna hope for. The other big trend is that we’re not in an environment of growing the number of disclosures. And when you speak with people at FASB, there’s been this big trend away from IPOs in America. Our market share of IPOs is just terrible. We represent a shrinking… And fewer and fewer companies are going public, and they’re really concerned about it. And so imagine if that’s kind of the backdrop and you’re like, here’s five other rules that you need to do, you know. Disclose all these other things and jump through these hoops. Well, they might say, well, now I’m even less interested in going public. So yeah, it’s a practical concern that… Yeah.

0:38:47.6 Michael Helbling: I was going to ask you about that, too. Like, the number of IPOs that have sort of… Like post COVID, it seems like they’ve been sort of dropping and you don’t see companies, especially like tech companies and things like that, jumping into IPO like they once did. I don’t know. I mean, so obviously, like adding more regulation might do that. But do you think there’s even like some chilling effect of like, even pushing the numbers they use internally out publicly is just sort of like, we don’t want to deal with it. Like, I don’t know what your thoughts are on that.

0:39:21.4 Dan McCarthy: Like, they don’t want to deal with just the issue of…

0:39:24.5 Michael Helbling: Yeah, just the scrutiny and like the headache, honestly, of like meeting guidance and like trying to get these numbers to work for them.

0:39:34.4 Dan McCarthy: Yeah. I mean, historically why would a company go public? You know, usually it would be for liquidity. But if they want someone… They want an open market for their stock and easier ability to raise equity down the line, instead of having to do the F round and then the G round and the H round. So to kind of have a market like that that can be a plus. But yeah, I think increasingly, I think companies are just raising more rounds and staying private for longer.

0:40:09.1 Michael Helbling: Yeah.

0:40:10.5 Dan McCarthy: But it is kind of interesting that of the companies that are going public, a lot of them are not profitable. Like, most of them are not profitable or weren’t profitable. I think we’ve had this chill, but kind of even before the chill, the companies that were going public were more mature and less profitable. So, yeah, I guess it would make sense for them if those are the people who can benefit the most from having a more liquid market for their equity. You know, that can be sensible. But yeah, I think it also is just incredibly expensive to go public. So, yeah, so I think the cost benefit may not be as good as it used to be.

0:40:50.8 Tim Wilson: Going back to the discussion of kind of the standards, just because my, my skin starts to crawl. Like some of the desire for having standards I think articulating, like all the reasons that it’s tough to try to define a standard. I do feel like there are cases where companies, they sort of hunger for this standard because they feel like then they’ll get a benchmark that they compare themselves to. Whereas it seems like, as you articulated it, the best thing to do is to try to have a metric that makes the most sense for your business that over the long term is a predictor of maximizing discounted net present value of cash flows, whatever. And when you start to try to put definitions, I feel like there’s a… There’s a tendency for internal. Like we did a whole episode around benchmarks, which, I mean, I get triggered on. Because people want to say, well, what’s the norm? And I want to be above or below that norm. And then I’m incentivized to tweak my definition, where I can, to kind of beat that norm. As opposed to really thinking about the fundamentals of what is my company’s strategy, what is the nature of my business, to how should I define a customer? How should I define… Yeah.

0:42:20.4 Moe Kiss: Can I throw in an example? I thought the one Dan, that you just referenced before, I really wanted to dig into, and it kind of ties in with Tim’s point. So I’m just going to like, jam it in there. To do with even how you count customers. Right? Like, let’s say we’re going to develop a standard or a norm for how you count customers. The point you made before is spot on. What happens if you have multiple business streams and a customer exists across different business streams. In your opinion, like what should the norm there be? Should it be that there… It’s like a unique person. What happens if a person has two different accounts? Like what… How do you create those norms? Because the reality is, is like those… Like you said, Tim, each company is so unique. Creating standards involves a level of commonality which doesn’t always exist.

0:43:13.0 Dan McCarthy: Yeah, it’s different. I mean, first, as a customer centricity person, the whole argument for customer centricity is to break down the silos. You know, we need to think about the holistic value of customer across the firm. And if we’re getting revenue from division A, but it’s just cannibalizing from division B didn’t do anything for us. So from a customer centricity standpoint, that would argue for a holistic definition of the customer that spans the enterprise, because that gives you more of a North Star that the whole company can be aiming for. We want to expand the overall value of this customer to the firm. So I think there is a real place to have that metric. But if you’re a conglomerate, does that really play out to the same degree? I mean, maybe. But like Berkshire Hathaway, you got See’s Candy on the one side, you got GEICO on the other side. I don’t know.

0:44:10.1 Michael Helbling: Does it matter if I’m a customer of both?

0:44:12.4 Dan McCarthy: Yeah. And how the heck am I going to tie that customer together? There’s just no way, you know. So…

0:44:17.8 Tim Wilson: Well but… Does that go down even to the level of like take JP Morgan Chase where they’ve got business banking and consumer banking. And while they on the one hand they want to have both, but one per the nature of somebody in their personal consumer life. I mean there are cases, super high net worth people, maybe it kind of blends. But like that feels like one of those where…

0:44:41.5 Dan McCarthy: May be separable.

0:44:41.9 Tim Wilson: A customer in business banking is a company, a customer in consumer banking is either a person or a household or something. Right?

0:44:50.8 Dan McCarthy: Yeah. Or you know, for Verizon they’ve got a small and mid sized business unit like that’d be very clearly different from consumer on so many levels and I don’t think that you would benefit and you’d probably be harmed by trying to lump them all into the same bucket because then…

0:45:10.3 Tim Wilson: Not that you don’t want to market and try to detect the ones who can be cross sold because they’re an individual and you can detect they also run a small business. But that feels like okay, well that’s a measure for how many of these people can I expand my share of the overall wallet or something? I mean this is hard.

0:45:30.4 Dan McCarthy: And you know, the other thing is you have people who are probably assigned to those units and especially if it’s a very large… If they have a very large portfolio of company. Like you have these apparel holding companies that have multiple different brands within them. Anthropologie, Free People and you know… You got the Free People people and you’ve got the Anthropologie people and they’re going to try to grow their business. But you know, I think that it can pay to have someone that kind of sits on the top of them that can help better understand what’s going on, overall, for the firm as a whole. And… But practically speaking you’re not going to have one person… You’re not going to have everyone who’s cross functional across every single brand within the apparel portfolio. So I think you end up having to at least keep track of… You probably ideally would want to keep track of both. I do think hopefully the… There could be some analytics department that is aware of kind of what’s going on across the whole portfolio at least to some degree. But you can’t really expect every, every division to operate like that.

0:46:43.1 Moe Kiss: Do you feel like you come up against strong opinions here? Like I feel like whenever how we define metrics come up, like I don’t know, I just feel like they get… Emotive is probably the wrong word, but I do feel like it’s an area where people have either very, very strong opinions or like it’s… They’re just like, oh, it’s so complicated. I just totally tap out. I find there’s like very rarely the middle room. Is that your experience, or am I like misreading?

0:47:15.1 Dan McCarthy: There are some people, and I’ll admit I’m probably one of them, so guilty as charged. But…

0:47:21.6 Moe Kiss: Plus one.

0:47:22.0 Dan McCarthy: Yeah, you get pretty strong opinions about my metrics and how things should be defined and… And so if you have two people who have different opinions, it can become almost like religion that it’s really hard to get anyone to change their opinion. But you know, there does end up being a point where if it is purely a definitional thing, oftentimes both of the things being argued for are valuable, but in different ways, and it becomes like a nomenclature thing. And then it’s like, well a rose is a rose. You want to call it this, you want to call that that. As long as we have different names for the things, then I think that that becomes really important. Again, that goes back to the bookings versus revenue. So a lot… This comes up a lot with customer lifetime value because there’s so many different definitions for it that are bandied about. And I want to make sure everyone’s on the same page if we’re talking about this, that we know what this actually is and is not. So there you really have to have some sort of hierarchy. You know, we’re talking about this one, this one, this one, this one.

0:48:31.0 Dan McCarthy: It could be revenue, it could be finite horizon, it could be undiscounted. This one could be contribution profit base, it has a discount rate involved. Goes out one year, goes out three years, goes out five years. But we explicitly build that into the term, the term’s name so that we know, oh, you’re talking about five-year revenue per customer. And I wouldn’t call that customer lifetime value. But you know, if you go to mobile gaming, oftentimes they just look at a measure like that as customer lifetime value.

0:49:02.5 Moe Kiss: It’s funny, we actually did that in my team. We created predicted subscription value for that reason. We needed like an LTV basically. But obviously I work in marketing, so we want to project it out from the time the customer’s acquired. And the thing that… We intentionally named it something very different. And we also wanted to have the word predicted in there to make it clear that it was looking forward and not based on historical data. And the bit that’s wild is like, we do often come up against people who are like, but just give me LTV. And you’re like, how do you want us to do that? Like, you described to me how we would do that. And it’s like this thing where sometimes yeah, I feel like there’s such an art in the naming as well, because people have this tendency to preference the metric that they think is super simple, like LTV or customer lifetime value, whatever you want to call it. And they think that that’s a simple calculation. And you’re like, especially when you have three users, it gets really complicated.

0:50:03.9 Dan McCarthy: Yeah. Yeah.

0:50:05.3 Tim Wilson: I love thinking of their brain melting when you’re… I mean that… I was like, okay, we’re just going to wait till they all die and then we’ll get back to you in 100 years and we’ll let you know. I mean, I.. Yeah.

0:50:17.3 Dan McCarthy: Yeah. That’s when you’d be able to actually observe it as you have to wait for every single person in the cohort to churn. Everyone.

0:50:25.1 Michael Helbling: Every single one.

0:50:26.2 Dan McCarthy: How long is that going to take?

0:50:28.0 Michael Helbling: We just start dialing. Hey, could you just cancel your account? There’s 10 of you left, and I need some…

0:50:33.8 Dan McCarthy: Yeah, you got to churn now. We need this metric.

0:50:38.7 Michael Helbling: All right. This has been such a good conversation and honestly, so educational as well. So thank you so much, Dan, for joining. And unfortunately, we do have to start to wrap up, although I feel like we could keep going for quite a while. Or at least I feel like I could. I’m enjoying this quite a bit. But one thing we like to do is go around and share something of interest, anything that might be happening in your world that might be of interest. And, Dan, you’re our guest. Do you have a last call you want to share?

0:51:08.0 Dan McCarthy: You know, I’d say the big thing that we’ve got going on at Theta right now, which is the company that I have that’s focused purely on CLV and customer base corporate valuation. It’s this new version of our model called CLV Ultra, which is really it’s like the best of both worlds, super accurate. So one of the standard limitations of these kind of parametric CLV models is that they don’t have the ability to capture a lot of kind of rich variation across the customers and across time. But this model is very, very, very good at that. And it’s both made the model more automated and more accurate. So we’re kind of… It’s the closest thing I’ve ever seen to a free lunch. And so our big push this year has been to really take advantage of the automation factor that we can now almost kind of take our hands off the steering wheel and be able to just let it run because they can kind of automatically tune itself. And so we have these clients now where there’s one that’s a big lender against customers. So maybe this would kind of be the second sub piece of my big one thing.

0:52:17.8 Dan McCarthy: What they do is they lend against specific customer contracts. And so if you have a very high growth firm, it might take a while to have that customer pay back. You can actually sell that customer for some amount of cash or lend against the customer as collateral, where the only collateral against the loan is the revenue coming from that specific customer. So we’re working with this one big company that does that for hundreds of companies. And it never would have been possible for us to have worked with a company like this before because running 200 models means like all of these detailed validation checks and temporal holdouts and things like that. Now we’ve got the ability to kind of turn around a model in theory in like a couple of days. And so we actually have the ability to pursue use cases like that. But yeah, just kind of also wanted to put that in as kind of this other cool thing that we’re starting to see a bit more of, is these kind of specialty finance companies that are explicitly kind of recognizing the long term value of customers and trying to find ways to help close cash flow gaps.

0:53:31.5 Dan McCarthy: That like for SaaS firms, oftentimes it can be quite a while before they get paid back on their customer acquisition costs. And so if you’re a very young firm, you can end up with working capital issues where you kind of have to raise the next round, raise the next round. But it’s because you have so many high value customers that you know that you should be acquiring. But you feel like you end up in this weird tension sometimes where if the payback period happens to be two or three years, you don’t want to give up all of the equity that you’ve been working so hard to maintain. So if you can have some non dilutive way of being able to lean on the customers that you’re acquiring, at least some of them, to help close that working capital gap, it could be really good for everybody. So I think large scale use cases like that, coupled with a model that gives you predictions at the individual level, I think will be really, really helpful as we kind of look at over the next five years.

0:54:28.9 Michael Helbling: That’s cool. Thank you. Awesome. All right, Tim, follow that. No, I’m just kidding. That’s your last call, Tim.

0:54:37.0 Tim Wilson: Yeah. Cool.

0:54:38.1 Dan McCarthy: That was a mouthful. But…

0:54:40.9 Tim Wilson: No, that was. Oh, that’s interesting. So mine’s also a twofer. I just want you to feel good about you had a couple, so I’ll… But they’re kind of related because it… One is a David Epstein piece where he talked to Alex Hutchinson, who recently wrote the Explorer’s Gene. So the name of the piece is why you should get lost more often. And it’s kind of like an overall exploration of kind of why doing hard things actually helps in kind of dopamine and all these different reasons that not getting stuck in a rut makes sense. And it reminded me of an old Cautionary Tales episode from 2021, which I probably last called back then. That was Fritterin’ Away Genius about Claude Shannon and some of the goofy stuff he would do. So they’re all kind of bundled up in a case for doing stuff outside of your comfort zone and how that can pay dividends beyond just doing that thing. So that’s a mishmash of two or three bundled together.

0:55:51.2 Michael Helbling: Awesome. All right, Moe, what about you?

0:55:54.0 Tim Wilson: I don’t know if that counts as one or three last calls. It depends on how we’re calculating the metric.

0:55:57.3 Michael Helbling: Oh, boy, There we go. All right, Moe, what about you? What’s your last call?

0:56:02.3 Moe Kiss: Mine actually is very topical for once. Mine’s almost never topical related to the show, but this one is. I recently read the Chairman’s Lounge by Joe Aston and it is phenomenally interesting. It was not what I expected. I thought it was going to be a book about all the gossip that happens. So Qantas is Australia’s national airline carrier, and I thought it was going to be all the gossip that happens in the private lounge. No, no. It goes into all of the financial detail and the workings of the company and the CEO and his performance. And the bit that actually is super relevant for this is it goes into… Alan Joyce was the CEO very famously, and it goes into a lot of detail about how he changed things that were reported publicly. So for one, instead of looking at traditional net profit, he started to change the definition to underlying profit and then would exclude a whole bunch of costs like redundancy payments and legal fees and other payments to make the company performance look more positive. He also started pulling things out of their annual report, like the average age of aircraft, which was like a statistic they’d always always shared.

0:57:14.4 Moe Kiss: And it was because basically he was running the aircraft down and so the average age was increasing dramatically. And anyway it was honestly just such a fascinating read for me because I’ve not really thought deeply about like the short term incentives for CEOs a lot in the past. And it just made me realize how like a really nefarious self-interested CEO can have such a detrimental impact on the long term performance of a company and I just hadn’t… Yeah, it was such a good use case and like it is very well researched, phenomenal read. So highly recommend the Chairman’s Lounge by Joe Aston.

0:57:53.9 Tim Wilson: Dan’s salivating. He’s like, oh, if I’d read their disclosures, I would have nailed them.

0:57:59.3 Dan McCarthy: Man that’s… That is very interesting. But that is what you typically see, that disclosure is strategic. I mean you kind of expect it to be. They’ll start disclosing things that lo and behold look better over time and they stop disclosing things that lo and behold look worse over time. Imagine that.

0:58:16.8 Tim Wilson: Michael, what’s your last call?

0:58:18.4 Michael Helbling: Well, mine’s a little bit of a departure, but it might be less of a last call, more of a public service announcement. I’m not sure. This is my Andy Rooney time. I’m, I’m a big fan of AI and I think a lot of us are. But I’ve noticed something when I speak to people and I’m out there in the world talking to people about AI. I’m noticing that people are putting a lot of personal connection and even sometimes emotion into their interactions with LLMs. And I guess what I’m using my time today to say is please do not do that. They’re not your friend, they’re not a person, they’re a computer. And I know it can be hard because I think there’s even like incentives on the LLM side to make you feel like you’re talking to someone but you’re not. And that’s it. That’s all I have to say. I’m sorry. It’s just on my mind a lot because I hear people… I hear people I think are actually really smart talking in ways about how they interact with LLMs that is just, I don’t think it’s good. I don’t think it’s healthy emotionally and so… Yeah.

0:59:25.8 Moe Kiss: Okay. Okay. But I… like, I use please and thank you when I talk to my Google home. Because that’s the behavior I want to instill in my children.

0:59:32.1 Michael Helbling: I’m not saying that.

0:59:34.9 Moe Kiss: Where’s the line here?

0:59:37.8 Michael Helbling: I think being polite is fine. I don’t know exactly where the line is. I just know that I’m a little alarmed because people I think… that I feel like are very intelligent are falling into something. I’m like, why don’t you see that that’s like basically a computer. You should just be giving it requirements, basically. And I don’t know. And I’m… Maybe I’m wrong and I just need to open myself up to embrace AI that loves me and deal with it. But anyways, that’s my last call. All right, that’s maybe a whole other episode.

1:00:10.8 Tim Wilson: The political… I’ll just say the Slate’s Political Gabfest had like two… the guests. One of the… or guest co host and one of the co-hosts got into a raging debate about that with different, different rationales for arguing proof…

1:00:24.6 Michael Helbling: Yeah, and I could be all off. I don’t know.

1:00:27.0 Tim Wilson: why and why not? I feel like there’s… Yeah.

1:00:29.3 Michael Helbling: But I’m pretty sure I’m right. And so anyways, let’s go on. All right, let’s wrap up the show. You’ve probably been listening to the content of this show and being like, wow, where can I learn more about this? Where can I interact with this conversation? Well, we would love to hear from you. And you can reach out to us on LinkedIn or the Measure Slack Group or on by email at contact@analyticshour.io. We’re also going to include in the show notes on our website links to some of the stuff that Dan, you’ve been talking about, like the CLV Ultra and Theta CLV and some of those things so they can get some exposure to your work. And I know you’re active on some social media somewhere. Do you want to share that with people if they want to follow you?

1:01:16.7 Dan McCarthy: LinkedIn is now my main platform.

1:01:19.0 Michael Helbling: LinkedIn. So you find him on LinkedIn. Perfect. And so that’s perfect. So thank you again, Professor McCarthy. I’ll use the formal one for this one. Thank you so much for coming on the show. It’s been something I’ve been looking forward to for a long time as I followed your work. like I used to follow you when you were… used to be at Emory and now at UMD. And so really cool, really cool to have you on. Thank you so much. And of course, no show would be complete without a huge shout out and a thank you to our producer, Josh Crowhurst, all that he does behind the scenes to make the show possible. So thank you, Josh. And no matter how you calculate your CAC or your LTV, remember, I think I speak for both my co-hosts when I say keep analyzing.

1:02:06.3 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.

1:02:24.2 Charles Barkley: so smart guys want to fit in. So they made up a term called analytics. Analytics don’t work.

1:02:31.1 S?: Do tahe 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.

1:02:45.1 Michael Helbling: All right. Right.

1:02:47.3 Tim Wilson: We’re no strangers to churn. You know the signs and so do I? A full LTV is what I’m dreaming of? You wouldn’t guess it from a CAC that’s high? I just wanna model your retention, gotta make you comprehend. Never gonna spike your churn never gonna tank your LTV? Never gonna blow your CAC and hurt you. Never gonna skew the base, never gonna lose the trace. Never gonna fudge the cohort curves that woo you.

1:03:26.3 Michael Helbling: There you go. That’s… That’s… Well, we’ll see if that makes it in or not.

1:03:33.8 Tim Wilson: Rock flag and read the disclosures.

1:03:36.8 Michael Helbling: and provide a definition.

Podcast: Download | Embed

Subscribe: RSS