Very True by Verissimo

Alex (01:20)
So again, just to clarify that one more time, you've got dependent variables, independent variables, and the mechanics. Those are your three ingredients for any good model. And you can do that all in one column to just validate and ask the right questions of what do we need this?

What do we need to believe for this to even be worth our time? Should we be quitting our jobs? Should we be incorporating an entity? Should we be taking funding? Like a lot of people are trusting that you already did that. And I'm always kind of unpleasantly surprised by the number of people who have not done that, which is a little bit mind blowing to people from outside of the venture tech ecosystem, but it's real. It happens all the time. And frankly, I'll just call it out.

Like for years and years and years, I was like, I don't understand this. WeWork thing. I don't understand this. WeWork thing. Like they're basically selling a dollar for 75 cents. And so yeah, it's growing really fast, but like it's completely stored, supported by, you know, zero interest rate equity capital and a really charismatic founder. And it was like, well, but maybe it can turn a corner. And I still remember like, actually, I think it was at the same event where I met that founder that asked me that question. was like 2017 or 18. And I was like,

maybe I'm wrong about this one. Maybe I'm just a luddite I maybe I'm too negative. Maybe I'm too pessimistic. And then like the minute I said that it was like two weeks later, the thing came like crashing down like SoftBank pulled their offer and like, you saw that like there was nothing sustainable about what they were doing. And like, yes, everyone jokes about community adjusted EBITDA. There is an idea behind that that if you can really create synergies and create a global dominant brand, which frankly they did.

and an excellent product that you can charge a real premium for and have some sort of monopolistic relationship. This, by the way, is exactly what Uber's done. Then it can work, like, that's probably not a bet that if you're listening to this, probably not a bet that you should be making. I hope someone like Adam Newman is listening to this right now and has that sort of charisma and vision and borderline megalomania. But ⁓ for most people, it's just not relevant. There's a lot of other ways to generate wealth in a more, well, let's say a

lower risk, more independent control your own destiny sort of way. So let's get back to this idea of the one column model. So the first thing that you do is you ask yourself, what are we, what are we, what's our, my favorite question is like, what's our unit? Is our unit a contract? Is it a customer? Is it a cohort of customers? Is it, a node? There's so many different ways to look at this. maybe I'll just, I'll give an example of,

of a model I built recently for a business that I didn't really understand before I invested in it. Frankly, I just invested in the founder. And that's often what happens. knew, so I'll just give this example. I knew that the market they were in was interesting despite not actually knowing or understanding the mechanics of that market. I knew a lot of smart people made a lot of money for a long time in that market, which tells me it smells like there's a good business model to be had here.

Now you always have to tread carefully with things like that because you know, sometimes a lot of those good business models are already eaten up and there's good business models and there's bad business models. And I'll explain kind of how I view that in the context of venture backed companies in a minute. But, um, I still wanted to double check with the founder. I had a lot of faith in that he knew, but we went through this exercise anyway. So we started with, okay, in our case, it was a, a customer contract.

for a service, let's say. And we were saying, okay, we think that this service, and we were basically running a brokerage model. So we think the service will price at $30,000 a year. And as brokers, we take 10 % of that, and we pay a commission out to our salespeople of 6 % of that. Those aren't the actual numbers, it's fine.

okay. So now, now that tells us like, this is how much money we can make. So in that example, it'd be, you know, 30,000, times, 10 % times 6%. Right.

We make $3,000 on that and then we have to pay out, 6 % of the $3,000. So it's really that times, 0.94. And that's kind of going to be our, we'll say effectively our gross profit on this, which is going to be $2,800. Now what's our unit? Our unit is that customer.

Right? Like we've got to go get that customer. So now if we flip it to the other side of the model, it's the, it's the CAC side, the go to market side, the sales side, the marketing, whatever you want to call it. It's the part that brings things in. And this is where I'll mention my other, one of my other key frameworks is that what do companies do? Companies all fundamentally do two things. They make stuff and they sell stuff. I remember years ago, I sat with a very smart kid who had no exposure to

the business world whatsoever, but really, really smart, savvy kid and very intellectually curious. And I said, yeah, I worked in corporate venture capital. I worked at GE ventures. And he was like corporate, like what does corporate mean? It's like, it's actually a good question. It's the word that we all use all the time, but like, does everyone really think about what it means? at that, again, at that moment I came up with the definition, which is companies fundamentally do two things. They make stuff and they sell stuff and everything in between.

That's not one of those two core functions is corporate. And when you think about a corporation, it ends up being mostly corporate, which is crazy. And I can say this when I worked at GE, most of the people, if not all of them that I interacted with were corporate people. They were all cross-functional this or that. Like I got to meet very, very few people who are actually out there like designing and then manufacturing jet engines or who are actually like owning the relationships with the airlines and the leasing companies and stuff like that.

Like that's just what ends up happening when these companies get like so, so big. So let's get back to this example though, which is, okay, now we have this item that we're selling and we have who we're selling it to. we know how much they're paying and how much we get as a, as a brokerage. our, our really unit is a transaction with that customer. Now each customer only makes one transaction and it covers them for whatever that is. So it's not like there's a

repeat thing. There is a renewal aspect, but there isn't a repeat. There is an annual renewal. But that is our unit. It's one item sold to one customer every year. And then we started looking at like, okay, how much do need to pay salespeople? Let's say they expect 100 % upside. Here's the base salary. Here's how many deals we believe they would need to close to hit a total compensation of double their base salary.

And then we kind of looked at it we're like, that's just not realistic. All right. Frankly, I didn't. The founder who knows this business inside and out was like, that's just not realistic. And in his mind, he had assumed we'll go after kind of like these mid market because we think like that market's softer. But as soon as he saw the math in this one column model, and just to give you an idea, like the number of rows we're talking about here is like 11 rows in all of this math. He was like, oh, this isn't going to work unless we go after customers where we think that the ACV is going to be at least like 60,000.

30,000 is just not going to work. It's a waste of our time. so even if we got some of those early, we immediately need to aim up. so, you know, I'll pat myself on the back as like my, you know, whatever value added investor, because this just goes to show that like that business model shapes the strategy of the business early and the strategy of the business early will straight the strategy of the business period. If you can't do it small, you're not going to be able to do it big.

Like I just see that over and over and over again. If you have a real strategy and find, but let's just be clear. There's a lot less exceptions to this than people think there are. Yes, of course there's always exceptions to everything, but a lot fewer than people think there are. So when you're starting a business, validate one column model, super important, independent variables, dependent variables, and then the mechanics. Now the model that I've spent the most time on in the last 12 years,

is the SaaS model. This is it, Like the SaaS model, the recurring revenue model. And when I worked for Harry Weller, may he rest in peace, at NEA, he was like, this is the thing. This was January of 2014. He was like, everything's going SaaS. Like, this is all that matters. And he looked at me and he was like, you're really financier and you're an engineer. I want you to learn this model inside out, upside down and backwards.

He gave me the most amazing resources. He gave me access to all of NEA's software company portfolio board decks, which he probably wasn't supposed to do, but I got them physical copies of them. got, he was like, here's some research from like Joel York and Michael Schock, who are the only people talking about this stuff back then. And then he just started sending me to founders who he felt like were either struggling with it or had really figured it out. And so I got this kind of, always say that's when I started my PhD in, in SaaS modeling, but

Really, that's when I started understanding business models. I had worked at Morgan Stanley for two years and 16 to 20 hours a day for two years. And we built a lot of financial models, but there's a distinction between a financial model and a business model. And I'll kind of illustrate that now for a second. A financial model and a good one is really like a three-statement model. So you've got the income statement, the cash flow statement, and the balance sheet.

Right. The income statement and the cash flow statement exist over periods of time, either a month or a quarter or a year. And the cash flow statement or sorry, the balance sheet is a point in time measurement. The cash flow statement explains the differences on the balance sheet from one period to the next. And the income statement starts with, you know, revenue turnover sales, whatever you want to call it, and goes down from there all the way to net income. Now what's important to understand about the income statement.

Even though that's where people I think most feel most like comfortable somehow when they're like learning finance, accounting, whatever, when they're an operator, that's actually like the least intuitive part. Um, because there's a bunch of stuff in there that's just made up. We've talked about this before, but like this stuff called gap, like the generally accepting accounting principles, they're made up and there's some genius ideas in there. I always talk about this, like, you know, revenue, that's an innovation, depreciation. That's an innovation. And what that allows us to do is economically align.

the quantitative aspects of the business on a time basis. I'll explain what that means a little bit more simply. Go back a thousand years. Things were like traded. Okay, fine. Go back 500 years. Everything was done on a cash basis, whether that cash was gold or coins or bills, whatever it was. It was just cash. So if you spent a bunch of money to acquire something, that was just cash out. And then if you sold it over time, it was cash in over time. Now,

a growing business, then we'll lose money. It's called networking capital consumption. And that's not necessarily a bad thing. This is again, going back to my point on capital structure. There are ways to finance this effectively to support the overall growth of the business and the profitability of the business, more importantly, but I'm not going to go super deep on that for now, but revenue.

which is earned and recognized when goods or services are delivered. That's an innovation. It used to just be like, when do you get the money? And that was the whole story. Now, when you get the money, still matters a lot. And you know, I mentioned this term EBITDA before. EBITDA is like the most upside down, inside out, backwards concept that kind of represents how we've effectively come full circle. So EBITDA is earnings before interest, taxes, depreciation, and amortization.

So taxes, fine. That's like an externality and they're different for all companies. And there's NOLs and all these different things. Fine. But DNA, right? That's depreciation and amortization, which are these non, these are time adjustments to costs. So if you buy a truck for your business, it depreciates over, let's say 10 years, it's a hundred thousand dollar truck. You recognize $10,000 of cost every year for that 10 years. But

The cash already left the bank. So again, what we're trying to understand is the economic viability. And that is the point that I'm to get back to over and over again is economic viability. These gap rules were created before innovative business models, frankly. Like this was innovation. Like there was no SaaS. There wasn't like usage base, all these complicated marketing tools and attribution and all this. Like this stuff didn't exist back then. And there's a strong argument that these things are not sufficiently well suited for the businesses that we're running today.

Could be, could not be. I don't know. We'll see. We just try to do our best to understand the economic viability of businesses. that's, okay, so let's get back to this concept of EBITDA, right? Which is like, again, maybe it's a meme now. I don't know, but it's been around for a long time. Why do we like it? If you ask AI right now, why do people use EBITDA? They're just gonna say, because it's a proxy for cashflow. It's like, okay, but.

Any company has a cashflow statement, like, it's broken into three sections, operating cashflow, investing cashflow and financing cashflow. So really don't we just want to know operating cashflow? And then don't we also just kind of want to know like free cashflow, meaning like free cashflow, levered, unlevered, meaning like, it paying interest? Is it not like what kind of cash is the business actually throwing off? Isn't that just what we know? And the answer is, and why startups use EBITDA is like, you can kind of

do a little like cash accounting and then add stuff back and it's a little bit hacky, but it kind of, again, it kind of gets you close to like, what is the economic engine of this business actually doing? And there's a reason why we have to like go inside out and backwards instead of just going back to the old school way of how much cash went out, how much cash went in. You know, a company could be EBITDA neutral, meaning it's not losing or making money, but it could be throwing off cash or losing cash, right?

I was gonna post about this, maybe I'll make a video about it, but I started going through the financial model of Figma. This is a confusing one, why? Because they got a billion dollar breakup fee from Adobe when the acquisition didn't go through. And the way that gets accounted for is just really complicated. And then they had all these non-cash expenses, which are stock-based compensation, which is a whole nother can of worms.

This is like flashbacks to investment banking. Like again, we're trying to really get to like, what is the economic viability of the business, but we're doing it through all of these standardized, often not often not particularly precise or maybe even accurate methodologies that come from old legal frameworks. but it's the best we got. So that's what we're to start with. And so I'll just, again, talked about a bunch of different stuff right there, but I'm to talk about.

The difference between a financial model and a business model. a financial model, it's not really about the assumptions. I'll just say it. The assumptions are usually pretty straightforward. You're like plugging a growth number. Usually maybe there's a revenue build, but it's like the deeper operational mechanics of the business are generally not represented. And so if you have, you're an early stage startup and someone, you know, one of these outsource CFO firms giving you a financial model, that's like three statements.

The most important thing there is the mechanics. And this is actually, again, the reason I was messing around with the Figma models. I have a tradition to keep me sharp and because I'm a nerd, which is that I build a three statement model from scratch, like once a year, maybe once every two years, I just open up a blank spreadsheet and I make sure I can still do it and model everything through. I figured this year, why don't I do it for real company investment banking analyst style, which is actually like much harder because instead of just making an assumption, you have to like dig through 10 Ks and S ones and figure out.

to the best of your abilities, what is actually going on here? So that's the financial model side. In the absence of, and I will also make it a really important point here, which is financial models get more useful the higher scale companies are operating at. And part of that is just operational financial inertia that like things are gonna change less. The mechanics are more well established.

Even if people don't understand them, the inertia just prevents them from changing too much. There's less volatility. An early stage startup, a financial model probably isn't going to tell you that much. Again, I've made some efforts here and there with like series A stage companies to build like really like rigorous, detailed operating financial models with three statements that all work mechanically. Probably not worth the time and effort, but can be an interesting exercise. The truth is that

If the business model works, that stuff doesn't make that big of a difference because what you're really trying to do, and I mentioned this earlier, is we're talking about the, I'm going to make up a word here, the capitalizability of a company. How do they extract capital? Whether that's throwing off cash and therefore you can do debt, paying dividends, and therefore you can, you you're kind of automatic in the, you know, getting equity investors or just having huge potential.

and, and, or illustrating that potential and validating future potential to, to some extent. So we'll get back to the business model. I've talked through the SaaS business model many times. Like I said, that's really where it started for me was understanding, okay, you've got this waterfall. And actually to me, it all started with depreciation on your balance sheet. have what's called PP and E property, plant and equipment. And you have a waterfall where you have the beginning number, you have the increase, the decrease, and then the ending number for every period.

And then the next period starts with that period's ending number. And you can pull the depreciation, which is the decrease in PP &E from the cashflow statement. And you can pull the increase in PP &E, which is cap X from the cashflow statement. And that just cycles through. That's a static thing. Like I mentioned, balance sheet is a point in time financial statement. On the ARR side, I thought a lot about what ARR is. And I think that again, that's just obviously like.

completely bastardized term at this point of what does it even mean? Who cares? There's stuff on my website where I break down all the definitions. If you want to go see it, verismo.vc, check out the founder resources section. I think it's the first thing in there, which is revenue related definitions in SAS. Annualized recurring revenue, not a gap term. Revenue can be or is a gap term. And again, like I said earlier, it is earned and recognized when goods and services

or services are delivered to customers. And COGS, cost of goods sold, is also recognized when those goods and services are delivered.

the power of that business model. the guys in acquired have talked about this, the enterprise software business model is the greatest business model in the world. Why? Because it accumulates. Accumulating business models are beautiful things because you can pay a hundred to get a hundred and you can net zero. But then next year you pay another hundred and get another hundred. Now you're at 200 and you're netting a hundred. Then the next year you pay another hundred and now you're at 300 and now you're netting 200. And no companies actually do that. But I remember calculating like

SaaS magic number, is kind of overall sales efficiency of like net new ARR divided by the associated sales and marketing expense. And I looked at some of these like scaled companies that weren't growing that fast and their SaaS magic numbers were like close to zero. But like that's fine when you're already at massive scale. When you're growing really fast, those benchmarks are completely beyond useless. They're actually hurtful and send you in the wrong direction. So I've got lots to say on

SaaS business model. Now people don't wanna talk about, SaaS is dead. No, the subscription business model has been around for ages. I have a presentation online. It's called SaaS 202. I've been working on it for over 10 years. It's all the accumulation of a lot of these details, both the theory and then the practice. So we will get back down to business model and how do you know if your business model is good? So we can talk about revenue.

And revenue again is this invention. It is an economic representation that is distinct from cash. Now, if you have problems converting your earned revenue into cash, that's an issue, but let's assume for a second that you don't and revenue is a great thing. You can recognize it on a daily basis, on a weekly basis, on a monthly basis, on a quarterly basis, on an annual basis for

what you've earned. Now, if you're providing a service or giving someone access to something, then it's just a very simple time-based thing. If you're ordering something on an e-commerce website for delivery, it gets delivered and the revenue hits the books. And there's lots of systems that automate these accounting realities, but revenue is a super valuable idea. Before you get anything to ARR and annualizing and da-da-da-da-da and why that matters, let's just talk about revenue. It is the top line. It is the economic

I don't know, gasoline that will say it goes into this engine. It's what keeps things driving. But what we really care about is cash. Cash is king. Now, people talk about profit, they talk about cash. They both matter. The reason these things come up is they both matter because you can have these aberrations that don't represent the sustainability of what you're doing, like the billion dollars from Adobe for Figma.

And that's why we want to understand net income, which again, that actually gets reflected there, but we want to understand net income. want to adjust these things. They're not games. They are an effort to accurately represent the economic reality and viability of businesses. So, but probably the most important thing is quality of revenue. Like I was saying, I referenced we work earlier, like if you're selling a dollar for 75 cents, then you're probably not going to make a lot of money.

And if you're a startup company, that's not going to work. Although I heard some like crazy take recently, which was like, if you're an AI model company and you have negative gross margin, that means people are actually using your product and you're just under pricing it and getting them addicted. And I'm like, yeah, okay. Again, if you're listening to this, probably not for you. Don't do that. You're not anthropic. So probably not going to work.

Like I always say, gross margin, and I'll talk about what that is definitionally and then what it represents. Gross margin, I always say, is margin for error. And if you're a startup, you need a big margin for error. You're trying to figure things out. You're making mistakes. You're iterating. You're hoping to grow really fast. It's like the opposite of like a 200-year-old business that sells, I don't know, ballpoint pens or ball bearings or springs. I don't know. You just don't have a margin for error.

And by the way, those companies don't have a margin for error because they probably used to when they got started and they started making so much money that then like competition rushed in and like prices flattened and then they're all kind of just run to sustainability.

Yeah, it's actually interesting. I just watched a video about how airlines all run cartels. I originally thought it was about like drug trafficking and how they're smuggling drugs by accident. It wasn't. It was just that literally airlines are cartels. And if you look at like Star Alliance and One World and I forget what the other one's called, Delta's one, like they are cartels and they have these agreements and they just keep getting stronger and not having to play by the rules. What they failed to mention

is that the OTAs killed this and consumer behaviors killed this and prices got crushed and it destroyed airlines and a bunch of them went bankrupt and all this craziness. And so this is kind of their way of fighting back and introducing stability to an extremely capital intensive industry. And it's not just that it's capital intensive, but it's also safety intensive. And so you don't just want the lowest cost things. Again, most of you are probably listening to this.

are flying around in Europe or the United States or maybe Canada or back and forth to Israel or whatever it may be. If you look at where plane accidents happen, they happen in like the third world and it's because they're literally not running maintenance properly. So you don't want a race to the bottom in other places and that's why the government can kind of intervene and allow these externalities to exist and even support cartel behavior. anyway, outside the existence of monopolies or cartels, which I heard another great take on monopolies, which is

Monopolies always try to position themselves as not being monopolies, which is like why Google calls themselves I heard this like on TVPN like and it was in the name of somebody else like Google always calls themselves a technology company and technology is very competitive even though like they completely own search and then other companies who like try to fancy themselves monopolies like they never actually are because They're just trying to pretend they are and that they own everything and they don't so

Revenue is the gasoline, but what we really need is results. We need profit. We need cash. Like that's what matters. Cash is king. And I'll talk a little bit here about the, I guess, the kingness of cash, which I'll do. I could do a whole episode on just the discounted cashflow model, but the value of something effectively maps to its ability to throw off cash in the future.

We take the net present value of that cash, again, the mechanics of the capital asset pricing model to the weighted average cost of capital to then figure out the discount rate and a market risk premium and a long-term growth rate to drive these things. It's all well and good. If company's not profitable, there's also what's called the terminal value in a DCF calculation. And that's often more than 100 % of the value of the business when run that way. In the absence of that,

We often try to exchange precision for accuracy, which is usually a good trade. Quickly. The definition of precision is hitting a target. The definition of accuracy is hitting the target. and actually you can think about that for a second, but it's a very important idea. Getting exactly the wrong answer doesn't help you. So quality of revenue. It represents a few aspects. Now in a world where growth matters a lot.

usually because we're hoping to reach profitability at some point. And we saw this happen in the public markets where like all these companies were coming out of zero interest rate where like was grow, grow, grow. And then it was like, wait, now don't just grow, grow profit, profit, profit. And then a lot of them did like they flipped a switch and they went more profitable and they stopped being inefficient about growth. And it worked. And then you got to kind of see how these things happen. I think that that muscle is underrated. I always tell companies

your ability to control your own destiny and run profitably and knowing what you need to do to actually make that happen, which involves running experiments and figuring out control groups and actual doing real attribution. Those are the exact same skills and tools that you need to grow as fast as humanly possible because efficiency is efficiency. If you understand the relationship between A and B on a really nuanced, detailed level,

then you can run it at one mile an hour and you can run it at 500 miles an hour. And it's really the same muscles and processes that allow you to do that. So that's super important. How we get from this thing called revenue into actual cash coming out of the business at the end of the day. there's like an MBA doesn't even scratch the surface on this stuff. People have gotten PhDs on this. Millions of people have built careers on this. You could argue this is CFOs jobs.

⁓ And even most of them are just hanging on by a thread trying to just get there and figure it out and make it quarter to quarter or month to month. But quality of revenue matters. And you hear this more and more. I would say there are two aspects of quality. Actually, we'll say there's three, but they're really, they're all part of the same idea of like converting revenue into cash flow. And those three aspects are

reliability.

Profitability, like contribution margin, which I'll get into, and velocity. And those are the three main things. Now, they are what drives value in this day and age. And because they're all proxies for cash. So if you have reliable revenue, again, the enterprise software model is a great example. A company pays you a year subscription in advance. A consumer can also do this. That's fine. That's also quality.

They pay you cash upfront, which means that basically even if they churn, they're not getting a refund. So you got the money and you know next month, you're gonna recognize the exact same amount of revenue as you recognize this month. Once you get into like governments and big enterprise deals, multi-year contracts, either paid upfront or again, the quality of those customers matter. So obviously consumers, SMBs, enterprises, governments, that's kind of like the stack ranking from bottom to top of quality of revenue.

And usually it's completely inverted with like the velocity of revenue. So if you can manage to get velocity, which means that they all come quickly with the quality or the reliability, then that's a hack. That's a, that's a winning formula. Like how do you sign a hundred thousand dollar contracts within two weeks? That's like printing money. That's having an ATM in the back of your office. Like it's wonderful. And this is also why enterprise software, especially like kind of the

A lot of the open source driven stuff and dev tools is so effective because I always say like the adoption patterns of dev tools are like 17 year old high school high schoolers like like, I, my friend said to try this. So like, I'll try it and like it's free. And then, and then they start spending money on something except like the budget of a 17 year old in a high school is like 10 bucks a month. And the budget of some of these software engineers is like a hundred thousand dollars a month.

And those adoption patterns and the virality and the network effects and like all that stuff can come into play and that's great. And that all speaks again to the velocity of the revenue and the quality of the revenue. I just, there's not, there's, I don't feel like I can say too much about these concepts. Like you've got durability, quality, reliability, all of these like positive adjectives that you can describe revenue with, which is really a description of

the customer itself, the tool that you're selling them, their use case for it, and your ability to deliver on that. And that's all like quality, reliability, dependability, durability of revenue. And that is often a characteristic of your business. And then what is your business? Again, it's your product and your customers. Like that's your market. It's product market fit. It's all of those things.

So if you have product market fit in an industry where the revenue is not reliable, I don't care. Like it's just, it's gonna suck at a larger scale. A lot of people want what I'm selling, but it's really unreliable and it's really expensive to sell to them. Yeah, that's not so exciting. So that's the durability. The velocity also matters. I remember we were working on an IPO when I was at Morgan Stanley, this was like 2012. And it was like a...

really unsexy business, they make like rack mounted hardware for midsize enterprises. Like it's like so boring. was like sort of security, also like storage, like network attached, like that whole jam. And the banker that I was working with who's still at Morgan Stanley, nice guy, ⁓ which is that's not to be taken lightly, he's a nice guy.

He was like, yeah, we do all this like technical stuff, but we have a high velocity business model. Like what the hell does that mean? And the answer is that just like the customers we sell to just move faster. Like they can sign in shorter periods of time and the revenue can come and we can earn it sooner. Now that has two aspects. There's always like before the sale and then there's after the sale. And this is like two distinct periods of time.

But oftentimes revenue growth is driven by a like new customers coming in, but then expansion of those customers. So the land and expand model has obviously been made famous by a bunch of different companies like, oh, we're going to get in with like a $2,500 a year thing. And then we're going to grow up to like a $250,000 a year thing or a $10 million a year thing in the case of like a snowflake or even AWS early. And that's all good. And that's like, that's high velocity. It's like, Hey, they'll buy one rack from us.

and we'll get them while they're small and then they'll keep growing with us. And that matters. They'll make a quick decision, we have a quick access to them, and then they'll grow with us. So that is the speed and the velocity of revenue. Is our business model designed to grow quickly or is it gonna take a long time? on how many vectors is it growing quickly? So high velocity business model, which this company didn't really have, but it was great marketing for a banker to come up with. It had both aspects, right?

relative to other companies in our sector and doing what we do.

our sales cycles are much shorter. Let's say they were three months instead of nine to 12 months, like high velocity, great. And then on the post-sale side, like it grew faster. know, like they would go from buying, again, $10,000 worth of stuff to $100,000 worth of stuff within the first 18 months. again, that's high velocity. It's multiple engines pushing this revenue number up and up and up.

And again, if you can accumulate revenue, like I talked about before, where it's not a one time sale, but it accumulates, then that's also gold. It speaks to the durability and the velocity. And this is why the SaaS model, the subscription model is like so, so prevalent. And then you, again, you layer on the upsells on top of that. Like it just becomes really, really powerful. When I was at NEA, I worked with actually three storage companies.

They were all like actually doing storage. One was cold storage. One was tier two object store and one was like, ⁓ you know, flash arrays and storage people like, we're going to sell storage as a service. And again, this is what AWS S3 is right. Storage as a service. But the truth is you don't have to sell storage as a service because storage is consumable. Like we've all been there.

We get new computers in a large part because it just says like the hard drive is full and then we buy a whole new computer. Or, you know, I used to take a ton of photos and this was, you know, 20 years ago when I was in high school and I would just fill up external hard drives. And I never like went and like deleted and cleared them out and reuse them. It was like, it's somehow different than floppy disks, but like it's not right. Like remember back in the day there was like CDRs and CDRWs like you could rewrite like

I would love to put up a survey of whoever rewrote anything on a CDRW. And I'm pretty sure, by the way, that it only like worked twice. Like, I think that was like a thing that they marketed, but like you could only RW like two times and then the quality would degrade so much that it was like unusable. So storage is consumable. Now, again, if you can sell a consumable as a service, which it kind of makes sense because you do keep accessing it and it is something that you continue to use and get supported. Again, that's...

S3, that's why it's blown up. like, that's why one of the reasons why AWS is such an incredibly massive business. So again, we've talked about durability, reliability, repeatability, and velocity of revenue. And you can see how they're all like, none of these things are really independent of each other. But it's super important to like characterize which of these things work. And as a startup,

Everyone talks about grow, grow, grow, and growth is all that matters and growth is king. like, yeah, we're talking about growth at all costs. Now, again, that's we all the ups and downs of that mattering growth is what drives value. Like, pump the brakes, you know, growth matters like, but only because of what I'm about to say, which is the other term that gets massively misused to the point that I basically stopped using it, which is unit economics. And I've heard this term thrown around.

years. Unit economics. Oh, the unit economics are good. I had a friend who was like, we invested in Instacart because we saw the unit economics were working. And this was like 10 years ago. And I was like, really? Like, are they? know, like, and like, how well and then again, what margin and like, again, it's a big business now. But like, come on. Like, how much more money did this thing need? How did you have to cook up what the unit was and the economics of that to

basically, they ran the same play as a lot of these other companies, which I call the Uber playbook, which is like, okay, we're going to operate at a loss because capital, both equity and debt is like super cheap right now. We're going to operate at a loss for like, I don't know, to the tune of like five or $10 billion of the capital. then, and then we're going to achieve global domination and become a verb status branded global platform. And then we're just going to like double or triple our prices, but people will be so addicted and you know, their lifestyles would be bound by what we do that we just have pricing.

control and then and then we'll be profitable. Again, like I if you're a founder that really wants to do that and thinks you can call me right now, like email me right now. Let's talk about it. But really hard to do. And there's a lot of factors that have nothing to do with your business, nothing to do with your market, nothing to do with you that make that possible. They're just economic factors and frankly, a lot of luck. But

The unit economics, again, when people talk about the unit economics, first question I always ask, what is your unit? And I mentioned this earlier, right? In that case, it was a deal. And the definition of unit, I've written about this, I've talked about this. The unit is where the product side, like the revenue side, meets the cost side. And where those things, like your least common denominator of those is your unit. Now, businesses have a bunch of units. Like I said, you could have users, nodes, customers, accounts, whatever.

volume, whatever it may be. Those are all subunits, but it always is like, what combination of those units boils up to the unit? So I will just run through quickly the marketplace example. I've talked about Uber before. Marketplaces are really complicated because you have two sides. I was advising the marketplace a year ago, or years ago, and the COO was like, well, I just want to know what one supply side member of our marketplace is worth. And I'm like, well, it depends. This is a load balancing.

This is why, for example, Uber used to have ads in San Francisco. like, we will pay a $5,000 signing bonus for drivers. Like what? How can that be worth it? Well, the answer is that if you open up Uber on your phone and there's no car available, you don't open it again. So the retention and the habit forming nature of that is so important and was so critical at that juncture. And they saw that, that it was worth it for them to load balance and really like go heavy on the supply side.

If you don't have a supply side, it's really hard to man side. But anyway, I'll run through this example and hopefully it will make sense. In most cases, a unit is a customer because that is what you are selling. You are selling to a customer. In some consumer style businesses or some e-commerce, it could be a cohort of customers where attribution specifically is very, very difficult and you just have a month of user, maybe for Spotify, that's what it would be, or Netflix.

And then in others, like e-commerce, actually ends up being like orders. Users matter, but like it's often actually just the order. then they look at repeat rates and they try to model that out. early on you want to, when you, when you talk to like adept e-commerce entrepreneurs, they're like, we're break even after the first order. And then, okay, now they're playing with how somebody, if they can get someone to order again, then like, great. Then that's pure like margin on top of that.

In most software companies, it's like, it's just a customer. Like you are selling an account. Like it's the Boeing account for whatever you want to sell. And like, that's the thing now, maybe it's a team inside Boeing, but you can think about it the same way you think about like Salesforce, you know, you have accounts, you have stakeholders inside that account, and then you have like the product that you're selling them. Um, and again, it's where those two sides meet. It's the product side and the cost side.

and like where they come together. So now I'm gonna run through really complicated marketplace example really quickly, which I probably shouldn't do here, but I'm gonna do it anyway, which is Uber. Okay, so first we need to figure out our unit. So our unit is gonna be comprised of drivers, of riders, and then we need to figure out the least common denominator between those things and put it over a period of time. So we are going to pick the period of time of one month.

And we are going to say, we know that the average driver, these are not easily knowable independent things. We know the average driver does 10 rides a day. So that's 300 rides a month. And we know that the average rider does one ride a day. So that's 30 rides a month. So our unit is going to be 10 is going to be one driver, 10 riders.

because that means one driver is doing 300 rides a month. 10 riders are doing also 300 rides a month over that one month. it's actually our unit is actually 300 rides per month. And that that's actually our unit. And then what we understand about that is we have the cost side. So, okay, we know now to support this unit, we know what it costs, which is we have to acquire one driver and we have to acquire 10 riders. And again,

The marketing teams are on top of this. like see how it comes through and all that stuff. And then we have what is that unit worth? So now we have, let's say 300 rides in a month. Let's say the average ride costs $10. That's $3,000. We pay, we earn 20 % on that, which would be $600. Right? I think I'm doing the math right. So we earn $600. That's what Uber, that's Uber's take rate. That's our, is our revenue. So we have

$600 per month of revenue on one side and that's on the revenue side. And then on the cost side, have X dollars to acquire a driver, one driver and Y dollars to acquire each of 10 riders. And that is where our unit comes together. I like the more complicated example because it makes you think a little bit more critically about this. It's okay to iterate through it and like not quite get it clear. It's okay to have it wrong and think you're selling nodes and then realize you're selling customers or vice versa. As long as you're thinking about it.

That is the most important thing, whether you're the founder, the CTO, the head of engineering, the head of sales or the HR person. I don't care. Like this is what you've got to think about.

So now we're going to talk about different kinds of costs. And basically it used to just be that there was this thing called variable costs and then this thing called fixed costs. So like when you learn basic finance and accounting, it's like, okay, you have cogs, have G and A, S and and R and D. So you have cogs which is cost of goods sold and that's a variable cost. like we are IKEA and we sell these glasses. So we have to buy those glasses and then we...

sell them or we don't whatever they make them fine but like our costs of goods and manufacturing and whatever and then and then we sell them and so we it costs us 45 cents and we sell them for a dollar great so like our cogs is 45 which means our gross margin is going to be 55 and that's what we're starting at this is also the superpower of the enterprise software model multi-tenant software the variable costs are super low

AI model variable costs are super high. Now that is a pure, pure like variable cost. But that's not all we care about in the modern business model. We care about much more than that. The other terms I use S and M, G and A, R and D. G and A is general administrative, S and M is sales and marketing. Oftentimes on public companies, you see S G and A, sales, general administrative. like, that's so many things. Like I don't know anything about this business. And then you have R and D, which is research and development.

There are again rules about how you talk about this. I'm just going to call out some of these big public software companies. They'll have like billions of dollars of R and D spend and they're at billions of dollars of revenue scale. I'm like, what is this R and D that you're doing? You ship application layer software. Like what are you spending billions of dollars on? And the answer is someone's job title and what unit they're in defines where they sit on an accounting basis on your income statement, which is kind of ridiculous.

A lot of those people probably belong in customer service, which then would put them into sales and marketing, but in reality should actually put them into cogs because it's a variable expense. What is a variable expense? It varies based on the volume of revenue that you have. So the more customers you have and the bigger you are, the more customer service people you need. It's a variable expense. The fact that it gets hidden in R &D,

then it doesn't even get attributed as a fixed cost, which is a fixed cost driver of growth is crazy to me. But again, this all comes back to this economic reality. what the accountants tell you, what they send you, the questions they ask you to categorize things, you actually want to be as specific as possible and categorize expenses to the best of your abilities, which is really hard when you're a small company because it's like, okay, I'm the CEO and I spent X hours on...

marketing and why ours and sales and see ours on writing code and this many hours and working with customers and this many talking to suppliers like how do I do that? And then it's, I'll just throw it in GNA. So it's very, very difficult to allocate these things, but getting in the habit of doing it properly early matters a lot. It's amazing how much power bookkeepers have. And this is why like ramp is such a powerful tool. I'm not sponsored by them. I am a small investor in ramp. Ramp tries to help you automate these things to like actually get better information.

It's like, great. And they're pulling from all of the kind of model training that their collective customer base has done over the life of all of their customers or all the data that they have to make that better. That's something that AI should be doing. It's something that should be kicking out to like an expert being like, hey, what was this for? Right? Like, what was this expense for? And they should just be telling it and it should be running it back through the algorithm. And then you shouldn't ever get asked again. Anyway, attribution of costs matters a lot.

variable and fixed matters a lot. And then truly variable costs. Again, when we start talking about recurring revenue, I would say there's not just like one, like, okay, is it fixed or is it variable? And by the way, there's some, it's a spectrum, right? So for example, I said variable scales with revenue, but what if you're a business that like has to hire an additional service rep that regardless of how big each customer is,

for every 10 customers that you have. Some customers might be at $10,000 a year and some might be at like $500,000 a year, but it doesn't change. You only need one for each marginal customer. So that doesn't scale exactly with revenue. So anyway, that's one example. But there's a whole other dimension here, which is the recurring nature of costs. And the example I always like to use is implementation versus service.

If you need to pay people to stand up your tool for someone and get it implemented so that it works, that's a one-time cost. Now, it would still generally go in sales and marketing, and we'll talk about, again, this idea of contribution margin, which matters a lot more than gross margin, to the point that trying to calculate gross margin in line with GAAP that actually represents economic, what's going on in your business, ends up being futile. It's just like, yo, let's just get the costs.

and figure out, you could basically score every cost on like, how recurring is this and how variable is this? And then you can start to understand like, how do each of those numbers scale as your top line scales? Like, how much more money do you need to put in to drive it? How much more money is the cost of maintaining it to keep it going? And then how much money does it throw off? It's actually really simple when you put it like that.

But once you start getting into like the nitty-gritty definitions of accounting, it gets like really, really complicated. the most important thing though is just thinking about it. Just for a minute, as a founder, as a CFO, whatever, any operator who's got a budget to manage, just think about it. Just think about, okay, is this cost going to happen again? Is this kind of going to get bigger with each customer or with the size of each customer? Like if you're hosting, like

Yeah, if there's like a million instances that cost more than having 10 instances. So, um, these are all really important things to think about the cost matter a lot. And then that contribution margin is key. And this is where the unit comes in. So you might have that. I'll go back to the Uber example. said, okay, this is your month. Okay. So now let's say it cost us $5,000 to acquire that driver. And let's say just for simplicity purposes, it costs us a total of $5,000 to acquire these 10.

drivers. Actually, no, keep it simple. It only costs $1,000 to acquire, you know, $100 each to acquire each of these 10 riders and $5,000 to acquire the drivers. So that's $6,000 of total CAC to acquire this unit. But now this unit moves through the fourth dimension, which is time. And it earns, it earns like margin for us over time. We decided it was $600, right? Per month. So now that tells us something really, really important about

contribution. So that contribution margin is really negative, but over time it becomes positive. And we know exactly when, right? It was $6,000 to acquire this unit. That unit is adjusted for all the whatever changes, because we see the date over time, is going to pay us $600 a month. So after 10 months, we break even on that unit. Now, I love, as an efficiency metric, love CAC payback, as we call it.

right, because you can look at it in two directions. You can say, Hey, we have all this company data and we see that we retain these people and they keep coming back for this long. And therefore it's reasonable to believe that we can invest this much in CAC because we're going to keep them for this long and they're going to, we're going to keep running from them for this long. And so we can kind of flex into a more inefficient part of the market and grow faster because we trust that they're still going to behave the way that our previous cohorts did.

That's great and well and good when you have lots of data to measure. Before you have that data to measure, you've just got to ask yourself, okay, let's say those numbers were different. Let's say it was $60,000 and it was $600 a month. Okay, so now we need to keep them for a hundred months before we pay back. Now, even if you're going to keep them for a hundred months, like that's a ridiculous amount of time. Like it's like eight years, you know, like you're going to keep these customers for over eight years. Now,

Then you introduce another whole problem, which again, it talks about the cost of capital, which is like, you had to fund $60,000 of spend, which you're going to earn back over eight plus years. Like that just is a really expensive way to fund and run a business. As soon as those distances get over like 12, 18 months on payback, like you start just having to capitalize the business in a really expensive way. So, this is a lot of information is moving on sorts of business models.

ways to think about it. But the most important takeaways are you've got to create the tools, you've got to think about it, you've got to figure out what questions to ask, you have to map the mechanics of how the business model actually matters, you need to think about least common denominator between revenue side and cost side to get to your unit, and then start asking yourselves, what do we need to believe? And as soon as you start asking yourself that, the answers will just fall out. I had this conversation yesterday with a young founder about decisions. And I said, look, great

Analysis results in non-decisions. Like you don't want to have to make decisions. Decisions are expensive. They're risky. They're like emotionally taxing. Like you don't want to make decisions. You should have the data in front of you such that you could just be like, ⁓ that's obvious. It's a non-decision. Like we obviously need to this. Like my example with this founder, he was like, ⁓ yeah, we can't target that part of the market. We got to target this part of the market because like that just won't work. There's not enough margin for error. We need a better contribution margin.

if we hope this actually works. So the last thing I'll say is that finance, which I talked about on my other podcast with Ariel at length, and there's more to come there. We'll get into more of the tactical stuff. Finance is data. It's just numbers. It's numbers data. I mean, most data is numbers. I guess we have LLMs now, but most data is numbers. It just has a little dollar sign in front of it. And you have to remember, I was using this definition. Now, finance is quantitative.

resource allocation and your job as a CEO, your job as any sort of operator is resource allocation. Where do you want to spend time? Where do you want to spend money to increase the value of the business? Grow profit, strategic, short term, long term, whatever it may be. But that is your job. That is what you want to do for yourself. That's what your shareholders are paying you to do. And that is what we all want to do in the name of great, efficient markets and capitalism.

But that's what we want to do. Finance is data. Data is finance. People talk about data being like the new oil. I actually thought of another more important thing, which is like data is actually like blood. And it actually like helps the entire system communicate, right? Like, and you go, you get a blood test, right? And it tells the doctor like an insane amount of information about how like every organ in your body is behaving and how they're interacting with each other. And that's actually what...

Financial and operating data tells you whether it's dollar sign data or user data or whatever it may be You want to be able to like that you want the system to work well enough that you can just like dip something into it Take a sample take a snapshot effectively and quickly understand like how healthy is this organism? How are its its internal pieces working together? Are they working the way they should are they not are things out of balance and? That makes it really clear like usually again results in kind of

obvious non-decisions. And so what you want is good data, good understanding of the mechanics of how that data interacts with each other. And that will help you do the analysis that you need to do. Again, the analysis will do itself. You'll just have a chart and be like, ⁓ we should do more of that. And it's non-decision. It's like things are just a realization. And that's the world you want to get to so you can start focusing on more interesting problems, frankly. And you can quickly understand if the ideas that you have are