Net Customer Base Expansion

Live from Schenectady, New York, it's SaaS Talk with the Metrics Brothers, Growth and CAC.

And I'm Growth, better known as Ray Wright, founder and CEO of Benchmarkit.

And I'm CAC, better known as Dave Kellogg, EIR at Balderton Capital, independent consultant. And the author of Kellogg.

And together we are the Metrix Brothers.

And we go together like popcorn and butter.

Mm, back to food again with the weekly simile, huh there, Kek?

Weekly simile, huh? We used to call it the weekly metaphor. I'm glad you've tightened that up. It is in fact a simile, Ray. And yes, we are back to food. There's just so many wonderful combinations from rack of lamb and thyme. Is it rosemary or thyme? Rosemary and rack of lamb, ham and Swiss, red beans and rice. I could continue.

Yeah, but why popcorn and butter for this week?

Ah, too subtle for you? Because popcorn, my friend, expands.

Ah, kind of like what SaaS ARR is supposed to be doing. And you wanted to talk about net customer base expansion as a metric, right, Dave?

I did indeed, Ray. It's a nifty little metric, nothing mind-shattering, but I do find it useful in certain situations. So today let's talk about net customer base expansion. But before we do that, let's hear a word from our sponsor.

The Metrics Brothers is proudly sponsored by Benchmarkit, the B2B AI and software industry's largest, broadest, and most contextualized benchmarking index. Dave Kellogg and Ray Groth Wright have put the Metrics Brothers podcast on hiatus for the remainder of 2026. All episodes we have ever recorded are still available on your favorite podcasting app. You can reach Dave @Kellogg and Ray on LinkedIn @RayReich. Thank you so much for listening to the Metrics Brothers podcast.

Okay, so hopefully the listening audience has went to their microwave, got their popcorn, hopefully it's buttered, but let's start with the definition of net customer base expansion.

Sure. So this is not a complicated metric, Ray. As I mentioned, it's one I call a nifty metric just because it's kind of like You know, so much of metrics is trying to shine lights on things. And a lot of it is about which metrics you pick and most importantly, in which combination to shine light on certain problems. So net customer expansion is simply expansion ARR minus churn ARR. Now, to understand what I just said, you need to adopt the Kelloggian taxonomy of ARR. So let me just remind people. I like to think very hierarchically. So to me, kind of the math I start with is starting ARR plus new ARR minus churn ARR equals ending ARR. That's Dave's words. And then new ARR has 2 subtypes, from new customers and from existing customers, right? The latter often called expansion, sometimes in my opinion erroneously called upsell.

Yeah.

Because I think upsell is a type of expansion. By the way, from new customers, people call it new logo. I call it new biz sometimes, but those are the 2 types of new ARR. So when I say new ARR, I mean both. And then churn ARR accordingly has 2 types. People are usually less confused about this. From lost customers, i.e., well, what some people call churn ARR, which I find incredibly confusing. I call it lost ARR or churn ARR from lost customers if I make the table. And then I say from shrinking customers, also known as shrinkage.

Okay.

So, given that as a reminder, net customer expansion equals expansion ARR, i.e., new ARR from existing customers, minus churn ARR, which means ARR from both lost customers and shrinking customers.

Okay, get ready for this one, Dave. It's a food analogy. You are the combo king because you always combine new logo and expansion ARR together to get new and lost customer ARR and shrinking ARR to get churn. You know that I always like to talk about all 4 together, but I think there's another important aspect of expansion ARR that you initially teased out. I kind of look at expansion as having 2 primary components: upsells, which I define as selling the same product to either new users or new use cases, So you're upselling from what you're currently having the product used for. And the second part of expansion ARR to me is cross-sells, and that's selling a different product to the same customer. And that's what you're talking about here, is expansion ARR includes upsells and cross-sells, Dave?

So, so look, to me, I like to come at it top down. Anything you sold to an existing customer is expansion ARR. It could be a price increase, it could be incremental seats, it It could be an entirely new product sold to a different division, right? It could be a new product sold to the same division, new product sold to a different division. It could be an addition upgrade from silver to gold, right? It could be usage-based pricing fluctuation upward or increase in minimum commitment. So, I always like to come at it top down. So, to me, expansion ARR is basically anything that makes an existing customer's ARR go up. And I like where you're coming from, Ray. I've just, I must admit, I've never, I've never defined a taxonomy at that level. I do agree with you that upsell implies like silver to gold and that cross-sell at Host Analytics, the classic cross-sell was we sold you financial planning and then we're going to sell you consolidation, i.e., a different product. So I agree in essence there. I'm not sure where in your taxonomy you put price increases or addition upgrades as opposed to just, hey, they had 10 accounts, now they had 12, so we sold them 2 more seats.

Yeah, I put that in upsells typically, but, you know, we do actually discuss our episodes before we go live here on the air. So this one I'm going to throw a curveball at you because we didn't just discuss it, but you mentioned usage-based pricing. So if people use the product more this month and it grows, it's expansion ARR. But what happens if they lose it? Yes, is that negative expansion ARR or is that churn ARR?

So this is going to come back to your definitions of what ARR means for a usage-based company. I'll just pick one standard example, which it means trailing 12-month actual spend. So I will agree with you, Ray, that if you're going to count it as expansion, I'm just going to call this Newton's second law of ARR. If you're going to count it as expansion, it has to count as churn, right? It can't be one or the other. My other is always, if you're going to count it as a new customer, it has to count as churn when it dies. Right. So maybe the answer is don't count it as a new customer if you're running like a trial program, right? Because the killer is that people do trial programs, they want to count the 1,000 people who signed up for the trial, and then they don't want to count the 800 who churn. So I agree with you on the symmetry principle. But to me, basically, if I just took the trailing 12-month spend, yeah, if it goes up, it's expansion. If it goes down, it's churn.

Okay. Got it. That's a good clarification. But I think one of the things I love is you always provide examples. So do you have some examples we can walk through here?

Yeah, it's going to be a pretty simple example. I'm not going to hit every element of the table that Ray gets to see. But I just want to talk about why I use this metric and I want to differentiate it from some other metrics. So the first metric I'm going to differentiate it from is net new ARR, which is another useful metric in my mind. And net new ARR, right, is just new ARR minus churn ARR. Period. In the Kelloggian taxonomy, right? All new ARR minus all churn ARR is net new ARR. I view this primarily as a financial metric because I like to use it for like burn multiple, right? Like, okay, how much do I burn in cash to get $2 million of net new ARR? So if I burn 4 to get 2, then I would have a 2x burn multiple. So I don't view it as a particularly good Operating metric, just for the problem you and I love to talk about, Ray, which is what I call the French railroad crossing, right? One train can hide another. There's a sign at every French railroad crossing that says that. So whenever you have these net metrics, right, that you could have one train hiding another, you have a great new ARR and horrific churn ARR, and you may look at a mild churn problem, right? But you can't see it. So, but anyway, I would differentiate net customer base expansion first from net new ARR, which includes in the calculation New customer or new biz ARR. I would also differentiate it from NRR because a lot of people might look at an example and go, let's just take a customer where they had in Q2, they had 600 units of expansion ARR and 250 units of churn ARR. Some people would look at that and go, oh, that's net expansion. Well, the way I'd look at it is say that's net expansion of 350. Then they divide the 350 by starting ARR for the period, say it was $10,000. And say lazy NRR is therefore, or NRR, I call it lazy NRR, would be 103.5% or just rounding up to 104%. So a lot of people calculate NRR this way, and I call that lazy NRR. And there's a Kellogg blog post on lazy NRR. But Ray, let me just punt it back to you so I don't talk too much. What's wrong with calculating NRR that way?

Well, NRR, and the SaaS Metrics Board agrees with your Kellogg blog post, is you should use a cohort-based, which means time-based calculation model. And what we like to do is look at a customer base from today against 12 months ago. So you're really looking at that exact same customer base from 12 months ago and how much has it grown over the last 12 months. So that's the main reason, because in formula-based NRR, you might miss some things.

Correct. Yeah. So, so Ray calls it cohort-based versus formula-based. I call it real NRR versus lazy NRR because real NRR to me, I don't need to say the word real. NRR to me is defined as a year-over-year metric that's based on a cohort of customers. Grab a group of people who were customers a year ago. How much are they spending? How much are they spending today? Divide the second by the first and you get NRR. So in other words, though, it takes you a year to figure out if you've got an NRR problem, in essence, right? If you fly off a cliff in one quarter, Well, I guess you'll know it, but it is kind of a lagged metric. It's kind of a lag metric. I'm not sure it takes a year to find out. But in any case, it's a lagged metric. It's not a quarter-over-quarter metric, right? It's a quarter-over-prior-year-quarter metric. And this, a lot of people, if you look at spreadsheets, will just call this row NRR.

Yeah.

And they'll divide it. And there's 2 differences to me with net customer base expansion. One, I don't call it NRR because it's not. And 2, I don't divide it by starting ARR to turn it into a rate. I just leave it out there as a number. So, if expansion is 600 units and churn is 250, I have net customer base expansion of 350. Just to go conversely, if expansion's only 400 units and I have 900 units of churn, the net customer base expansion is -500. Okay.

So, I get the calculation. But what I guess I— and by the way, one clarification point is this is really Good for quarter-over-quarter customer base expansion, where NRR is year-over-year. But my question is, why use this versus just doing year-over-year NRR every quarter? So I look at Q2 of '25 versus Q2 of '24, and I still know what my expansion is, and I go below that and say, okay, how much was my expansion ARR? How much was my churn ARR? And I still know the same thing, don't I, Dave?

Uh, I don't think so. And I'm not 100% sure here, Ray, but I feel like basically you're comparing to something different. In Q2, you're comparing to Q2 a year ago. I would rather compare to Q1. And then NRR just doesn't do that. By the way, not I would rather. There are times when, comma, there are times when I want to compare to the prior quarter. And let's come back to where I think the use case for this metric is. I've only used it with one company. Because in general, you know, NRR is good enough and you just look at churn ARR and you get a sense for it. But I had one company that was kind of fluctuating around. Some quarters, the customer base would shrink. Sometimes it would expand. And it was never really very consistent. So, we decided to put a light on that and say, you know what? What happened last quarter to the customer base? And by the way, it was kind of a monthly SaaS company. So, this was meaningful. I didn't need to look year over year. Like, you could churn from month to month. Right? This is a month-to-month SaaS company. So, we said, you know what? Every month, I want to look at, or every quarter in this case. But again, with monthly cadence, you're looking at this metric quarterly, right? With NRR, if you're doing a monthly NRR, you're still comparing to year ago February, right? Whereas in February with this, you're comparing to January. So, I like to use this when there's kind of a problem with the customer base. That problem would be it's shrinking. And either because it's a monthly SaaS company or a velocity SaaS company, I don't want to do a year-over-year comparison. I want to do either a month-over-month month or a quarter over quarter comparison. And I just feel like this is a better— this is not a super rational metric, right? It's just called a nifty metric because what's nifty about it? I could just see in a dollar figure what happened to the customer base. Oh, it went up by $350,000 or, oh, it went down by $500,000. And sometimes all these rates and percentages, you kind of lose that visceral touch.

Dave, you know what the light bulb that just went off for me was? And that's why I love having these conversations. If I'm a heavily lean upon usage-based pricing, this would be really good to see because I'm looking at every quarter, if not every month, how's my existing customer base usage of the product and the associated ARR growing?

I love it.

Absolutely. And by the way, the company we use this with is that— it's kind of user-based, usage-based, which I think we talked about last time. But basically, if you need 5 seats in January, use 5. If you use 8 in February, then you can use 8. And if you need 5 again in March, you go back to 5, right? So it's kind of user-based, usage-based, but it has that frequent fluctuation in the customer base. So I like this metric. The other thing I like about it, Ray, again, this is not mind-blowing, but this is just a different way to shine light. And that's what we do here with SaaS metrics, right? We're trying to shine lights on problems. And for this scenario, I think it's a great metric. The other thing I like about it, so one, we talked about kind of the velocity or month-to-month or the usage-based or the kind of Monthly fluctuating, pick your adjective, SaaS company. The other one is just, I always like visceral metrics where I can feel them. And my favorite thing to do with either churn ARR or net customer base expansion is to multiply it by the CAC. So if you tell me the customer base shrunk by 500 units last quarter, how big a problem is that for the business? The answer is CAC ratio times that negative 500. So, if my CAC is 3, say I have a high CAC, that's a $1.5 million problem, i.e., it's going to cost me $1.5 million simply to refill that 500 units of ARR that leaked out.

You know, that is such a good point. And with all the customers we work with, less than 10% actually take, once again, different metric, but 1 minus GRR. So let's say that's 14% times how much revenue that really is, and how much is it going to cost me to get back to equal to 100% NRR? Most people don't do that. They don't have an idea, Dave. I'm amazed at that. So a little tangential, I know.

So, okay, so let's come back and say, what does this actually look like? So this looks a lot like lazy NRR, right? Because lazy NRR, the calculation is starting ARR Plus net customer base expansion divided by starting ARR, right? That is lazy NRR. You do it kind of quarter over quarter typically, and you take this metric that I'm talking about, net customer base change or net customer base expansion, and you simply add it to starting ARR, then divide by starting ARR. So we've talked already about how NRR is supposed to be a cohort-based measure. So that's— it's not real NRR. I call it lazy NRR. But basically, lazy NRR turns net customer base expansion into a percentage, right, based on starting ARR. So, if the net customer base expansion is positive, then lazy NRR will be greater than 100. If net customer base expansion is negative, it'll be less than 100.

Okay. Hey, Dave, since we talked about this preparing for this episode and I saw your LinkedIn post, I'd like to take us down just a little bit different path, but it's related. The importance of Using the word cohort versus segmentation, can you just explain a little bit about why a cohort-based calculation is so important here?

Yeah, so what I did— Ray's referring to a LinkedIn post I did the other day where I basically said a lot I increasingly hear, and I do increasingly hear this, people use the word cohort when they mean the word segment. And I think frankly they do this to sound cool. I mean, I got to be honest with you, Ray, if I have to guess why people do it, It's because cohort kind of sounds cooler. But cohort analysis actually has a definition that's different than segment analysis. Cohorts to me always invariably involve time. And typically, like a cohort example might be, let's get all the people who turned into stage 2 opportunities in January. So that's the definition of the cohort. There's a time element. There's a thing and a time. So something happened. They became an S2 opti. And that happened when? In January. And now let's watch them month over month. What happens to them? Right? And cohort-based metrics are great. And cohort analysis basically looks like you might do stage 1 opties by quarter going down the side of a table, and then ARR resulting, or deals won if you're doing count-based, deals won by quarter across. Or you could do down in months and across on months. Doesn't matter. Depends on the cadence of your business. But that's what a cohort analysis looks like. It's one of those kind of hanging triangle tables, right? Where there's this kind of hanging Right triangle, right? Where you feel like you have the upper half of a right triangle filled in and the bottom half isn't filled in. That's what cohort analyses look like. And segment analysis is different. I might just say, hey, what's our win rate in retail versus manufacturing? And that would be kind of slicing and dicing, right? So cohorts to me invariably involve time. And the thing I like about NRR is that it uses cohorts. And to me, cohorts in that particular case get rid of a lot of the noise. Because look, companies I've run, Ray, they make churn so— they're doing everything they can not to call something churn. So I got tired of fighting about what happened along the way and was it churn or was it expansion or was it reclaimed ARR? I haven't heard that term. I've heard recycled ARR. Like, I've heard all this crazy shit that people make up to avoid saying that, you know, the C-word, churn. And I just got— I like cohorts because in this case it's snapshot-based, right? Just Get all those people. What were they worth a year ago? What are they worth right now? I have 2 numbers to divide one by the other. So part of the reason I was drawn to NRR was I was really tired of the swamp in between then and now. But so I liked the notion of cohorts. I love, look, the best, even better than NRR is that hanging triangle, the cohort-based chart. You can just look at what's happening over time. So that's where it came from. And I can't remember where you wanted me to go with it, but that's why I posted.

No, you explained the difference between cohort and segment, and that's what I wanted to do. But I do have something I like to add, because even with a cohort-based metric like NRR, it's really important to me. And this is something I've always done, whether it's an operator or consulting to companies today, and that is look at that cohort metric by segment. So I want to know what my NRR or my net customer base expansion is for my enterprise customers versus my mid-market, or for product A versus product B. As soon as you get to 20 million or higher, I think you have to do segment-based analysis also.

Correct. Another good reason not to confuse cohorts and segments, because you might be doing them at the same time and it could get very confusing. Yeah, so net-net, cohorts means time-based. And the other thing I'd say, Ray, because Ray and I talked about an interesting example with NRR, which is if you're doing kind of a lifetime history cohort analysis, you actually go all the way back to when you signed your first customer, right, in the kind of first column of that table. But the issue is a lot of times now, and certainly I believe even in your standards board, Ray, you don't look at the cohort. The cohort's not defined by when you signed. It's defined by when you were a customer, right? So get all the customers who were a customer a year ago, regardless of when they signed, whether it was that quarter or a year ago that quarter or 5 years ago that quarter, grab that cohort. That is a valid definition of cohort and then watch them over time. But The better cohort analysis would actually go all the way back and break them into the— this was the 2014 cohort of customers, the 2015, the 2016, etc.

And by the way, the reason I love to do that, because I do ask people to do it by time, is if I try to improve my onboarding process, if I did product enhancements, so my 2015 cohort performs better than my 2013 because I changed the onboarding. It's really important to see the effect on NRR based upon that time-based cohort.

Yeah, it's also important to see how they die off and warning signs. There's a lot of value in a full cohort analysis. I would just say, Ray, look, all of this analysis is predicated on kind of the wannabe miracle of SaaS, which is that the customer base automagically grows, right? That is kind of the unstated premise of SaaS, that these companies are worth so much money Because not only because this is kind of an annuity, right? It endures, but it's actually better than an annuity because an annuity is fixed payments over time. This is even better. What's better than an annuity? Growing payments over time. And that's what NRR is trying to track. And that's in some ways what net customer base expansion is trying to track because your company's worth more if you can look an investor in the face and just say, hey, kind of look, no hands, and the customer base is growing. Right.

Okay, so let's, let's wrap up here. But unlike CAC ratio, which I do think changed my life when I started using it, is net customer base expansion really going to change someone's life in the SaaS company?

Look, Ray, I don't think so. I, I use it. I mean, it's a metric you should have in your tool bag, and you should whip it out when you need it. And when you need it will be a function of the company's situation, like if the base is fluctuating and you want to keep an eye on it. And when else do you need it? You need it if the base is fluctuating. Oh, and used in conjunction with other metrics. I need to see the full dashboard to know if you should put it on yours, right? But that would be an interesting topic one day, Ray, to talk about how you build the dashboard. Because to me, it's never about one note. It's about the chord, right? Like, what are the 5 metrics we're going to put on this dashboard? Yep.

And I would say the other time that I think it's really valuable is when you've got a rapidly growing usage base customer or you have high volatility month over month, I think it's a really valuable way to look at month over month or quarter over quarter trends.

Awesome. Well, thanks a lot, Ray.

Okay, I like this one, net customer base expansion. I wouldn't have done it without you, Dave.

We got matching shirts today too, Ray. We'll have to see if we can continue that tradition.

Welcome to the team.

The Metrics Brothers AI, SaaS, and More is a production of the Metrics Brothers Growth and CAC And a member of the Benchmarket Podcast Network. By accessing this podcast, you acknowledge that the Metrics Brothers make no warranty, guarantee, or representation as to the accuracy or sufficiency of the information provided or the humor content of the jokes told. Ahem.

Ray?

Information, opinions, and recommendations presented are, according to our spouses, probably wrong and provided for information purposes only. This podcast should not be considered professional or, for that matter, unprofessional advice. We disclaim any and all liability for any direct, indirect, indirect, misdirect, incidental, special, ordinary, consequential, inconsequential, or other damages arising out of any use of, or God help you, reliance upon the information presented here. Ray Grothreich is based in New York City and available on X @rayreich. Dave Kellogg is based in Silicon Valley and available on X @kelblog. Schedectady, which is French for unspellable, is not our actual production location. Thanks for listening.