w/ Daniel McCarthy, Emory University
Get ready for a deep dive into what really makes customers stick around or peace out. Professor Daniel McCarthy unravels the myths around churn rates and why a low number isn't always a high-five moment for businesses. If data drives your decisions, this episode is an indispensable resource for anyone navigating the complexities of subscription businesses.
Kurt Elster
Today on the unofficial Shopify Podcast, we're going to unravel the mysteries of churn. Yes, we've got uh Daniel McCarthy, a professor. Who goes deep on these things, has published a research paper, he's been on the show before. Um last time we talked about LTV? Dan, what did we talk about last time? All things LTV. Okay, that's what I thought that it was LTV. Alright, so what's this? This is essentially a a dramatic recreation of an academic paper structured as an interview. What you you recently published this thing, what is it?
Daniel McCarthy
Well it it's all about this uh phenomenon that we find that You take a startup that is subscription-based, and what you will see is that its overall churn rate, which is kind of the most commonly disclosed measure of a company's retention trends, Uh the overall trend rate just tends to move down over time. And I think a lot of people they'll point to that as evidence that the business is gaining traction, it's achieving product market fit, you know, people really like it. Uh but in fact uh what we show is that it's not the case. It's just something that always tends to happen, even when um the customers are getting less sticky over time. So uh so really you you don't want to get faked out by that. You know, but it kind of speaks to that broader point that uh yeah, churn is really something that does matter. It's a measure of of how uh happy people are with the service, the all else being equal, if you're more satisfied with your the okay to the month or your AMA subscription, crunchy roll, something like that, then you're gonna stick with it for longer. You're gonna renew for more periods, and that's gonna mean, you know, more revenue and ultimately profit in the company's pocket. So So it is something that matters, but you really want to make sure that you're kind of measuring it the right way.
Kurt Elster
Okay. So I if I'm I'm not canceling my Netflix subscription because I'm so thrilled with them, right? It's if I'm I'm canceling it, it's because I'm not happy with the value I'm getting would be the most charitable way to think of it. Um but give us uh Dumb it down for me. Just give me the layman's explanation. Define churn for us, since that's that's the topic of the day.
Daniel McCarthy
So a churn churn the event is canceling the subscription. So if you have the Netflix subscription, you fully cancel. So you're no longer a member of that of that company. Um and the churn rate then is taking all of the people who had been subscribed in the previous period It's saying, well, how many of them churned this period? It could be this quarter, it could be this month. And just let's look at the ratio of the two. So yeah, if I had hundred a hundred subscribers in my base Uh last month and ten of them churned, then my churn rate is 10 over 100 or 10%.
Kurt Elster
All right, so stupid question. Churn is exclusive to subscription businesses. Non-subscription does not have a measurement for churn.
Daniel McCarthy
Uh so churn it's a great question. In a non-subscription setting, we know that people will churn in the sense that they will end their relationship with the firm. You know, if you take a hundred people who buy you know their first pair of shoes from Allbirds, Uh some of them will leg it and will continue to repeat by. Some of them will say, I'm never gonna buy a pair of shoes from this company again. And uh And that is effectively churned. Like they're they're ending the relationship, but the the main distinction is that we don't have the ability to actually see the cancellation vent, you know? They're just silently attriting is what they'll call it, you know, that they just kind of silently uh silently exit. So Saturn does exist in a in a non-subscription setting, but Uh, but it's an unobservable event. And so um I think for that reason, there really isn't a way for non-subscription companies to report any sort of a churn measure. And so you will often Um you only see these churn type disclosures uh coming from subscription firms.
Kurt Elster
Okay. And we can throw a curveball in here. When I go to cancel they'll offer to to retain me as a customer. They'll say, Hey, instead of canceling, especially if it's like a a physical good that's getting mailed to me, you know, like cat litter, multivitamins, whatever, they go, Well, maybe you have you already have enough. You don't need You just don't need this month's shipment. Rather than cancel, can you skip? Sure, I'll skip this month. That throws a monkey wrench into it. How do we Think of m well how does uh a subscription model where you could skip affect churn measurement
Daniel McCarthy
Yeah, basically it to me what it introduces is more variability in the monetization of customers that are with a firm. So you may have a whole bunch of people who are with you, but some of them will never skip and some of them will skip very frequently. And basically, I would say that they're both, you know, all those people are still with the firm. They haven't churned yet. But now we need to keep track of this extra process, you know, which is the you know, essentially it's the probability that someone's gonna skip or not And that can lead to to meaningful differences across the customers, you know, which can lead to again, we go back to customer lifetime value, which is what we spoke about the last time, it can manifest in pretty big differences in customer valuation then. But Um, but from the firm's perspective, it makes a lot of sense. You know, oftentimes um yeah, the the trade-off that a firm will make when allowing for skips is, yeah, hey, if If I didn't allow for skips, there may have been some people who had all that cat litter, you know, in in the in the pantry and the next month rolls around, they don't need anymore, but they they don't want to cancel, and so they just accept it. And uh and so you you monetize those subscribers better. Uh if you allow for skips to occur, you're gonna basically allow people to downshift a lot more. And so Yeah, the the revenue per subscriber will probably decline, you know, but by doing that. But then the hope is that you retain you retain your base for longer. And so I think companies are making this explicit trade-off between modernization and retention uh when they when they do things like that.
Kurt Elster
Hmm. So you your paper is called the c The Declining Churn Fallacy. Tell me what you what you did here. What was the research?
Daniel McCarthy
Yeah, so the research is really to um to kind of point out this phenomenon that um A lot of companies will say that their churn is declining when it's actually not. And yeah, it kind of speaks to this difference between uh the churn rate as I had defined it before and uh keeping track of the stickiness of customers over time that I would say If a company is acquiring customers that stay for longer and longer and longer, all else equal, that company actually has declining churn. Like the retention patterns are getting better over time. Um and that is not being measured by the churn rate. That that's kind of the main thing that we show in the paper. So Yeah, nine times out of ten, the churn rate will move down. But what we show is actually if you were to kind of look at uh a large number of month-to-month subscription businesses. that over time the the customers tend to get a bit less sticky over time. That more people in that cohort they acquire a year from now are going to churn out, you know, after six months than the cohort that you just brought in right right today.
Kurt Elster
Why is that? It's not is they're like an inherent unsolvable death spiral to a subscription business?
Daniel McCarthy
Yeah, it does. Yeah, at first. Now there's a question of magnitude. So yeah, we do find that, you know, first, you know, what we call cohort specific churn, you know, just the stickiness of the customers the acquire. you know, acquisition cohort by acquisition cohort, that stickiness does tend to decline, but it's more variable. You have some firms where yeah, the customers get more sticky. Uh you have other firms where Yeah, the customers get less sticky. And so it's not it's not nearly as universal. And it's also not nearly as strong. But Yeah, I think what it's picking up on is this um this thing that we kind of intuitively may expect that yeah, the customers that you acquire at the very beginning of your corporate life they're probably more the true believers. You know, they're they're jumping in early. And the fact that they're jumping in early is probably telling you something that, yeah, they they like the model better. It kind of it it jives with them more. And so they're much quicker to adopt. If if it takes many more years to bring in a customer, that's telling you that they're probably, yeah, again, they are literally jumping on the bandwagon late. So Um, so it it kind of makes intuitive sense that they just may not all else being equal be into you quite as much as seeing the current the kind of early true believer
Kurt Elster
Interesting. Cause yeah, initially I was like, I have no idea how or why this would be. But what you're saying is I start early with a company, I have loyalty to it. That's like, oh I You know, I was uh a true believer, I was there at the beginning, I've been around, I've seen it grow, I've seen it improve, versus late the company's later stage, there's less of that rapid growth, they're more established, maybe they cost more I've already heard about it a while. I sign up. Eh, I'm not that impressed. Right. Well I stick to it. Yeah.
Daniel McCarthy
You may jump to their competitor more quickly.
Kurt Elster
You know, it's just uh and yeah, the market around it probably has matured as well.
Daniel McCarthy
Mm-hmm.
Kurt Elster
Yep, that's definitely. Earlier you said the you know company will report declining churn as a positive thing and you're like it's actually you know um it's the other way around. The are they intentionally cooking the books or misunderstanding? Or is it because they're not they're looking at it overall or they're looking at a like a segment over time?
Daniel McCarthy
I think it is genuinely just misunderstanding. It's one of the most popular disclosures of subscription firms. And And so naturally, if you see that that disclosed churn rate is has moved from, you know, twenty percent to ten percent You call that a wind, you know, unless you knew better. Now if you're within the marketing department, you really should know better. So um Yeah, so you know, it it actually goes back a bit to how we define lifetime value. Yeah, the lifetime value is you take a cohort of customers, you look at how much you spent to bring them in the door, you look at how long that cohort stays with you after acquisition. And then you look at the amount of profit that you get while they're with you. And that, you know, you kind of smash it all together, that gets you CLB. Um, you'd want to kind of look at retention in the same way, you know, that You have your cohort, you just acquired them, they're all with you. You're gonna probably lose, you know, 30-50% of them in the first month. Oftentimes it it wouldn't be out of pattern to see that sort of thing for a a meal delivery type business. And then after that, you know, the The co the the customers that stay, they tend to be around for a while. And so oftentimes what you'll end up seeing is this retention curve that has this very sharp decline in month one, month two, month three, and then it kind of goes flat. And that if if they're with if you have 10% of your subscribers with you a year later, you probably have 10% of your subscribers 13 months later. So So in general, you know, that cobhort specific churn rate declines. You know, that by the time you get to month twelve Yeah, the churn rate of customers in month 13 from that cohort is going to be pretty small. Uh, it's gonna be a lot smaller than that, you know, 30, 50% that you'll often see in month one.
Kurt Elster
So overall churn is like f lifetime of the entire subscription program? Or is it per like twelve month just everybody in there over twelve months and then the cohort specific is groups who signed up in like X period?
Daniel McCarthy
Yeah, the overall turn is taking. Imagine we think about Netflix. It's taking every single subscriber that was subscribed last month. They could have joined a year ago, they could have joined 10 years ago. Uh and it's saying, let's just take them all together. Let's see how many people churned this month and let's just compare the two.
Kurt Elster
Okay.
Daniel McCarthy
That would give you the churn rate. And that kind of gives you a sniff of what the issue is that um if if you had some startup, most of the customers that were subscribed last month were probably young customers by construction. You know, if the business just launched a year ago, then they literally can't have any customer that's been around for more than a year. And if they're growing really strongly, most of them probably came in a month or two ago. And so you have a lot of these customers, these subscribers who were just testing the waters for the very first time recently. And so as a result, you'd kind of expect that the churn rate's going to be higher because yeah, they have a lot of nibblers and uh and they don't have a lot of die hard loyals in that uh in that in that base. Whereas Netflix, yeah, I actually I don't know what the number is right now, but I would just have to imagine that most of their current subscriber base has been with them for at least a year. Probably at least a couple years. And so, you know, those people, they've they're they're locked in. Yeah, they're they're not gonna leave Or it's gonna be much less likely that they're gonna leave. And so you just would kind of intuitively expect that the term rate's gonna be much lower for Netflix, all else equal, just because the base is older, you know? But it doesn't mean that you know, if you compare it, you know, a customer that they acquired today to a customer they acquired a year ago, that that those customers are somehow better. You know, it just means that The base is matured. They have a lot of old, you know, highly tenured customers in it. And so just mathematically speaking, that's going to drive down the the stated disclosed term rate.
Kurt Elster
In your paper, you've got the Netflix example is good because I think we can all relate to you know the the customer experience with Netflix. Um the but in your paper you've got the you've got the Hubble as an example, as a a case study. Can you tell me about that as it applies to this?
Daniel McCarthy
Yeah, Hubble, uh, they're a a contacts company. Uh they're kind of direct to consumer selling contacts. And uh And they they sell on a month-to-month subscription. So you kind of sign up, you get your subscription, you get your your product each month. And uh And they're a pretty startling example of this trend that if you look at their month one churn, which is to say, take everyone who acquired today, what proportion of them are still with you next month? And just look at that that month one churn rate over time. What you'll find is that It's kind of steadily moved up. Yeah, it's basically moved from 40% to 50% over this first 12 months of the company's life. But if you then look to the overall churn rate. It paints a much more sanguine picture. You know, that the the overall turn rate has declined from call it 40% down to around 20% and then stayed relatively stable. And so Again, you kind of have this tale of two worlds going on that if you look at that overall turn rate, you see it going from 40% to 20%, you say, mission accomplished. You know, we're we're we're doing a lot better than we were. But in fact, yeah, that month one shurned would have you being always people a little concerned, you know, that it does seem like the customers are getting less less and less sticky over time.
Kurt Elster
This episode is sponsored in part by Clever Riffic. Hey, ever wake up to customers canceling orders? Maybe they said, oh I'm bad at math and I shouldn't shop when I'm high. This was a gift, but we broke up. You're taking too fing long. And why should the customer have to ask you to edit their order for them? What if they could do it themselves? Enter Cleverific. With Cleverific, you can empower your customers to make their own changes through a self-service portal or let your team handle complex changes directly from the back end. CleverIffic is designed to reduce customer support tickets, ship orders faster, and even reduce returns and lost packages Don't just take our word for it. Peter Manning New York cut their order support queries by 99% with Cleverific. And they're not alone. Don Get this, exclusively for unofficial Shopify podcast listeners, you can get 50% off Clever Effix Pro Plan, now just $49 a month giving you full access to everything their app has to offer. Just go to cleverific. com slash unofficial and use the promo code Kurt50 at checkout to claim your discount Take control with Clever Iffic and watch your customer support tickets plummet, shipping speeds increase, and package returns decrease. We've heard a lot about how we're we're doing things wrong. I want the right way. How should we be using churn as a KPI to gauge business health?
Daniel McCarthy
Look at the cohorts. I'm a big believer in cohorts. And so you acquire customers It starts the clock. And you want to think about the performance of your customers as a function of their tenure with the firm. So yeah, the way I would conceptualize a subscription business is every month you're acquiring some people and You then want to track measures more like month one churn, month three churn, month six churn, month twelve churn. And that's saying Yeah, just kind of apples to apples. If I took customers that I acquired at a certain point in time and rolled forward the clock, you know, one, three, six, twelve months, I want to know what percent of them are still with me So that is what what they call a cohort specific measure because it's not aggregating across everyone in the base. It's saying, let's look at people from these specific acquisition cohorts and just see you know, how long they've stayed after a certain amount of time has passed. So you really want to look at those cohort specific measures and track those over time. Uh you can roll them up into yeah, I think kind of derive predictive measures that could be trackable in a singular way. Yeah, so again, the nice thing about the overall turn rate is it's a single measure. And we can just kind of look at that one thing, different periods of time, and say, um, yeah, am I doing better or worse or not? Uh as soon as you start moving to cohort specific measures, there's the question of which cohort specific measure to focus on. You know, 'cause in theory, every single point on the retention curve could be disclosed. You know, there's just a lot of them. So the other thing that can be important to track would be just the expected lifetime of a customer. Yes, if you took uh a cohort of customers acquired two years ago, one year ago today, what is the expected lifetime of that customer? You know, how long on average will they stay with you before they churn? And that is a pretty good standalone measure where you don't need to track a zillion different things. You can just kind of track that one thing and uh and pound for pound it's gonna be pretty pretty diagnostic and relevant and won't won't kind of fake you out as long as you're kind of doing the the measurement correctly.
Kurt Elster
All right. Walk me through Either step by step how I'm supposed to do this or like give me the the actionable strategy here.
Daniel McCarthy
So for the churn measures, for the enlightened kind of cohort specific churn measures, uh just look to the transaction log. So the very first time someone buys, that is when they're born. I guess in this case, the very first time that someone subscribes And the month one churn is basically, you know, take all those people who subscribed at a certain point in time and look at the churn, the number of churns relative to the number of people who subscribed one month after acquisition. So again, you can kind of do that just with some spreadsheet math. And uh it's actually something that uh I do in my CLB class. Uh yeah, I teach a a whole semester course just on customer lifetime valuation. Uh this is one of the things that Yeah, I'd I'd have the kids do. Um so you know, it it's relatively simple math, um, but very, very powerful. Uh but yeah, all you would need to do these calculations is pull up your transaction log and uh and then just look at some ratios. Now it gets a little trickier when you move to um We move in kind of two directions. One is if you don't have a month-to-month subscription business, you imagine that you had a business where you know people could only have the opportunity to turn after a year. Or if it's like Dollar Shave Club where you can kind of pick, you know, like you you prepay for a certain number of months then you you want to account for that in uh in the churn measures. Um by kind of holding aside those complications, Uh the other thing that gets a little more complicated is to compute the expected lifetime. So if you have a really, really old cohort, yeah, if if you acquired them You know, 10 years ago, it's probably the case that every single customer from that cohort has churned. The expected lifetime then is just take the churn time of every single one of those subscribers from that cohort and just average the churn time. You know, so if you had a cohort of two people And one of them stayed with you for one year, one of them stayed with you for three years, then the average lifetime of a customer from that cohort was two years, if that makes sense. Now the tricky part is if you acquired a cobhort three months ago, you will have some portion of the cobort that churned. after that point, but you probably have a whole bunch of people who are with you right now. And so you kind of need to predict when it is that each of those customers is going to uh det churn in the future. And that requires a retention or help.
Kurt Elster
I was gonna say how do you predicting customer departure? That's a cool trick. How do you do that?
Daniel McCarthy
Uh the easiest way honestly is to look to the retention curve. And again, kind of what you'll find is that the retention curve has this pattern where it kind of sharply declines declines and then starts to go flat. And there's a very, very simple model that is implementable in Excel where you kind of take that historical churn pattern And it can basically uh allow you to get a very good estimate of what the future churn pattern is going to be. So um, yes, I'd be happy in the show notes if you're interested to to share an Excel file. And uh yeah, PowerPoint. Please do. Yeah.
Kurt Elster
With these equations and you can simulate this churn model.
Daniel McCarthy
I even have a YouTube video that takes you through the spreadsheet to say, put in your data here, click this, press that, and you get your estimate.
Kurt Elster
So useful. Yeah, for sure. I need that in the show notes.
Daniel McCarthy
Yep. Yeah, so uh just remind me, I'll I'll make sure to get that over to you. And yeah, I think that the beauty is um it it really is very simple. So yeah, it'd be hard to Or to do the the deep dive here on today's show, I think it just would be a little it'd be a little bit much. But um yeah, if you kind of sit down with that YouTube video and the spreadsheet and the PowerPoint deck for about hour or two. Yeah, I think that uh you get it. So yes, I'd kind of give you a tool to be able to to project out retention for your, you know, for your business, assuming that you're a subscription business.
Kurt Elster
And okay, this is I think all of this kind of assumes you're an established business because that way you have data to work with. If I'm just a new business Do I is there anything I could do? Is this useful at all to me, or is it like just hey, wait until you've got the data, then you could figure it out?
Daniel McCarthy
Yeah, it's a great question. If you're a month-to-month subscription business, um Even if you don't have that many months of operating history, you can still use this to get, you know retention trends for the the customers that you've acquired so far. So and you don't actually need that many months to get a pretty good estimate of what those cohorts will do. you know, probably, you know, four or five months. So not not much. Now it's a lot less data than you're gonna have for a much more mature business like Netflix. So, you know, it's uh you would kind of expect a lot more uncertainty in your projections. And if you're a young firm, I'd say, yeah, the other thing that can happen is it's more likely that you're gonna dramatically change up the model somehow. You change up your pricing, change up the product. As you get more mature, the likelihood that you make some groundbreaking change to your business model goes down. You know, so so inherently you'd expect that that business will become more predictable. But yeah, I'd say the the the main type of business where it's gonna be hard to like use that spreadsheet. It's gonna be impossible to use a spreadsheet if you don't have any subscribers. So um Yeah, so you if you're kind of at that stage, then the best that you can do is look to to peer companies. Yeah, so pull up data either from uh yeah, companies like Netflix, but earlier on in their history and see what is good, what is bad, what is kind of what is normal. And you know, if you're putting together investor materials, you can run different scenarios to say, hey, Uh yeah, I know that you don't have any subscriber data for me. I can't share that with you yet, but this is what tends to be normal, you know? And um and that can at least give investors some sense of Uh yeah, if they were to kind of deliver a relatively normal quote unquote amount of performance, um we can kind of expect this from them from a retention perspective. Now, I I would say just as a final point on that, even if you have a pretty decent amount of uh subscriber data, I think it can still be very helpful to run those same sort of competitive analyses. Uh because yeah, it could be that you've you've been doing something wrong for a while, you know? Um so if your retention is way lower than every single one of your com major competitors, yeah, that might be telling you something that Yeah, you're you're leaving money on the table somehow. And so um you so I think that that sort of analysis, it can have value whether you're young or you're old, but it when you're old, at least you can also look at your own subscriber data, whereas if you're young you know, the competitive data, maybe all that you've got.
Kurt Elster
Quantitative Is there any value or any method to add some qualitative data in here? Can we get feedback from customers? Hey, why'd you cancel?
Daniel McCarthy
That could be very helpful information to know. Now, I am a a numbers geek, so my bias is always towards numbers. Uh, but what I would say is Yeah, I think that qualitative information can be helpful and in fact can be fed back into the quantitative model. So um so if you had information on on the reason that the customers churn, for example, what you may find is There are some customers where it really wasn't you, you know, that you'll you'll kind of want to to distinguish, did the customer churn because of something that I did, or was it because of some external circumstance that the customer was exposed to. Like, um, yeah, I'd say that the the most standard example of this is in the t in the telecommunication space, that they actually have pretty decent data about the the causes of churn. Yeah, they they can kind of broadly be lumped into voluntary versus involuntary. And involuntary churn would be things like they moved out of the area and they changed states, you know, or they moved to a different country. And sure, you know, that that kind of stinks. I mean the Verizon won't get that revenue anymore, but it Verizon wouldn't say I did something wrong, you know? Um they didn't really do anything wrong. The person just had to move Um, whereas if it was something like I had that service, then that would be a very different story altogether. So um Yeah, so what you can do is you can kind of add that into your analysis and be able to say, well, I may also have data about, you know, customer service and customer service touch points. to the extent that I can improve my customer service, I can then relate that to the probability that customers churn because of customer service related issues. Those customers may still end up churning for other reasons like moving out of state. But at least then you can get a much more refined view as to like what's the money that could be in it for me if I were to to dump more money into customer service and get people to to churn less often because of uh being unhappy with customer service.
Kurt Elster
Hey Shopify sellers, listen up. Get ready to unleash your creativity with the brand new landing page builder from Zippify Pages. This isn't just Zipify's biggest update ever. It's a revolution in building Shopify store designs. Say goodbye to Clunky Designs and hello to Total Customization. With Zippify's new builder, you're the boss of every pixel on your site. Drag, drop, swap elements, it's all in your hands to craft the most effective layout for conversions. Zippify pages come straight from the founder of a $180 million e-com empire. These templates aren't just pretty, they've been tested in the arena with a nine-figure Shopify store. And it's easy to use. A few clicks in your product pages marketing pages, blog pages, even your homepage will be looking sharp and conversion optimized for maximum sales. And because these templates are tested in a nine-figure Shopify store first, you know they actually work. So why settle for the standard when you can have a site that's uniquely yours? You can try Zipify pages and their brand new builder for free for 14 days when you go to zipify. com slash Kurt. That's Z-I-P-I-F-Y. com. Slash K U R T to try Zipify pages for free. And if you tried pages in the past and thought it wasn't for you, you owe it to yourself to try it again because the new Zippify pages is a game changer. Zipify. com slash Kurt. Check it out. You've got your finger on the pulse here, right? What are You're seeing the similar data across a variety of industries. Are there trends we're seeing across all subscriptions? Are there industry-specific patterns? Anything unique or interesting stand out to you trend wise?
Daniel McCarthy
Yeah, one of the big ones is uh B2B versus B2C. So B2C, you know, that's business to consumer. So that'd be things like the Hubble example we talked about where they're selling contacts direct to consumers. Uh it could be Blue Apron, HelloFresh. I guess Blue Apron would be not a good example to mention anymore because they were acquired. Uh but yeah, any mail delivery business uh that operates on subscription. Um They tend to to observe that pattern that I just described of sharp drop-off, then kind of goes flat. But one of the the other kind of key distinguishing features of a B2C business is it typically the amount of money that you get for them while they're alive tends to be pretty stable. Again, just think your gym membership or your your cell phone subscription or Netflix. You know, it's that same monthly rate. And yeah, yeah, it's possible that people can kind of move up to that higher service tier. Um from SD to HD or or what have you. But typically, you know, on average it stays relatively flat. And so if the retention kind of sharply drops and then goes flat, then the revenue is going to sharply drop and then go flat. And so you kind of would just have to expect that. And B2B, it's it's really different. Um Yeah, for one, they're acquiring much fewer customers, because their customers are businesses. And there's just much fewer businesses in the world than there are humans. Uh So uh and then oftentimes yeah that they will have uh a lot more modernization while customers are alive So they'll keep track of this thing called logo churn. And logo churn is churn as we've been talking about thus far. Yeah, if they acquire a hundred businesses and ten of them churn, their logo churn is ten percent. But the other thing that they'll often track is what they call the revenue churn. And revenue churn, it's kind of um It's kind of a bastardized metric, in my opinion. It's because it's kind of gloming together multiple different behavioral processes into one thing. But what it represents is Basically the same thing but for revenue. So I acquired a cohort in the first month I got a hundred bucks of revenue from them. If I now am getting 110 bucks of revenue from them, then my net dollar retention is 110%. Because I'm making Yeah, 10% more than what I started with. And so revenue churn would be negative in that case. It would be negative ten percent. Um Now that's gonna be a function, it's gonna be driven by both logo churn, like if you have customers churning out more frequently, you got less money coming in the door, right? Because they're not gonna be bringing any revenue to you. But the other thing that kind of drives it is the amount of revenue that you get while customers are alive. And in B2B settings, what you often find is the amount of revenue that a company will get. from its subscribers tends to dramatically increase over time. And uh and the the reason is there's kind of a handful of different reasons, but I don't know how many people in the audience have used Slack before, but I think they're a great example. They've got a whole bunch of people who uh are startups that start on that that lowest tier paid plan. And if you're a startup, you might have yeah, 10, 20 employees. But what happens is as As that company grows, it raises the next couple rounds of funding. Uh, it's gonna have more employees. And so maybe it's gonna have 60 employees instead of, you know, 20 employees. And because Slack charges per seat, then all else being equal, that means triple the revenue. Uh but the other thing that happens is As you grow more and more established, you start having to pay for all this extra stuff like enterprise grade service and data security. You know, the list goes on. So typically companies like Slack will have kind of their basic plan, but then they'll also have the enterprise plan. And the enterprise plan will not only it's gonna be more expensive per seat. You know, so in addition to having the 60 employees, you're getting more revenue per employee uh from that company than uh than you had gotten originally. So yeah, so again, Slack is gonna have customers churning out either. They may churn out because they moved to teams, but they may churn out because You know, a lot of startups go bust, but from the companies that do you know end up succeeding, they get so much more revenue from the succeeders that uh their revenue churn uh becomes pretty steeply negative. I know it was a mouthful, but hopefully uh you know it kind of gives a sense of the difference between those two settings.
Kurt Elster
So I've noticed in in marketing discussions Over time I've been doing this long enough to over time KPIs change. Like this is the right KPI. Or like don't use that one, use this one. Yeah in like certainly we've we've seen fr go from ROAS to CAC to like there's just different way to MERS different ways to get similar KPIs to measure similar objectives. In churn analysis Do you think that that KPI will change? You know, in the future, where do you see this type of business analytics going?
Daniel McCarthy
I do see uh either fewer companies disclosing the churn rate, or at least if they disclose it, they're holding their nose appropriately. So Netflix, for example, I keep mentioning them. They've done a great job of poo-pooing the mag trick in public. I I I kind of say that lately. They'll very often say, hey, look. The composition of our base is maturing. That's going to push our turn rate down. Yeah, so they make it very clear that they kind of know what's going on here. So yeah, I think it at least at the very least we're gonna see that. Um I think any one metric can be gamed. And so uh I like the the appeal, the simplicity of having a single metric, yeah, that kind of summarizes success. But I think it it it's very dangerous and can lead to some of the cycling that you're alluding to. Yeah, imagine for SAS SAS firms, probably the the the most popular disclosure besides paying customers is net dollar retention rate, you know, which I just mentioned. And there's just all sorts of ways you can game that if you really wanted to. You know, you can have everyone come in, but when they come in, they're they're spending a dollar. And then you just immediately ratchet them up in month two to some higher number. And it sets this really low base, oh, you know, uh, for for the net dollar retention rate to expand over time. It basically would encourage lots of companies to move towards land and expand business models relative to uh trying to monetize their customers um earlier on. So yeah, so I think yeah having a holistic approach is gonna be important. Yeah, I think Yeah, these different cohort specific churn measures are important. I think the the net dollar retention rate can be helpful. Um but you know, you also need to keep track of CAC Yeah, if you're spending a zillion dollars to bring customers in, you could have the greatest net dollar retention rate in the world. You're still gonna be a very poor business, you know? So And then the other thing is the contribution margin. You know, you can bring in all this revenue, you can keep all your customers, but movie pass, right? The whole issue with movie pass was It basically burned money. You know, it it sold customers a product or a certain price. And every time those customers did even a little bit of activity, they lost money.
Kurt Elster
And so until they went out of Until they the whole thing fell apart. It worked out pretty well for the customers.
Daniel McCarthy
Yeah, they went from um uh it was something like 10,000 subscribers to a million subscribers in a really short amount of time. So if you think about it, they were hitting their ball out of the park on like every single dimension. Customer acquisition, their cack was probably nothing. Uh the retention was probably great. Their monetization was fine, but their contribution margin was negative. Was negative. Yeah. So yeah, that one measure that that one measure just screwed 'em.
Kurt Elster
Um it's funny We're essentially talking about like trends in and favorability uh as it relates to business KPIs. I mean like pretty esoteric stuff, but financially important. I think a good example of that that I failed to mention was contribution margin. Like suddenly we're seeing education around contribution margin and brands adopting contribution margin as this this North Star for advertising campaigns for exactly the reasons you laid out. Um and so I think it that's the reason to pay attention to these things is when they give you insight into your business, that means you have confidence. And yeah, everybody tells me, I will I want to make data-driven decisions. Yeah, okay. Um, in this case, you know, when you have that that quantifiable info and know what it means and know what to do with it, there you go. Now those are your data-driven decisions, can now be made.
Daniel McCarthy
And see it and it does I I'll be the first to say it comes to the expense of complexity. So suddenly now you're managing to ten measures and uh not two or whatever it might be, but I would argue that you have to. So, you know, otherwise, uh, I think you you could have these blind spots where you think that you're doing well when in fact you are either not doing well or you could be doing better. So yeah, yeah, again, another measure like that is the you know gross profit or contribution profit payback period. Yeah, how many months is it till you get your customer acquisition dollars back and hey, yeah, that's got the the CAC in there, it's got the contribution profit in there. That probably would not have looked good for the movie passed. So that would have, you know, shown some some red warning signs, but um but even that could be gained. if that was the only measure that you were focusing on. You could just kind of start skimping big time on CAC and maybe you should be spending a little more on CAC. Yeah, the customers you're bringing in are are very good. So Yeah, so sorry. Rant rant over.
Kurt Elster
No, I I appreciate it. Uh okay, give me the takeaway. What's the what should people They end this episode. What should they go do? Give me one actionable step.
Daniel McCarthy
Uh look to your transaction log. There's a wealth of data in the transaction log. And one of the key drivers that you want to be very mindful of is how valuable are your customers. And the key driver of that is how long are your customers staying with you. So you want to track that. Don't look to trends in overall churn rate as being indicative of much of anything. What you want to look at is you want to look at your retention curves, which are equivalent to looking at your cohort specific churn measures. And track those across all of your cohorts. And that's going to give you a much better sense of how you're doing along the retention dimension. So you look to those. Uh obviously if you're watching this episode and you have the show notes available to you, you can start running some of these models to be able to project out. what future retention is going to be. Yeah, so you can also look at measures like expected lifetime over time. So look to those. Um I'd say the other major takeaway from from the end of uh the end of this episode is This will give you a clean read into retention, and that's really important, but you want to then insert that into the raw context, you know, where you're again looking at that alongside CAC contribution margin. um and revenue per user. And uh if you kind of keep track of all those things, then you'll have a much more uh holistic sense of of how you're doing from a customer value perspective.
Kurt Elster
I love it. Where and you've got you've got tutorials and examples that we're gonna include in the show notes um and an Excel spreadsheet. Ooh, I love getting other people's spreadsheets. I've discovered Probably in the last four years, and I've been doing this a little over ten years, in the last four years is where I'm like, being a spreadsheet jockey, that is the value. That is where you can really uh make a difference and do some impactful things. So exciting to get a a fancy Excel spreadsheet from you. I'm in no way kidding about that. Where can we go uh to learn more about you. Like you're you're a professor, but you've got other things going on as well.
Daniel McCarthy
Yeah, so we yeah, we do kind of enterprise grade customer value modeling, uh both in subscription and non-subscription settings for my company, Theta. So thhetaclb. com is our website. Um so yeah, if you want to have uh yeah, if you want to kind of bring in the pros, um Yeah, we'd love to speak with you. I think there can also be value to having a third party perspective, you know, because you know, we have no conflict of interest. We just want to get to the truth. And so, you know, we'll just kind of tell you what we see. Um another company I would mention is uh Earnest Analytics. You know, that they offered the competitive benchmarking type data that could be used to fuel that type of analysis where you're looking not only at your own company, but uh some of your peers. We've got a a benchmarking product that we're releasing with them. So uh I think that that would be kind of another another potential resource. Uh on LinkedIn and Twitter. I think that's actually how we f had first connected was on it might have been on Twitter.
Kurt Elster
Um I think Andy Bedell.
Daniel McCarthy
No, I thought Andy Bedell introduced us. Might have been. Yeah, it's been a while. Yeah, part of the context behind that is I'm constantly talking about this sort of stuff on social media. So you know please do connect with me and yeah, happy to share my my LinkedIn and and and my Twitter handle now with you in the in the show notes too.
Kurt Elster
Fantastic. Uh yeah, I'll put all that stuff in there. Um okay, uh Dan, thank you so much. I gotta go be a spreadsheet jockey now.
Daniel McCarthy
Well, you're here. That's uh that's that's that's where it's at.
Kurt Elster
Crowdfunding campaigns are great. You can add social proof and urgency to your product pre-orders while reducing risk of failure. But with traditional crowdfunding platforms, you're paying high fees and giving away control, all while your campaign is lost in a sea of similar offers. It can be frustrating. That's why we built Crowdfunder. The Shopify app that turns your Shopify product pages into your own independent crowdfunding campaigns. We originally created Crowdfunder for our private clients. And it was so successful, we turned it into an app that anyone can use. Today, merchants using Crowdfunders have raised millions collectively. And With Crowdfunder, you'll enjoy real-time tracking, full campaign control, and direct customer engagement. And it's part of the Built for Shopify program, so you know it's easy to use. So say goodbye to high fees and hello to successful store-based crowdfunding. Start your free trial and transform your Shopify store into a pre-order powerhouse today. Search Crowdfunder in the Shopify App Store to get started.