Advanced Customer Acquisition Cost (CAC)

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

And I'm Growth, also known as Ray Reich, the founder and CEO of Benchmarkit, formerly RevOps Squared.

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

And together we are the Metrics Brothers.

Today's episode is the 6th in our recently launched podcast, SaaS Talk. In the last episode, we talked about the CAC ratio. What's on tap for today, Ray?

Well, Dave, as we were wrapping up last week, it sounded like we needed to dig a little bit deeper and dive into advanced topics on CAC. with an eye towards how to correctly interpret and use it. What do you think of that?

Yeah, I think it's a good topic. You know, last time we talked a lot about kind of calculating it and the basics. I think there is another layer that I'm excited to explore today in how to interpret it. What does it mean? Some of the nuances. Okay.

But right before we dive into that detail, let's go and get a quick word from our presenting sponsor, Maxio. 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-Reich 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 at Kellogg and Ray on LinkedIn at Ray Reich. Thank you so much for listening to the Metrics Brothers podcast.

Thank you.

So Dave, since a lot of people, maybe they don't remember last week's episode, or God forbid, they actually haven't listened to it yet, can you summarize a little bit about what we're not talking about today because we covered it in the last episode?

Sure, Ray. Good idea. So the first thing we talked about— well, first, CAC means customer acquisition cost. We talked about the difference between customer acquisition cost, which is kind of a per-customer metric, and And the CAC ratio, which is a per unit of ARR metric. The second thing we talked about was how to calculate the CAC ratio, which is prior quarter sales and marketing expense divided by this quarter new ARR. And we talked about the variations on that calculation. And then we hit what I consider to be the top-level questions, you know, to include or not to include customer success, to phase shift by a quarter, by nothing, or by more than a quarter, and by your sales cycle length, and to what degree. We talked about acceleration of commissions expense, which we both agreed should be done because expense by default means GAAP, and we think it should be GAAP expenses plus sales commissions accelerated into the calculation. We talked about new versus expansion CAC, i.e., segmenting the CAC for new customers versus existing customers, which works well the more separated your sales model is. We talked about line of business segment level CAC, like a CAC for your SMB line of business versus your enterprise. We talked about whether the CAC should be fully loaded or not. Do you include the rent for the sales team in the calculation? We talked about new versus net new. And just said basically, hey, beware that public companies typically will do it off implied ARR, which is based on subscription revenue, which is in the end net new. And then finally, we talked about land versus expand, what gets put into CAC versus what goes into expansion, and particularly the kind of NRR ratio. What's new ARR versus expansion ARR with an eye towards consistency? And you took us down memory lane to talk about VSOE, vendor-specific objective evidence. How'd I do in my summary, Ray?

I don't know. I don't have my headphones on. I don't know if I was listening very well. No, honestly, there was so much in there, Dave, that, you know, I really encourage people to go listen to episode 5 if they didn't have the chance to, because we spent 20 minutes talking about all that. But one of the things we just touched upon was, you know, CAC, especially the CAC ratio, is really not a great standalone metric. It's usually in context like It's a function of growth. What do you think about using CAC ratio in combination with growth rate?

So I think it's absolutely essential. Look, probably most benchmarks are kind of cut by company size. And for a lot of them, it actually works. Company size is not a bad way to look at— I don't know, ARR growth, for example, is typically a function of size. But the CAC ratio is really a function of growth, i.e., the faster you're growing, the more ramping resources you're carrying, right? Kind of the faster you're growing and the longer your ramp times, I— and the longer your sales cycle, the more time it is between when you hire a salesperson and when they close a deal. And in some ways, you can look at every SaaS company as kind of a steady-state sales operation plus a ramping sales operation, and the CAC ratio definitionally includes both. It includes all sales and all marketing costs, whether or not they're actually productive. And look, if you're— to take the extreme cases, if you're a flat company, you're not growing sales at all. So say your sales and marketing expense is fixed, everything is fixed, you're going to be carrying effectively zero unproductive resource, right? Whereas if you're a 100% growth company or a 200% growth company and you have a long sales cycle, you're going to be carrying a ton. of unproductive resource. So I think this makes CAC benchmarking inherently flawed or inherently dangerous because in the end, the CAC ratio is a function of growth.

Well, first of all, I'm so glad you talked about being able to look at CAC ratio in context of growth rate. We just published the 2023 SaaS Benchmarks Report. And we actually have that data, but we didn't include it in the report. So what I'll do, Dave, is I'll make sure we put out with SAS Talk exactly what those correlations look like by band of growth rates and the associated CAC ratios. That's a great idea. But you mentioned something really important because most of our listening audience aren't at 0% steady state. Let's hope they're growing. And you brought up an advanced topic people really need to better understand. That is, if I have my fully ramped reps, let's say I have 10 of those. But I just added 3 more and it's a 6-month ramp. How do you actually calculate CAC ratios so it's meaningful, Dave?

Yeah, so, so one of the most painful conversations in any board meeting goes something like this. How many reps do you have? 15. How many are fully ramped? 12, 7. Well, there's 2 guys on the half ramp and 3 people on a— wait a minute, how many reps do you have? And it can take 15 minutes of a board meeting like How many reps do you have? And it should be a simple question. And the answer is ramping potentially makes it hard to answer because everyone knows how many fully ramped reps there are. That's pretty easy. But you might say, well, 10 of the 20 are fully ramped, but that's, that's not enough information to analyze things, right? So, so what we need is some way that— look, there's, there's 2 ways to solve this problem if we're trying to understand sales productivity. We could say there are 10 reps who are fully ramped. They sold this much ARR, and therefore we take that ARR, i.e., ARR from ramped reps divided by the number of ramped reps, and that will give you a consistent notion of sales productivity, what I might call terminal sales productivity, right? Once people are ramped, and the way we calculate that is by counting how many are ramped and then counting how much they sold, right? Do not include the ARR from ramping reps. That's one of the more common asymmetric SaaS metrics where people will calculate sales productivity per ramped rep but include ARR from all reps. So that's one way to do it. The other way is what I call RREs. But before I do RREs, what are your thoughts on that first question?

Well, but you didn't really give me a CAC ratio. You gave me terminal sales productivity. That is how much ARR was closed per fully ramped rep. But now if you want to get a CAC ratio, you have to put the fully loaded cost of those fully ramped reps and then divide that by the ARR from those reps, correct?

And yes, you're correct, right? This is why I don't actually love this as a metric. I like the concept, right? But it's a tough metric to calculate because you need to do the loaded cost of all the ramped reps only, exclude the unramped ones, right? Arguably exclude the cost of unramped support resources like SEs and BDRs, right, if you really want to do it pure. And then you need to allocate marketing and say, ah, this much marketing went towards ramped reps this month went to unramped. So I think it's a kind of theoretically interesting but practically difficult metric, which is why I went to sales productivity, because that— we'll come back to the CAC ratio in a minute. But, but if you're trying to get a sense for how efficient is our sales model at steady state, one of the biggest drivers of the CAC ratio in your model, and just go build a driver-based model and play with this sale, is terminal sales productivity. How much does a ramp person sell in a year? That's a highly leveraged driver. So that's why I went that direction.

But I totally understand that. You also said something else last episode that I want to make sure we slow down and really define sales CAC ratio, because you mentioned sales CAC ratio, and I don't think a lot of people are using sales CAC ratio. So can you talk about the difference of sales CAC ratio versus CAC ratio?

Sure. Look, I like sales CAC. It's not good for benchmarking because no one calculates it. It is good for historical analysis. To say, hey, last year we spent this much on sales, the year before we spent that much on sales, here's how much ARR we got for it. And you can calculate its sibling, the marketing CAC ratio, right? Uh, basically, if the CAC ratio is sales expense plus marketing expense divided by new ARR, we could just simply segment that in 2 pieces. The other way to look at that, Ray, is actually my even preferred way, which is look at the sales to marketing expense ratio. So again, it's not a CAC ratio, but if I'm trying to watch How much bang do I get for a dollar of sales versus how much bang do I get for a dollar of marketing? I would tend to look at the sales expense divided by the marketing expense and watch that trend over time. The alternative way of doing the same thing is to look at the sales CAC and look at the marketing CAC.

Okay, so let's summarize and be very clear. So if we were to provide a recommendation to the audience, here's the best way to actually measure CAC ratio when you're ramping reps? What say you?

I would say first, obviously, you calculate the standard CAC like everybody else, which will include the cost of ramped and unramped. If you wanted to try and capture this difference, well, there's a couple of moving parts, right? Personally, I would— you're not gonna love this answer— I would either just use a sales CAC ratio, or I would look at— I'd probably and look at the sales to marketing expense ratio. Because I'm trying to make sure we're not getting out of balance. But the actual thing I would look at, I wouldn't actually try to solve that problem with the CAC ratio. I'd look at sales productivity because I know the sales productivity is a massive driver of the CAC ratio. And there I have a definite point of view. You can— and it's one of two answers. It's just that you say tomato, I say tomato. You can either look at ARR from ramp reps divided by the number of ramp reps, right? ARR per fully ramped rep, in effect. Or you can look at ARR per RRE, which is— a math nerd is the one I prefer. And let me just be clear on what I think an RRE is, a ramped rep equivalent. To give you a simple example, if I have 16 reps in total, 10 are fully ramped, 6 are 50% ramped. So if you say, Dave, how many reps are there? I'd say 16. If you said, Dave, how many RREs are there? I'd say 13, right? 0.5 times 6 equals 3. So I have the 10 fully ramped plus the 3 ramping equivalents. So I basically, I have the sales capacity of 13 ramped reps. I actually have 16, right? But I have the capacity of 13, and I'd look at that basically ARR per RRE, and that would be my sales productivity. I mean, I know you keep wanting to talk about CAC and I keep talking about sales productivity, and we can talk about both, but I'm a big believer that just because of its effect on the model, that sales productivity is a very important metric in When talking about the CAC ratio.

Here, here we are. We have a former CMO wanting to talk about the SAC sales productivity number. I want to talk about marketing CAC ratio. No, seriously, let's talk about another nuance though. And that is, you know, a lot of companies, especially enterprise companies, Dave, are selling multi-year agreements, 2-year or 3-year. And they're like, how does that impact how we should calculate the CAC ratio?

So great. So first thing on the prior point, I mean, yes, my calculations always tend to have marketing's back, and that's a fair observation. But the primary way I protect marketing here is actually just looking at the sales-to-marketing expense ratio. That is the way you cover marketing's back. Because simply put, over time in most companies, that goes up. You spend more in sales, spend per unit of marketing, you spend more in sales over time. Why is that? Because the sales leader is a better negotiator than the marketing leader, and they have more leverage. So as a CEO, if you do nothing, I guarantee you that ratio will go up. And that's basically saying you're putting the burden of productivity improvements in the CAC ratio on marketing, right? Just say you want to improve your CAC ratio. Whose problem is that? Does the sales CAC have to go down, or does the marketing CAC have to go down? And how much are you asking each team? How much are you putting that improvement on one team's back versus the other? That's the way I think about that one.

And that's why it's good to have both the sales CAC ratio and the marketing CAC ratio. But I know you wanted to go back to that because you want to make sure you had marketing's back, but I want to have the audience's back. So what about that multi-year booking structure? Does that impact how you would recommend calculating? Yeah.

I mean, so, so first let's do, let me just eliminate one that people sometimes do. Which is some people might do a book— what I call a bookings CAC ratio. Say you sign a 3-year deal for 100 units of ARR, so it's 300 units of TCV. Some people try to calculate a CAC ratio based on TCV, and, and I think both of us are rabidly opposed. Correct.

Please do not do that, please.

Yeah, and the reason is that the CAC ratio is supposed to be about the subscription base, and, and that, that person That customer, if there are 100 units per year, there are 100 units per year, right? If we want to know what they're worth over time, we can look at NRR or customer lifetime. But if we— the CAC ratio is how much did it cost to buy somebody for a year, right? So you don't get to multiply that 100 times 3, and that will be reflected in other metrics.

But let me ask you another advanced— this might even be a master's level. So we do a 3-year deal. And we're really good at building in price escalators. It's $100K ARR year 1, $200K ARR year 2, and $300K ARR year 3. So it's $600K ARR. Some people, I've seen them divide that by 3 and say it's a $200K ARR deal, and that's what I'm going to use in my CAC ratio calculation.

Yeah, that— so that's what I call price ramp deal, and we've locked in good escalators. I mean, frankly, as a CEO, I like that deal because it's giving me a high NRR, right? Because that would count. Well, well, it depends. It depends how we count it. But, but ostensibly, well, there's going to be a consistency issue here. But on the surface, it's a great NRR deal because the guy starts out as 100 units, goes to 200, goes to 300. It's all baked in. I have to basically kind of do nothing and I get a great NRR. That's the good news about that deal. Look, I'm not a CFO. You're not either. I think a lot of CFOs would look at that deal and say it's 200 units of ARR per year, particularly if the cash goes that way. Like most CFOs I know say, follow the cash. Like, that's great that we've structured it that way. Maybe even the usage goes that way. But if the cash goes 200, 200, 200, it's 200, 200, 200. In which case, And that was a subtle issue, by the way, because I thought the CFO would have gone to usage, but they don't. They— usage might be secondary or tertiary. First place they go is cash. So look, to me, I would call it a $100K deal and get a great NRR off it, and it would hurt my CAC, obviously, because I'm going to take whatever sales and marketing expenses, divide it by 100 instead of 200. I don't think you'll burn in hell if you divide by 200. The only place you burn in hell is if you divide by 200 and then use the escalation in your NRR cap. But what do you think?

So I've been in that situation. And because, you know, I always wanted to be a CFO, though I never became one.

I'm sorry.

I actually did. I know I did CAC ratios.

We all have dreams.

And I blew away my CFO. I said, here's— I have 2 CAC ratios for you. I've got one for all my using single year. And here's my CAC ratio if you include all our multi-year deals, because sometimes we had 2-year deals, sometimes we had 3-year deals. He's like, But you're kidding me. You did 2 CAC ratios. So I would say do both. Here's when I only count the first year and here's when I count the multi-years.

But you're not doing TCV, just to be clear, on the multi-year. You're doing an average across the duration of the contract, correct?

Average across the duration.

And are you going to go upstairs or downstairs? When you did NRR, tell me about how you calculated NRR in that context.

I wasn't responsible for that at the time.

Ray!

You know, I was just trying to make my CRO numbers look good, but, but Dave, that's really important for the audience, right? You've got to be consistent. What are the metrics you're trying to really highlight and promote to the board and investors? What do they want? Our investors wanted— they were okay with a little bit higher CAC because we were growing so fast, Dave, but they wanted to maximize and optimize NRR. So we used a CAC ratio limited to the actual first-year deal value, and NRR then kicked in second year and third year.

Yeah, to me, and again, there are presumably accounting rules that govern some of this on the revenue side, but CAC is an ARR metric, not a revenue metric. So the accountants may help us less than we think. But look, I believe it's fine as long as you're consistent. If you got the $100K, $200K, $300K deal, I think it's okay to say you bought 100 units of ARR in year 1 and then show a great expansion. I think it's— personally, I think it's also okay to say you bought 200 units of ARR, but then you don't get to count it as expansion. And that would be my worry, is that you'll do both. So I tend to default to you bought 100 units of ARR. Great. Your answer was to calculate both. Let me spin on you, Ray. I want to take your question and make it harder. It's no longer a contract expansion. Just say we have a bunch of data that says your average customer comes in at 100. And just, I'll make it a very simple example. It's unrealistic, but simple. Then like everybody becomes 200 in year 2, and 300 in year 3. So it's not contractual. But it is the observed behavior. When now the question is, how much ARR did you buy for your sales and marketing expense?

I'm a purist, Dave, 100. I don't like to get ahead of myself, even if it's consistent. And it goes the same way, by the way. And for the listening audience that has usage-based pricing, this is a very important part of this conversation. Even if I got observed ramp to commit to $100K, but everyone within the first year gets to an average of $200K, I don't calculate the CAC ratio on that $200K. I do it on what their minimum commitment was. And then at the end of the year, I will do a second CAC ratio, which I said is based upon the actual realized ARR from those usage-based pricing deals. So the initial commitment is one. The second one is based upon actual ramp incurred in the first 12 months.

I think that's a pretty reasonable approach, to be honest. And you've, you've taken us to the, to the area I want to talk about, which is consumption pricing. I guess the question I was going to ask you, and I no longer think it's a great question, but maybe you will. If I said those were months, not years, how does it change your answer? Or if I said they were quarters, not years, how does it change your answer? And I don't think it does, because you're basically saying— let's do the quarters case. You're saying, hey, I'd only look at the first quarter revenue, subscription revenue, and then in my second metric, I'd look at the whole year.

That is correct. That's what I do.

Yeah.

Even if it fully ramps in the— if it fully ramps in the first quarter, Dave, You could do it, but most deals close at the end of a quarter, then it's the second quarter, the next quarter that the ramp actually happens. I think it's too short of a timeframe, so I do it upon close in the quarter that it closed and then for 12 months.

And the only, the only case where I think it gets sticky with your formula is say it's an extremely fast sales cycle, like we close deals in less than a month and they can expand in the next month and then the next month. I think, I think You know, months are a lot like quarters are a lot like years unless they're not, right? And some of these metrics, we need to say, hey, wait a minute, if these deals are— most SaaS metrics to me are silently predicated on 6-month sales cycles, right? And deals that expand over years, not months or quarters.

Yeah. Well, I wonder how much we've confused all the listeners out there. Dave, is there anything else you think we should talk about with this kind of advanced CAC and CAC ratio discussion?

I have more faith in our listeners than you do, Ray. I think they're hanging tough. I think they're ahead of us. So yeah, we should talk about consumption pricing. And I'm gonna tell you my view on consumption pricing, which is pretty simple. And this is the way I start to approach consumption pricing with SaaS metrics in general. There's a body of work around SaaS metrics that Ray and I are very familiar with and talk about on this podcast. And that body of work is all based on ARR, right? Not all of it, but most SaaS metrics, ARR is popping up somewhere. It's kind of a fundamental unit. And the problem with consumption-based pricing is you don't actually have ARR, right? In a pure consumption-based model, you don't have it. So, and Snowflake, you know, I haven't read their stuff recently, but they were in the beginning, they were pretty unabashed that they were not a SaaS company. They'd say, we are not a SaaS company. You know, we don't try to do SaaS metrics on us. We're a usage-based company. So I think the answer though still is if you want to do SaaS metrics, You need to have a proxy for ARR. And the reason I like it is it lets you reuse all the existing work. We can go invent a new set of SaaS metrics for consumption-based companies, right? We could go do that. But I think it's far, far better to say, wait a minute, if I don't have ARR, what can work as a proxy for ARR so I can shove that proxy into all my formulas? And I'm going to give you a couple different proxies that I think people use. One is trailing spend, right? What did you spend last quarter? What did you spend last year? Probably my favorite. Another is for— is basically snapshots. So let's just say your ARR for last year was 12 times your December usage, which in my mind is a pretty aggressive way to do it. But some people take a snapshot, some people do averages, right? They'll say, oh, your ARR for last year is 4 times your Q4 monthly average. So those are the 3 proxies that I see. The other place people look at it is they look at the contract and they might look at floor value. They'll say, ah, well, your contractual minimum is 100 units per quarter, so I'm going to count that and anything else is overage. But I still need a proxy. I mean, I guess you could just do it off the contractual minimum, but you're going to really skew the metrics because you might be ignoring a material amount of overage on top that really should be counted. So I don't know, what are your thoughts? I'm answering the more general question of how do we map SaaS metrics? to a consumption-based world. But to me, in order to answer how do I calculate CAC ratio in a consumption-based world, I need to answer that question first. What do you think?

Well, the first thing I heard, you were talking about Snowflake, who highly promotes their 155% to 168% net revenue retention number. And they used to maybe unfairly kind of look at first year Revenue and then second-year revenue, and they would use it on a revenue basis, I believe, Dave, if I'm not mistaken.

It was a trailing— the way I would say it, Ray, is it's a trailing 24-month lookback, right? They look at actual spend last year, actual spend the year before, divide the, the, you know, last year over the year before.

On a GAAP revenue basis?

Correct, on a revenue basis. Yeah.

So that's the other thing you could do And I actually recommend to all my clients who have usage-based pricing is to calculate retention rates, both GRR and NRR, on GAAP revenue also. Say, this is my net revenue retention on a GAAP revenue basis year over year, period over period. But that's a different podcast, I think.

I think it is a different podcast. I have to weigh in on one thing. So I like what Snowflake does with the trailing spend. I think it's generally correct. I think the NRR metric is particularly vulnerable when done that way. And I would actually argue that Snowflake's NRR is not really comparable to typical SaaS companies. Because let me just remind you what NRR is. I stand at this moment in time, I look back 12 months, then I look at their forward commitment. So it's half forward-looking, half backward-looking. Snowflake's is 1 year backward-looking divided by 2 years backward-looking. It's actually very different Like if they have a decay in their expansion, I think it will show up slower. I'm not sure. I need to think about that. Well, they won't know because they don't have forward-looking commitments. All I can say is it's different. I'm not exactly sure if it skews the calc, but, but it's one looks back one year divided by 2 years ago. The other one looks at this year's commitment divided by last year's commitment.

And that is, I think they call it 2-year lookback. I think they even named it. 2-year lookback. So, well, we're gonna have to end this episode, and we didn't get to talk about how all these nuances around CAC and CAC ratio impact CAC payback period. You think we should do a whole episode on CAC payback period, Dave?

Yeah, we absolutely— I, I have called and will continue to call CAC payback period the most misunderstood SaaS metric, and we have a lot to talk about on that one. So yes, hopefully next episode we could do CAC payback period.

Okay, Dave. Well, with that, our 20 minutes are a little bit past and gone. To our listening audience, if you like Dave and I's conversation about SaaS metrics and our little bit of banter and a lot of deep details, it would mean the world to us to go ahead and subscribe to SaaS Talk with the Metrics Brothers on your favorite podcasting app. Go ahead and give us that 5-star rating. Most important, because Dave and I, this is really a show we wanted to do for the industry. And that means we want your input on what we should be talking about next. What metrics do you want us to talk about? And for CAC payback period, do you have any questions? Send those to sastalkpodcast@gmail.com, sastalkpodcast@gmail.com, and we'll include at least one of those questions in that podcast. Thank you, everyone. Thank you, Dave.