Ep 13

Your SaaS Metrics Are Lying To You

Which SaaS Metrics really matter? What are good Startup Metrics or Startup KPIs? Corey Haines, Founder of Swipefiles and Marketing Lead at SavvyCal shares his hot takes on SaaS metrics that don't really matter and what to focus on instead. Listen in as Corey shares his experiences from working at Baremetrics and consulting other startups with their SaaS growth metrics.

Presented by
Ryan Hatch
Head of Product Strategy & Innovation
Robert Kaminski
Senior Product Strategist
Corey Haines
Founder of Swipefiles and Marketing Lead at SavvyCal
Corey Haines
Founder of Swipefiles and Marketing Lead at SavvyCal
Andrew Verboncouer
Partner & CEO

Ryan Hatch: Welcome to exploring product everyone. Uh, today we're excited to have Cory Haines on Corey is a founder of swipe files and the marketing lead at savvy counts. So Corey. Welcome to exploring products. 

Corey Haines: Yeah. Thanks for having me really excited. This is a topic I've been noodling on for a long time. And, um, one of these days I'll write a tweet thread.

So I'm hoping that this conversation will be sort of like the open, extra relation of all my thoughts related to this. Cause we'll have a lot to unpack. 

Rob Kaminski: That's going to be our whole marketing platform for getting people on the show. It's just come battle, test your content. We'll get you nice and sharp before you take it live under your own name. Right? 

Corey Haines: I would love that. Yeah. Perfect. Great. 

Ryan Hatch: Corey, we're super excited to have you and SAS metrics are a big deal. Everyone's talking about different ways to look at it. There's tons of tools out there, but really kind of excited to have you on to talk about how your SAS gross metrics are aligned to you and really interested in why you're passionate about that topic and how that came to be.

And how did you come to this realization that there's something wrong with most people's SAS metrics and how people are. 

Corey Haines: Yeah. Well, I think, I think the whole backstory here is that, uh, previous to what I'm doing now, I was the head of growth at metrics, which is a SAS metrics platform for, uh, first startups.

And for anyone with MRR, really, I was looking at subscription-based metrics and analytics. So I got to see all sorts of metrics, all sorts of different startups and sort of their mix. Uh, I think over the almost two years I was there, I talked to about 10 to 20 founders and operators a week and got to sort of like dig into their barimetrics instance and talk about their metrics.

And there was, everyone had always had questions around, Hey, is this good? Is this bad? What does this mean? What do I do with this? What does everyone else's look like? What are they doing? How do you improve this? Uh, where should this be? What's the benchmark. What does good look like? And so been in this for quite awhile only recently though.

So the other part of this story is that after I left bare metrics, I did a big consulting. And I guess sort of like coaching slash mentoring for a while. So at one point I was working with 12 different startups, uh, basically on like marketing growth capacity, um, to help them and would talk with them once a week or once every other week.

And a lot of these same questions came up around metrics and how to improve things. And how do we drive growth? What are the key levers and kind of KPIs that we're going to be. Uh, we're going to be focusing on. And it was through that experience afterwards, really being able to be in the weeds that started to kind of like piece some thoughts together on like, well, let's start this thing.

This thing over here is like kind of useless or like this thing is actually lying to you. You needed to actually like dig into this more. So that's kind of like the backstory of why this matters and, uh, why I've kind of culminated these thoughts around why some of your SAS metrics can be lying to you.

What, what to do about it, how to dig into. Um, which ones are telling the truth, which ones aren't. And also, cause there's a lot of like kind of old advice out there these days, like SAS used to be very new and nascent. Um, but now it's like a pretty mature industry and it's well-known, but we still kind of use some of the old, uh, basic ways of thinking in some ways.

Um, and or it's just like, it's kind of that, that game of telephone right. Where it's like, oh yeah, You know, someone tells you that you should be at this metric, or you should be focusing on this thing, but they don't tell you why. So then that person. Pass onto the next person pass onto the next person.

And then pretty soon it's like, why do we really measure that thing? Or like, what does this actually mean? And pretty soon it's kind of taken on a life of its own and it's lost its real substance. So anyways, that's kind of the backstory 

Rob Kaminski: peer metrics. It's interesting, like you said, Thorough thought and insights around what they should be and could be came after bare metrics.

I'm curious, what was bare metrics approach? Was it more of, you could do anything you want and like, did you guys really not have a great story for these are the absolute best metrics and these are the ones you should avoid. Um, so know anyone watching, who was maybe consulted by Corey years ago at Baremetrics, you may not have got the best advice.

And what did that look like for you? 

Corey Haines: That I tried to be as opinionated as I could be, but what does that for your metrics? It's very, um, you know, I just want to be. Uh, fair. I can't really divulge a lot about other people and sort of like, what does good look like? And it's hard to really be specific. So it was afterwards that I was like, no, no, no.

Okay. Now that I've left, you know, I don't really have like any strings attached to what I can, and can't say about what I think about certain metrics. And so I can be a little bit more of a loose cannon. Um, but also. After hearing the same thing over and over again, it started to, you know, piece those thoughts together around, you know, what, ignore this thing over here.

Or actually we should really dig into this part. Whereas before I'm trying to be a little bit more consultative about how to use the prototype, to get the most out of the product. Not necessarily like being very cut and dry around what actually matters at the end of the day, very metrics I would say measured most of, you know, probably like 80% of the core metrics that you want to be.

Uh, do you want to be measuring on a given day, week, month or a year? There's another 20% that we're missing. And I would say out of those 80%, really only about 20% of that is what I really think is important. And what is interesting. Um, and that's mostly kind of what I'll be talking about. That's related to.

The main metrics that are lying to you in your instance of what you're likely looking at, uh, like a growth rate, for example, or like turn lifetime value. I was writing per user, like those, the types of metrics that are going to be top of mind for everyone. And that there's more under the surface to, 

Ryan Hatch: yeah.

I remember when we were looking at, um, you know, different platforms and then like, you know, what platforms do you pull it, pull it in what time do I go? A fully integrated approach, one tool that can. You everything like more active campaign or more, more holistic, or do you piece together your own segment?

IO data pipeline funnels. And you're, you're going, you're like best of breed, but you have to stitch all these integrations together and it's. I know there's there's I'm sure you've got lots of questions on, Hey, what other tools am I missing? Well, I don't, I measuring what's important. What's the right time for the size of my startup, you know, based on the phase I'm at and peer stuff.

I'm curious like how your conversations, like, it feels like you were in tons of conversations all the time. Did you already have these realizations when you were at bare metrics of what was, what should or shouldn't be. No, you already, you said you were already opinionated, but I'm wondering, like, did you have, did you have a different perspective on things when you're able to really look at the whole picture from a business standpoint, as opposed to, from a product pushing a product standpoint?

Like, how did your perspective change? And then when was the wake up? When was the like aha moment for you? 

Corey Haines: Um, I think a lot of it has to do with what people are actually. Looking to get out of their metrics when you really kind of dig in. So I get questions all the time around, you know, you strike up a conversation and people are asking you how to make the most of my account.

What's good. What's bad. How should I improve this? What do I do? But then if I, if I really kind of dug in and tried to understand. Why are you asking that question? Or like, what is it about this metric in particular that, that peaks your interest and then you dig any ask why, you know, five or six times, and they really get to the root of it.

And then he realized that maybe they're looking at the wrong metric altogether. So I'll give you an example here. Uh, here's my, my F my first hot take on I'll come out swinging here is that a lifetime value is a completely useless metric and concept for SAS. Utterly useless, completely. This is one of the things that one of the first realizations for me, when I was at parametrics was people would ask all the time.

And by the way, this is one of the core metrics that Baremetrics tracks and, you know, uh ProfitWell and ChartMogul utterly useless because what's happening is that people are asking, Hey, what's, what's my left and value the, is this good? And this is bad. And I started answering the same question over and over and over again around, well, you want it to be around, you know, two to three times.

Uh, your first year is customers' revenue on average. Um, and then we started to dig into people are asking, well, how do you calculate lifetime value? And I'm going through. And, you know, it's really just a function of, uh, your average revenue per user divided by your churn rate. And it's essentially a function of how much can I expect to collect, uh, from a customer over on average, over the lifetime before they turn.

But it makes a lot of really risky assumption. For example, that every customer will turn over time, which is not true at all. In fact, if you go look through real accounts over time, you'll see that there is almost like layers. Like there isn't like the crust of the earth where early customers, the, the, the layer might get smaller and smaller over time, but there are still customers from 5, 6, 7, 10 years ago.

And then there's a second layer over 20 13, 20 14, 20 15, 20 16. And not all customers churn over time. And so what we found was that if you actually measure the lifetime value of certain customers, Uh, for example, been metrics, you know, we would have a lifetime value. A customer has been with us for a year and it'd be about a thousand dollars for example.

But then if you really look at our, our best customers and the customers who have been with us for, uh, for years, it was 50, 60, $70,000, nothing, you know, it's completely different, right? So if we're measuring lifetime value, we're talking apples and oranges compared to how long customer's been with us also, what is their plan?

These days and SAS, there are a lot of different price points that you can offer. You might have a freemium user, which their lifetime value is going to be zero effectively. You might have a starting plan at nine, 19 or $49 a month. You might have a sort of like an SMB plan around a couple of hundred dollars a month.

You might have a enterprise plan at a few thousand dollars a month or tens of thousand dollars a month. So if you're just, you know, then you can break up your lifetime value by segment, maybe for example, But then even then, what do we really want to know at the end of the day about what lifetime value tells us?

Well, when I actually asked them they actually poked and cried, people would say, well, we need to know lifetime value for acquisition purposes so that we can measure it against our customer acquisition costs because the startup culture and sort of the, the advice that's been passed along through the millennia is that you're a CAC to LTV.

Should. You know, three to one, four to one or five to one or something like that. Right. And that's how you know that your, uh, the acquiring customers for the right price and that you can scale effectively. That's really interesting because if you actually look at that and you say, okay, my lifetime value should be, let's just say it's a four to one.

So cash should be one fifth of the lifetime value of a customer. The thing is that lifetime value grows over time for each customer. So what point do you measure the lifetime value of a customer when you acquire them? Let's just say, for example, it costs a hundred dollars to you to make the math easy.

Hopefully I'm really bad at public math here, but things don't make sense. You can change it. Let's just say that it costs a hundred dollars to acquire. And your average lifetime value is $500. He might be thinking of. Okay, great. It's actually less than five to one. It's more like six to one. So we're, uh, it's going to be profitable on average, but that doesn't take into account things like.

So let's just say, half of your customers turn out within the first month, which is actually pretty common, uh, which a lot of people don't realize, well, then you effectively have to cut your lifetime value in half because you can only expect to collect a much lower amount from your customer base as a whole.

Right. If you're only collecting up to half of that or much less, right. Uh, even then. Again, when, when we take into account things like pricing plans, let's just say our starting plan is $9 a month, but your lifetime value, because you have cloud enterprise clients is skewed up, right. It's skewed to be larger.

So that $500 lifetime value. The customers that you're acquiring for a hundred dollars, their lifetime value is actually like a hundred dollars or $200. Now you're kind of screwed. You're lighting money on fire, essentially because you're using a metric like lifetime value to try to see if this is useful or not.

So anyways, I could talk more about lifetime value, but if you're really digging into what people are actually trying to get to, what they're trying to get to is some sort of measurement of how profitable it is to acquire a customer. On average. And in which case I started telling people, I started telling bear much as customers look don't use lifetime value as our measurement against CAC, use a payback period.

Use an average revenue per user. Because when you actually get down to it, you want to know how long is it going to take for me to recoup the costs of acquiring this customer with the revenue that they generate and so much better, it's actually much better to use it a metric like Azure and per user, and then use that as a function against CAC.

Right? So in this case, your CAC is a a hundred dollars. Your average revenue per user is $10 per month. On average, on that case, it's gonna take about 10 months. To recoup the costs, not accounting for churn. You can basically round up to like a year. Are you comfortable with that or not? Now you have a much more realistic.

Uh, example and scenario and equation to work with about your scaling economics of customer acquisition. Uh, so I'll pause there. Yeah, no, the 

Rob Kaminski: other thing, cause I had a question coming in, like my assumption, when you talked about for acquisition, it's all about positioning for funding, right? Like that the lifetime value is really about showing predictable growth.

So like give me money. But it blacks in your like, articulation of it, it lacks the time horizon, like, 

Corey Haines: okay. Like 

Rob Kaminski: five months, five years, like lifetime value, like the, what you talk about in terms of having a value per what would you call it? Revenue per user and knowing when it's paid back. It actually, if it's still for acquisition and investment, that makes a ton of sense because it's, it positions the conversation to be like, when might I have to raise, or if I raise, when can my investors start to see a return, come back.

Yeah, you just doesn't quite capture it because it's missing that variable 

Corey Haines: of time. Yeah. And in fact, I would say, um, that CAC to ARPU kind of a ratio or a factor is actually a really good way to create a marketing budget because one, you need a pile of cash to work with. If you have a. Uh, a payback period of let's just say three to six months, which is like pretty normal and acceptable from a marketing perspective to scale pretty effectively.

But that means that you need about a three to six months buffer to work with before the cost starts to recoup themselves, um, through. Right. And so you need to have, essentially, let's just say you're an attorney per user, or is it your customer acquisition cost is a hundred, a hundred dollars. And you want to start acquiring a hundred customers a month.

Again, trying to do public math here, make myself, make it easy for myself. Um, that means that you're going to be spending about $10,000 a month. And if your, if your payback period is about six months, that means you need to have about $60,000 in case. Set aside just for customer acquisition costs before that starts to sort of replenish itself in theory, right.

Again, you have to also account for things like churn, uh, accounts, growth, or accounts, um, uh, uh, expansion or contraction. Um, but that tells you, okay, now I can actually back into a number here. We need $60,000 set aside to work with. In order to scale this effectively, what a lot of companies do is they start just lighting money on fire and they start pouring money into Facebook ads and LinkedIn ads into sponsors.

And they don't know their payback period. And so they're waiting for the cost to be replenished. And then it's like, oh, we have to raise six months earlier than we thought. Right. Or we have to turn it off right now. Otherwise we're going to run out of cash because there are costs aren't being replenished as quickly as they thought regionally.

Rob Kaminski: Are there metrics that like, it's interesting when you say that rather than having a, a metric that kind of tells the whole picture or your cause you're getting into cash. Um, my assumption is that when they're just asking about CAC, are they separately monitoring cashflow, and then having to do this math in their heads of like, 

Corey Haines: my cashflow is 

Rob Kaminski: this and my CAC and LTV average.

Is this like, um, what have you 

Corey Haines: seen that? Yeah. If, if there are really finance savvy founder, maybe if they have a CFO, probably cause they're going to feel a little bit more accountable to. You know, the CFO, his job is to not run out of money, right. Or to basically tell the founder if they're going to run out of money.

So they don't want to be responsible for that. So they're going to be modeling that out. In fact, I actually created a, um, uh, a summit template it's called , it's basically like a Excel on a whiteboard kind of idea. Um, but those things are, are really useful to model it because you actually, again, you do want to know, okay, if this is the cash that I need, how quickly is this gonna be?

Uh, replenished, if things slightly changed, for example, if you want to count for like a 5% monthly churn rate, right. That means that you have to multiply everything by 1.05 in order to account for the, uh, the effect of churn in the future. Right. Um, but to answer your question. No, a lot of people aren't thinking about this at all, which is why I like to talk about it, because this was one of the recurring things that comes up with founders.

And when people are asking questions, but there are for sure ways to model that out and to effectively kind of run the numbers, sorry, 

Rob Kaminski: for context in the discussions that you have. And I'm sure it runs us spectrum. Um, my assumption is that it's like B to C like early stage. Like, I dunno if maybe you can like level set the conversation where this most often comes up, um, by type of 

Corey Haines: model and where they're at in their, whether it's revenue or it happens across all of them, B2B enterprise all the way down to like a, B to C really, really low RPM.

To, to pro-sumer everywhere in between. Um, again, with SAS these days there isn't like a, usually like a strictly B2C price point. They'd be like a, like a calm app, for example, it's like what, like a hundred bucks a year or something. Um, they're a little bit more of like the exception than the rule. And also there's a lot of like big enterprise startups who are charging, you know, seven figure contracts, uh, you know, paid upfront, right.

And then their three-year commitments. Most startups are somewhere in between and we're going to have like a blend of premium, low price point, mid price point high price point. Um, and in that case, you have to account for all of those. And, you know, you're effectively B to C. B2B B2C and B2B. Um, and so it happens for everyone.

You have to run the numbers, whether you're charging a lot or a little, I would say that there are some nuances in there, for example, for a higher price point, usually it's. It's going to be more expensive to acquire the customer. So your CAC is going to be larger. Um, and usually there's going to be a longer sales cycle.

So, you know, it really starts to get messy when you start thinking about payback period to the sale cycle, to close to first revenue, and that can be stretched out. Whereas if you're working with like a B2C product, like a savvy Calla, for example, I can run. Get a sign up and get revenue right away. Right.

And so there's, the sales cycle is effectively zero days or whatever the trial length is. And that makes it a lot easier. But again, the numbers can be a lot smaller. So our, our, you know, we're running Google ads right now, for example, and our CAC is around 50 bucks, but the average first year revenue, average contract value.

Cause I, you know, I'm not going to use lifetime value is somewhere around $150. Right. So we're working with. Three to one ratio ish right now around CAC to average contract value for a savvy cow, but the revenue is, is right away. Right? So we start recouping those costs when we do it on time within three months.

Whereas if you have a really high ticket, uh, revenue model, and if that's a longer sales cycle, and if there's even like a long cycle from like, um, you know, from like close to revenue collected, because. Of procurement or whatever it is that can also add months and months, months on there. So also have to account for those things, right.

Account for those buffers account for those delays. But at the end of the day, they're all effectively. 

Ryan Hatch: Yeah. And you're really coming out, swimming and swinging here. Right? LTVs it isn't worth tracking.

Um, I know when I was studying, you know, David Scott's work and looking at like SAS metrics, 2.0 type calculations, and he really has like two main guidelines to abide by. One is LTV to CAC ratio. And another one is really the CAC payback period, which is really what we're talking about here. It's all about cashflow management, right?

It's like, it's like, it's a, it's a J curve. How fast can I pay back? Is it three months, six months or a year. And he really talked about when you run the financials, um, you're always going to be, you know what? You look like, you're losing money. Unless you're really analyzing from a unit economic level. If you're just doing the whole company, it looks like you're always, you're always going to be losing money.

So how do you get these early indicator metrics to know that you're not lighting your money on fire? And I think payback period. If you, if you're at more than 12 months, like you're sunk, like it's just, the cashflow is just too heavy. And so he really encourages to get under 12. And, and so, so I think about, okay, payback period to me makes the most sense to measure if I can get payback in three months, boom, I can turn that over and I can raise money on that.

And you know, I got something going here. I think then the interesting thing is you, you, what you've done is you've just paid for the marketing department or the sales group. But then you haven't paid for operations or anything else. So like, um, are you, are you thinking about like a couple of questions?

Are you thinking about the financial financials in that sense? Like how do we pay for operations in the certain, certain way? Like once you've paid for the marketing and sales group, or another question I have is like, how are you measuring our poo? How are you measuring, you know, average revenue per user?

You know, it's gotta be in cohorts, but it's gotta be. It will talk about that. What's what's our poop in your mind, you're doing this cat. 

Corey Haines: Yeah. So for that first one, I'm measuring like operational costs and R and D and engineering and all that kind of jazz. Um, for the purposes of, uh, you know, effectively, what we're talking about is CAC to LTV or essentially your unit economics of scaling customer acquisition.

The operational costs R and D engineering is effectively not important because it's an entirely different equation. Essentially. What happens is, for example, if you have a, uh, let's just say that you have a six month payback period, the payback period is for the marketing costs. And if you're, especially if you're blending in, into your CAC, uh, basically like labor and salaries and marketing technology, and anything else associated with marketing and sales, then, like I said, That part of the company is accounted for.

So then anything after six months goes to the rest of the company effectively, right. To operations R and D to engineering. And that's essentially what you use to, uh, to model out and grow the rest of the head count. Right. And how you scale up the company. Um, but you know, when we're talking about things like gross margin, um, and things like that, man, that gets really, really.

Complicated very fast. I don't know if it's very helpful in, in many ways I'm sure later on. I think when you're like series B and later, anything about. You know, getting ready for exit IPO, things like that. You, you want to start thinking about gross margin. You actually want to, once you have like a very predictable, scalable acquisition model, then yeah.

You want to think about, okay, let's get our gross margin to, you know, 70%, 75%, 80%. You want to like keep bumping that up, keeping, making it more profitable, more effective, but for really early stage startups. And, and for even like, just getting this going in the first. I would say, don't worry about it.

Completely forget about gross margin. Just focus on those core unit economics of CAC to, uh, ARPU or what's the faculty, the payback period. Now on ARPU, uh, to answer your question there. I guess complicated. Um, but I would say that the, the more price ones that you have, the more you want to separate your ARPU and kind of create these cohorts for our pu uh, if they're, if you only have one plan or if the plans are really close together, then.

It doesn't really make a difference. You can blend those together and average them out, and that will still help you with figuring out that whole equation of cactus ARPU. Uh, so for example, if you have a $9 plan, a $99 a month plan and a $990 dollar a month plan, it makes sense to create an ARPU effectively for each one of those cohorts.

And then to measure CAC against each one of those. You know, using some sort of attribution model, which I can also get to a little bit later if we have time, um, because the way that you acquire them is probably going to be different and you want to make sure that you're not overpaying for the smaller cohorts and you're not underpaying for the larger cohorts and missing out on, on, on, uh, basically the opportunity cost of revenue that you didn't pay for, because you could have acquired a larger customer that way.

Um, but if you had. $49 a month. Plan $9 a month plan and $200 dollar a month plan. It makes sense to average those out into one ARPU. And then that's kind of the main one that you use. Yeah, you 

Rob Kaminski: brought up attribution, right. Uh, and I go into the why right where this matters, um, that you brought it up in your example of where do I double down?

Where do I just completely ignore with your know, my overpaying for the 9 99 user? Or am I underpaying? And when you say underpay, I want to know that because that's where I should be dumping my money. Um, What are your thoughts on attribution? Like, does that get into tooling or you brought up models, I'm really curious on how you 

Corey Haines: coach and approach that.

Yeah. So here's my experience and why I bring this up is because attribution was by far the biggest pain and the blocker when I was at parametrics, because what happens is that you start to look at, okay, where do we invest money? Where do we put more time and resources and budget into? And if you don't have any clear sense of.

Attribution then it effectively becomes a ginormous question, mark. And you're sort of stuck until you get an answer or you just start throwing money at things and seeing what changes and you need to have a big budget to buy high, basically to have a really high risk tolerance. So what we found was I started really looking into.

And initially it felt like, oh, we have some pretty like clear ideas of what's working. Um, because I can look in Google, Google analytics and see where the conversions are coming from. And then I really started to dig into the world of attribution to learn about different attribution models. Um, there's like five or six kind of main ones, but really to kind of boil down to first touch last touch and multi-touch and multitask can be anywhere between like, Uh, there's sort of like, you can give more way to the first and last ones and then like blend out the rest of them.

You can have every touchpoint kind of give equal weight to the attribution and everything in between. There can be linear, there can be decay and, uh, kind of gets, uh, complicated pretty quickly. But, um, what I learned was that Google analytics by default, and you can't change, this is a last touch attribution tracking plus.

So effectively, what they tell you an attribution is that when you look at Google, Google analytics and you see where their conversions are coming from, it's only from the last known source from whatever page they converted from or whatever was the last known refer effectively. I'm like really, really simplifying this.

Um, but that was, that was questionable because most of the marketing that we do is. Last touch. It's not remarketing. It's not, um, some sort of like conversion campaign. It's not even like sales focus where I'm like working leads and I'm trying to get them to start a trial or to close from a trial. Most of the marketing that we do is first touch it's top of funnel.

It's how did you discover us? How do we get in front of more people? So it became, you know, it was horrible to learn that it was last touch because then it was like, oh, Frick, well, this doesn't tell me anything. This is not helping. At all. I don't know where to put our marketing dollars because I can't tell you what the first touch was that got someone who eventually converted into the funnel in the first place.

And so this is one of the other things of your SAS. Your SAS metrics are aligned to you because if you just look at conversions from the last known source from last touch, it's going to paint a completely different picture. If you were measuring from a first touch, which first touches probably the.

Useful attribution model that you should be using day to day. Um, multitouch is more of like a realistic, but if you actually thinking about again, what is the purpose of measuring attribution? It's where do we invest our marketing dollars? How do we get in front of more people? How do we know that something's going to result in revenue later on?

And first touches like the most pure form of measurement there. But first I is the hardest one to measure. And it's also not the default in the platform that most people use, which is Google analytics. 

Ryan Hatch: How do you, how do you solve for that? Like do you, um, those are the best way I can think of to actually solve for that.

Isn't actually data at all. Isn't analytics and all it. You know, randomly survey people on this plan versus this plan per cohort and figure out, Hey, how'd you hear about us freeform text, not drop the science pre-form texts. Like I'm trying to figure out where, where the energy is coming from, or at least where, where they felt like they became convinced or the biggest impression was made.

Right. Otherwise we're just talking about. 

Corey Haines: Yeah. So, uh, I tell everyone to do that. And this was one of the reasons why is because you don't know what you don't know. And the only re the only way to really start to uncover that is to ask people, Hey, how'd you find us? How did you discover us? Where'd you first hear about us?

It's it's one of those things that sounds too simple. You're like, nah, we got to have a better way, but like, no, really that's one of the things that you need. It's not the source of. But it's one of the things you need to start to piece things together for your attribution model to know what's working, what's not working.

So in your onboarding forum, in your demo form in whatever it is that sort of, uh, gets people into the product, ask a simple question, don't make it, uh, like multiple select or like one of the, like, you know, just click on this and select them. Have it be an open text box, a little comment box and how people type in, uh, how they heard about.

You can make an optional that you make, make it required. I don't really have like a strong opinion there. Obviously, if you make it required to going to get more responses are going to be lower quality. If you make it optional, you're going to get less responses, probably higher quality. So kind of pick your poison.

But, um, that's one of the big ones. Number two is you can do some like custom attribution tracking yourself, uh, by using some really simple. Kind of cookie in the browser, uh, native tracking, if you have like service side control of your website, and if you are able to basically make some customizations to track, uh, for every, every visitor that comes to the door for the first time you grab the, the known refer and the page, the landed on, and then later on, you can hopefully track that back to.

Uh, when they converted, now, you can pipe that into your segment, into your parametrics, into your Intercom, into whatever else we ended up custom database. I don't really care. It doesn't really matter. It's up to you, but if you can basically attach cookie on the first known visitor and say, what was their first known refer and first known page that landed on, that'll help you a ton.

Um, and this is more for like the technical folks out there. Threads that can link to and ways that this can be done, but I couldn't, I couldn't really explain it to you. Uh, and then the third way is through some other tracking analytics software. Um, personally, I don't like Google analytics. I don't use it.

I think it's useless. I think it's outdated and. You know, it might even be misleading and a lot of ways, in fact, it's funny because Google analytics now through all the like privacy and data loss stuff, they don't track personally identifiable information. And that is what you need for attribution, unfortunately.

And so by default, they're becoming less and less useful over time. Now there's a whole other conversation around data and privacy and all that jazz, but, uh, I'm not super sensitive to that stuff. I think that it's useful and you should be tracking. As you need to. Um, so I like platforms like split B personally.

That's what we use for Sadie Kell. Uh, you can use fathom and you can send customer events. It's like a privacy first. If that's what you really care about. Um, there are some other really fancy platforms out there, like, uh, dream data and attribution nap, and all sorts of other ones like that, where you have to like spin up a custom kind of data warehouse and track everything in house.

You can go that route. I don't know if it's going to get you that much more than a split B plus custom attribution, plus the, uh, onboarding form where it's asking you how'd you hear about us? So those are like the three ways that I would say you're, you're never going to have like a perfect attribution with picture.

It's never going to be solved. It's never going to have all the answers, but you're going to basically try to triangulate through those three sources. All right. Where do we spend our marketing dollars? Where is like, The, the most known places that people are finding us from and how can we do more of those things to get top of funnel.

That's awesome. 

Rob Kaminski: So Corey, we have about 10 minutes left. I'm not sure if you want to throw another, uh, hot take out there that you had in mind, um, with metrics 

Corey Haines: or, yeah. Well, here's one more interesting one is that this was another one that I realized when I was at Baremetrics was, um, the higher your growth and especially the higher growth rate.

The higher you can expect your turn to be. And this freaks people out, because what happens is that it starts to work and you start to grow and you're adding revenue every month and more and more customers are coming through the door. And then all of a sudden, seemingly out of nowhere, you have a churn problem.

And then people freak out and they stopped the acquisition. They stopped the growth and try to figure out their turn problem. When in reality, that's not what they should be doing. You should actually be, uh, you should be either one working concurrently to fix return, uh, as your growth, but also too, you should actually kind of ignore it because by definition, essentially what happens is when you grow, you're adding more people, top of funnel, where people are describing you, learning about you and more people are becoming more interested in you.

More people are coming through the door, through the form of trust. There are signups through demo requests and more people are coming through the door, into your product who are paying, who actually might not be a great fit at the end of the day, just because they're excited about the product or maybe the product isn't quite at that product market fit for that particular part of the, of the market.

But that's okay. You can expect higher term. When you have higher growth, so I've, your metrics were growing, right? And I'm like, whoa, things are working. I'm getting my bonuses. Uh, you know, things are going well. And then all of a sudden churn, uh, almost doubles, right? And so we stop everything. We're like, we have to figure this out.

Otherwise it's just not going to scale. And so we tried to do that growth completely stalled. And then once we ramped growth back up again, the exact same thing happened again. I was like, wait a second. I thought that we have fixed term. No, no, no, no, no, no. It, it sounds very counterintuitive, but you have to ignore it in some ways, because you can expect for more people who are not good fits to come through the door and you just have to do a good job of filtering them out earlier rather than later.

And you have to be okay with that. In fact, if you look at, um, You can't. But again, I got the sort of inside scoop into a lot of SAS metrics and in particular companies and what I noticed over timing, and one of these things was the high growth startups are worried about churn and the low growth startups are worried about getting more growth.

And it turns out that that's, that's not a coincidence. Those two things are related because you find that a company is growing very, very steadily or. Basically plateauing, but their churn was immaculate. It was like, you know, 0.5% or 2%, which is like basically nothing. And then the super high growth startups, their turn was.

8%, 10%, 15% sometimes. And they're freaking out. It's like, well, you gotta pick your poison. Right. Do you want high growth or do you want really low turn? It's really hard to have both of those. Um, unless you have a really, really strong market, really, really strong product. And you're able to counteract those things at the same time, which a lot of startups can't do those at the exact same time.

And then as a part of that as well, speaking about. A lot of people freak out about sharing, because I think that it's way too high. And if you just run the numbers, uh, for example, like a 5% churn rate, I think that the math works out to where you would turn through about 40% of your customer base every year, but that's actually not true.

And your turn might not actually be as bad as you think, because you need to account for reactivation. Reactivation is essentially when a churned customer comes back and it happens more often than you might. In Baremetrics one of the, so we have that, we split it out into, into five different parts of your MRR.

There's new MRR expansion, MRR, existing MRR contraction, MRR, and then lost MRR, right. Or turned MRR. Uh, oh, and then there's, and then there's reactivation MRR, right? Uh, and that's the one that everyone always forgets about because I see that little thing that they're at the bottom, they're like, oh, interesting.

You know, reactivation, like these customers came back, so reactivation essentially eats into the LA. MRR. And it cancels out some of that last MRR because they turned last month, but now they're back. So like, can we, is it really turn now? You don't want to like, make your own formula and equation for term, but basically let's just say that your, your turn rate is 5%, but the air reactivation rate is 1% that effectively gives you a turn rate of 40.

On average, because you can expect some of those to come back. So again, after metrics, you know, we're growing, then turn goes up when we're freaking out. And then, uh, and then we stop the growth and we try to figure out the turn at the same time the reactivation went down after we stopped growing. Why is that?

Well again, because people are coming through the door, they churn, but then they actually come back. And so I found that our turn wasn't as bad as we thought we can just completely ignore that we need to account for the reactivation. Um, and in fact, you can have pretty high churn if you also have a high reactivation rate, because it just means that your customers are.

Um, for example, you see this a lot in B2C or B2B to see if you're working with like consultants or side projects or, uh, anything to do with like the creator economy, honestly is, uh, people are finicky. They're like, oh, I have a website. I don't have a website. Or I pay for this tool. I don't pay for this tool.

Or I picked up this gig, this contract, this consulting gig, whatever it was. And they just use it for that period. In fact, one of my favorite startups is a spark tour. Uh, founder was Mo uh, founder of Moz Rand Fishkin. Um, and he talks about this because they basically talk about how they know they have a higher turn rate than normal, and they don't care because they expect customers to come back and to only pay for it when they need it.

And so, again, you're turning might be lying to you. You need to account for reactive. Your SAS growth 

Ryan Hatch: message aligned to you. Your LTV is not that we're not worth tracking. Your, your turn rate is lying to you, right? Your attribution's lying to you. And this is, this is all coming together. 

Corey Haines: But there are ways to counteract it.

Right? I hope it wasn't just like a, all these things are wrong, but really like LTB is a sham, but instead you should use payback period. And so you use ARPU that to compare against CAC instead, um, higher growth equals higher return, but also if you have a higher reactivation rate, you don't have to worry about churn.

As much attribution is going to be really tough, but try to get to those first touch attribution rather than relying on last touch attribution. Um, All the stuff is really hard, right? This is me speaking from like four years of experience now and just talking to a whole bunch of startups. Um, I personally, I'm not even like a really big analytics guy, like David Guy.

I, as I look at my own metrics, like once a month, pretty much, um, because a lot of them are lagging indicators. There are very little leading indicators. Um, and so even if you look at things like signup rates, uh, you know, number of new customers, there's really only one. Eight to 10, you need to be really tracking at any given time.

And I would say you really only need to look at them pretty much once a month. Otherwise you're kind of waste your time, just staring at a screen and looking at a graph and pondering, well, what do we, what do I make of this? Right. Um, but if you really understand the core mechanics of what's happening and you'll understand, what's okay, what's not okay.

What are you doing to improve those things? The metrics will take care of them. So. 

Ryan Hatch: Yeah. And again, it's different for every business. I know we're talking about a general SAS here. These are generally applicable to subscription type companies, but like these are all lagging indicators and they don't matter until all the leading indicators are solved.

And in all the other micro customer behaviors, right. It's really been looking at, are we delivering on value? Are they getting value time to value all these other things that are going to be different for every single client? Would you, what do you think about this? Like, you kind of talked about higher, higher growth, higher churn.

Um, and is, is there kind of like, is there a parabola, is there like a sand timer where it's like, you have high turn in the beginning because you don't have product market fit. Maybe you close that gap and then churn goes down. Your growth is really accelerating. Cause you've you hone into this target market segment and you you're able to raise a ton of money, but then.

You need to grow, you need to grow faster. So you kind of like you lose you, you, you expand the top of the funnel really fast again. And then, and then you kinda like the same time or inverts again, where, because you're chasing higher growth, churn, churn goes down again. Is that kind of like, is there a sweet spot in there where you think like, because you have to know when should I pour fuel?

When should I pour more gas on the, on the, on the fire? Right. And when, and when, and when not to do that, they're like product market fit seems to be like, Sweet spot in the sand timer model in my mind. But is, is, is that kind of resonate with you or is that not how you're looking at it? 

Corey Haines: Yeah, no. A hundred percent growth normally comes in S curves, not.

It is linear progressions going up into the right, you know, in a, in a perfect kind of algebraic formula. Right. Uh, it comes when, like I said, you focus on one thing in particular. So maybe, uh, maybe for a while, you're really focused on retention. You're trying to find product market fit and you're working on kind of the core mechanics.

And then you start to scale things up and things start, you know, that's where you get kind of like that first curve of the S but then churn catches up to you. You need to, uh, Maybe some incumbents come by and they copy you. Some competitors pop up. Maybe there is a really fundamental issue with churn the way that people are using it or even pricing.

And so then sort of, it starts to starts to even out again. Right. And then you get kind of the top of that S curve again, but then you want to repeat that S again, you need to be able to figure it out and then create another curve up again. I permit. It was never linear. I think another one of those like outdated models or sayings and SAS is, uh, the long slow SAS ramp of death because in the early days, things were pretty slow moving and it was just sort of like, yeah, you have your, your, your SEO going and you have like your word of mouth going and just kind chugs along, going up and up and up and up, up to the right.

But it's hard to really make things go. Parabolic, like you said, or, or, or to really flatten these days. That is not the case. Um, it's usually flat or it's going up a lot. It's hard to really have a very consistent, slow growth. Um, Not to say, I mean, savvy has experienced something like that, but we still had some plateaus, uh, parametrics had plateaus, but then there was some, you know, parabolic, uh, growth periods, definitely more common than not to have like an S curve rather than just like a straight line graph up into the right.


Ryan Hatch: I love that gross comes and ask hers. I think that's, that's super true. You think of like a. You know, whether you want to even corporate innovation, you think about how do we get to the next level? Is it gross requisition, like not CAC in this case, but actually acquiring a company versus starting a new company, kind of like same thing with a series B kind of startup it's like, do I expand to a different customer segments?

Do I go up to up in the enterprise? Do I go down into B2C and do this SMB play? Do I come with product two? Do I expand inside my, my current clients? Like there's a lot of other ways to get to that next to the, get to the next curve. Right. 

Corey Haines: Oh, yeah, yeah. A hundred percent. I would say too. You know, one of the ones that is largely overlooked is expansion revenue.

I don't know very many companies that intentionally try to grow their expansion revenue where that's like a key KPI of, we want to, we want X percentage of our MRR growth to be from expansion revenue. And here's how we're going to go and do that. But that's actually, I mean, it's one of the huge advantages of having a SAS product running a SAS business is you can have expand.

You know, where a customer starts paying you, you know, twice as much as they were before overnight, they just clicked a button. All of a sudden your RPO doubles from that one customer, right. And more companies aren't, we're coming to should be focused on how to do that, how to move their customers up the pricing tiers of their own structure, rather than just focus.

Bringing more customers into the door. This is actually one of the things that I focused on a lot when I was at parametrics was, uh, upselling. And cross-selling pretty much because I created sort of like a suite of products between analytics, which was like the flagship. And then there was recover, which is a Dunning tool.

And then there was cancellation insights, which was like a turn prevention tool. And then there was messaging, which was, I don't know where it's at today, but it was sort of a Intercom competitor in some ways like a messaging tool. And for about a good year, I focus pretty much exclusively on increasing the adoption of recover and cancellation insights from users who are already using analytics.

And that's where about half of the revenue growth came from that year was just from a running campaigns doing sort of like some inside sales, uh, creating some like product led growth. Upsells and pop-ups and ways to expose the product within there where, you know, there's a, you know, a micro ad in there or there's, Hey, how do we improve this metric?

Or what does this mean? And then there's takes you to a page where it sells you on the, on the product. You can try it out, um, campaigns around, Hey, we'll give you your first thousand dollars recovered for free. If you want to try it out, or, Hey, I'll consult with you for three months on how to improve your retention.

If you also use cancellation insights and that'll be like the core, uh, research. Sorry, a dataset that we use in order to make product suggestions and, uh, fix onboarding all sorts of things like that. Right. But not other people think about expansion or contraction either. If you're, if you have a lot of contraction revenue, man, that is a, a horrible sign.

You have to fix that right away. But, um, more important thing about expansion. 

Ryan Hatch: Yeah. It's always easier to, um, sell the S tell us someone who's already getting back. 

Corey Haines: Right then, then start from 

Ryan Hatch: scratch. Right. And I was reading an article on these, these, you know, Tom companies that raised VC versus top companies that bootstrapped, I was just comparing the two it's really interesting takeaway that you, you know, the top companies in both actually had a negative term, right?

Like it was like expansion in inside, you know, existing contract expansion was cute. And the interesting takeaway is that the bootstrap, the bootstrapping companies had outperformed on the, on the expansion side, um, which is really, really interesting. And I think like there's so many implications on the expansion side of even you talked about, Hey, I had to work with the existing product team, with the existing product to put an ad inside the product for this other product.

And it becomes this really interesting. Like how do you organize teams to work together? To, you know, you know what I mean? Cause you're, you're crossing product experiences at that point. 

Corey Haines: Yeah. That's was a lot of fun. Yeah. And net revenue retention is probably another, another one of those like core metrics that I would say you have to pay attention to that's that's the dream really?

I mean, if you have. Uh, if you have negative churn quote unquote, or essentially positive net revenue retention year over year, um, that is an amazing business model where basically cancels out every one of your turn customers, because their current customers are outpacing the revenue that you've lost from lost customers.

Um, that's, that's a dream. I mean, there's nothing there's no, no other like a business that can be. Right. Um, that, that's what kind of brings me to the SAS where I've kind of fell in love with the business model and tested products. That's really magical, 

Ryan Hatch: great stuff. Corey, this has been so awesome. We talked about payback period.

We talked about first touch being really important. It's so interesting that there's not a lot of. You know, even help on the, on the first touch design, especially that's I think it's going to be an interesting thing going forward with the privacy plays coming. And how do you, how do you capture that? Um, but you covered a lot of great stuff today and I'm so excited because this is stuff that people obviously don't talk a lot about.

I think it's really easy for us. You know, as product people or, you know, to get tied into tooling, right? Isn't that always it off. I just hook this up to this and look at these cool look, what I could do and Intercom this or that. And it's, it's actually less about the tool. It's more about what method and the process and your mental model that you're bringing into the business.

Right. And like you to use what, whatever tools you have, but, but view them the right way and measure the right things. 

Rob Kaminski: Corey. Where can people hear more about the stuff that you post and share and talking about metrics, um, in your community or what's best for people to get at you? 

Corey Haines: I'm on Twitter all day long at query hands co sharing things mostly about SAS marketing.

Um, also if you just go to dot com, you can sign up for my newsletter. It's a cutting edge, sometimes crazy marketing ideas, straight to your inbox. I share a lot of case studies, thoughts like this, um, tactical stories and things that I've experienced framework. Um, strategies that I've noticed. There's also tiny marketing ideas, which is like a drip of, I think maybe close to a hundred, uh, sort of bite-sized marketing tactics that you can implement for your business.

And really it's just like, get your brain going. That's that's why it fails.com/tiny, I believe. Um, and took us to avocado as well. If you're, if you're tired of Calendly or if you don't like the power dynamics there, uh, we've got some fun stuff. Derek, the founder is an amazing product builder. It's a fantastic product.

I was one of the first customers actually. That's how I got connected with him in the first place. But, um, mainly Twitter just connect with them on Twitter and you'll find links to everything else.

show notes
  • Critical Thinking for Startup KPIs
  • Hot Take: LTV is a Useless Metric
  • Why People Think They Need LTV
  • Instead of Lifetime Value - Use ARPU
  • CAC to ARPU for Marketing Budget
  • LTV vs CAC Payback Period
  • When Does Gross Margin Matter?
  • Truth About Attribution Models
  • Google Analytics Is Misleading
  • The Importance of First Touch Attribution
  • How to Get Better Attribution Data
  • Google Analytics Alternatives
  • Relationship Between Churn and Growth
  • Customers Change Their Mind
  • Lagging Indicators and Analytics
  • Growth Comes in S Curves
  • Revenue Expansion Opportunities
  • Net Revenue to Retention