Say goodbye to relying on “magical numbers” and instead focus on a thorough understanding of your market and customers for better results. We discuss key business metrics for growth, such as customer acquisition cost, campaign effectiveness, and churn calculations. We also stress the significance of segmentation, clear customer profiles, and willingness to pay for each segment.
So don’t miss this insightful episode full of valuable advice and perspectives to help you navigate the challenging world of startups.
Listen to the podcast to learn about:
01:30 - Common Pitfalls When Calculating TAM
07:20 - What is a Good TAM from an Investor’s Perspective
09:00 - Calculating TAM for a Category-Creating Company
11:00 - The Right Way to Calculate CAC
18:30 - How to Think About Churn
28:00 - Change the Question from LTV to Payback
30:15 - Revenue, Retention & Cohort Based Analysis
41:00 - Returns, Promotions, Discounts & Refunds
46:00 - Averages Vs Medians
50:00 - Smile Curve for SaaS Companies
Read the complete transcript below
Shripati Acharya 00:25
Hello, and welcome to a very special edition of Prime Ventures Podcast, where I’m here with my partner and colleague, Amit Somani. I’m delighted to be here. Welcome, Amit.
Amit Somani 00:40
Thank you, Shripati. Looking forward to it.
Shripati Acharya 00:45
Today, our episode is Bad Math in Startups. We are going to talk about how, in various ways, we see math being twisted, turned, otherwise butchered, perhaps in inadvertent ways by founders. Our goal in this conversation is to highlight some of that and perhaps offer our point of view, which might be helpful to our listeners.
So without further ado, let’s jump right into it.
Amit, the first thing which I would like to kick this off with is a conversation which we have in almost every pitch, and we open with that, which is the market opportunity or TAM. What are some of the pitfalls and mistakes we see in that when we are talking to entrepreneurs?
Amit Somani 01:35
Absolutely, Shripati. The TAM, as it’s called, total addressable market, is definitely one of the several pet peeves. We’ll cover many of them as we go through the rest of the podcast. My biggest challenges with TAM is that most people will just do a very simplistic top-down model.
Let’s say you are a food tech startup, you’ll be like, “There’s 1.4 billion people in India, only 10% of them ever use this, and they eat three meals a day. That is 140 million times three, that’s 420 million transactions per day. Even if I only got a dollar per transaction, this is a trillion dollar opportunity.” Absolutely bogus, right? Because that is not really the TAM.
The ideal customer profile may be an Indian who eats three times a day or is mobile and wants to order online or whatever it is, but you really have to take a bottoms-up perspective to TAM, saying, “What’s the ideal customer profile? What is it that their willingness to pay for this is? And what is your ability to address that customer?” In that total addressable market, the addressable part is very interesting and relevant.
In fact, if I go a step deeper, I would even say that the serviceable addressable market, which is to say that the addressable customer segment, the willingness to pay and your ability to reach that customer eventually acquire that customer, is very important. That is one that really gets me in terms of-
Shripati Acharya 03:10
Let me ask you, are you really distinguishing between a top-down and bottom-up, or are you still talking about how to do a better top-down analysis here?
Amit Somani 03:20
No, I am actually distinguishing between top-down and bottom-up. I think top-down can be indicative, that in terms of a litmus test or a laugh test, that there’s something here.
When you are building a revenue plan or an annual operating plan or a three-year how am I going to build this business for this round of financing or the next round of financing, you’re not going to say, “I’m going to serve India.” You’re going to say, “I’m going to start in Bangalore or in Delhi, NCR. I’m going to start with this demographic. I’m going to use this channel to get to these customers. I believe customers in tier one will happily pay a 99 rupee subscription or transaction convenience fee.”
I’m really talking about building a TAM model that bottoms up. I think the best models are triangulated, just like an annual operating plan, so it’s totally fine. I’m not dismissing or poo-pooing top-down at all levels, but the best models will be let’s do it top-down, let’s do it bottom-up, let’s do it channel-based, let’s do it based on the competitive or the existing money that people are spending on eating.
I just had randomly hooked onto food, but in that category. But I would say if you have to pick one model, then I’m definitely a big fan of the bottoms-up model.
Shripati Acharya 04:30
Got it. Let’s take this… Just the food subscription business here. So would you say that if a company is coming and pitching and you’re listening to their pitch, would you also like to combine the company’s go-to market when they’re actually thinking about the TAM, or you’ll say, “Hey look, in an optimal go-to market case, I would actually do it”?
Amit Somani 04:55
Excellent, nuanced question, Shripati. I would expect no less. I am looking for initially the customer segmentation. So for the TAM, actually just the segmentation is enough. Saying tier one India, who makes more than 10 lakh rupees a year, that’s my target demographic for this particular subscription offering. That’s fine.
I don’t need to know how will you read them, whether you’re going to use Google or Instagram or MyGate or whatever. That’s a second order of discussion and detail. So not looking for the full GTM, but looking for a sharp customer segmentation and clear profiles in each of those segments and the willingness to pay, because that really gives me a proxy for how much the total price here could be in terms of the addressable revenue opportunity.
Shripati Acharya 05:45
Okay, got it. What would you say makes for an attractive TAM?
Amit Somani 05:55
Actually, basically a summary of what I just said. Clear segment, willingness to pay, and then the second order thinking is the ability to reach those customers. My simplest rule of thumb that we share with the team at Prime internally is if you are going to need a PhD degree to compute TAM, it’s probably not a good TAM. If you’re frivolously saying, “Look, billion people or 10,000 startups will buy my SaaS software.” Only if 10% bought, that’s also not a good thing.
But the simpler the model is in terms of… Usually, I’d say fit it in one line with a 40 pt Calibri font, saying, “X multiplied by Y, multiplied by Z, unambiguously is a billion dollar market or a half a billion dollar market or a 5 billion market.” Those are the easier ones to… Of course it doesn’t mean that you won’t dig into it further, but just a simple definition or a formula is what is a good TAM.
Shripati Acharya 06:50
Which means it’s very directly connected to your price points.
Amit Somani 06:55
Yes. Price point, total number of customers or businesses you’re going to reach and some notional ability to reach them. Because you could say, “Look, I’m going to address everyone.” Everyone is a target customer. Usually a bad idea, not just for GTM reasons, even for TAM reasons. Because even to address everyone, you need to reach everyone. You need to have a brand for everyone, you need to have a distribution strategy. So it becomes more challenging.
Shripati Acharya 07:15
And so, what would be a good number to make it attractive from an investor perspective?
Amit Somani 07:20
Look, every investor will look at it differently, but one thing that’s common to all venture capitalists, people are looking for large, addressable markets, because venture is not a business that is a cashflow business or a day-to-day profitability kind of business to begin with. You’re building for a large outcome.
So rule of thumb I would say is anything less than a $500 million TAM is probably a non-starter, to begin with, especially if you’re at the early stage of your company. Because one thing that we won’t get into today is some of the best entrepreneurs will also expand the TAM by going into sister markets and all that. So for today’s podcast may not be in scope, but we can cover that later.
But I’d say at least $500 million to $1 billion of bottoms-up, addressable TAM is a good number. The reason I say that is for most companies to get to 15, 20% market share is a daunting proposition. Microsoft in the late ‘90s got to 20% market share and got the FTC after them. Everybody says, “Of course I’ll get 20% of the market because I’m the best,” but doesn’t really happen. So I would say that basically makes it that your obtainable revenue is at least $100 million in some scenarios, say, over five, seven years.
Shripati Acharya 08:30
Before we get off the TAM subject, what if I come in and say that, “Hey look, I’m actually doing a category-creating company,” something which at Prime we really look forward to evaluating and funding? We have done that several times in the past. So in those cases you could say, “Hey look, I’m creating a category. Maybe the TAM is only $250 million, but I will still actually reach $100 million of my revenue in some reasonable period of time.” How would you think about that?
Amit Somani 09:05
Yeah. No, wonderful question. Actually, I’ll take it back to product for a second and I’ll come back to the TAM. Jack Dorsey has this code that nothing is invented new on the internet. It’s all the same human needs, the same business needs. I need productivity, I need creativity, I need quality, I need measurability, I need transparency, whatever it is. I’m mixing B2B and B2C companies here.
So really, there will be some proxy for what you are thinking you’re unleashing, whether it is an iPod or an iPhone or even a Sony Walkman from days past, to say, “Okay, there’s people listening to music. And therefore, they’re already spending X, Y, Z.” Now you could say, “I’m going to expand the number of people that get to listen to that kind of music or use that kind of device or are much more mobile than they were before.”
To say that there is zero proof and this is a brand new thing… By the way, that does happen from time to time. We’re sitting on the ChatGPT revolution and there’s so many products, right?
But for every ChatGPT or something else, there’s also a Segway scooter that there was no real demand. So I would just say look for proxy signals for… Is there any spend on that category? Is there the customer’s economic background? Is it relevant, whether it’s on the business side or on the consumer side, to be able to pay for that?
You’re right, there are some absolutely truly category-creating companies where it’ll be difficult to compute the TAM. Even as us as VCs, we have to go and back them because they’re solving a meaningful problem, and someday we’ll figure out how we’ll make money.
Shripati Acharya 10:45
Fair enough. I think I would say one thing to avoid for entrepreneurs is to not do a goal seek on Excel.
Amit Somani 10:55
Yes. We’ve seen that. I have seen magical numbers, like 1.18 and then you do some regression and it’s like, “Oh wait, this is just extrapolated from this thing.” So yeah, I would avoid that too.
One other thing I’d want to talk about Shripati, is the CAC. If you go down and then… On the CAC, maybe I’ll start and you can chime in, because related to the TAM question, and then you said, “Hey, what about GTM?” Right?
So on the CAC one, I think the challenge that I find is that, obviously initially, to bootstrap the funnel because demand you have to create and you’ve to leverage. You’ll go spend some money on performance marketing. Now it could be Google or Instagram or whatever it is that you’re spending money on.
However, one of my pet peeves about CAC for a company that has a little bit of traction, it could be a few $100k ARR or half a million ARR or whatever it is, is people always talk about blended CAC, like, “My CAC is $200.” Okay. It’s $200 for what segment? For what channel? For what price point of the product.
And typically it’s like, “No, no, no. It’s everything. All.” But no, you’ve been around for a while. You’ve just started spending on marketing three months ago. Earlier you were acquiring customers organically or through referral or through your Facebook group or your Insta channel. No, no, it’s all blended. That’s not really your CAC.
We’ll talk about averages and medians later. Because the cost of customer acquisition is very relevant to a channel or a segment because it depends on where you’re acquiring.
You might actually find that your paid CAC, to distinguish the term, is actually $600, because it’s very expensive to get people on these channels. However, your organic thing is zero. By definition, it’s organic. Whereas your SEO CAC, which is a longer level gestation period, you build on SEO or ASO. You’re spending money on content marketing, on SEO, but you’ll realize the results six, nine months later.
So it’s very, very different. It’s very important to know what is… At the very least I would suggest to entrepreneurs to say, “What is the paid CAC for all the paid channels if you must blend it? And what is all the organic channels? When did you start what channel?” At least at a broad level, at monthly, quarterly, whatever it is. That is one thing that I would strongly recommend. That really bothers me.
Related to that is the fact that CACs don’t stay stable or grow linearly. So it’s not like, “Well, my CAC used to be 500, now it’s 300. So therefore once I grow even more, it’ll be 150.” No, it won’t.
In fact, if anything, unless you’re in a massively monstrous category, which is completely untapped, and in a category-creating startup, you will find lot more competition comes in, you will find some harder to reach customers, you will find customers that are the late majority, to use Geoffrey Moore’s framework, et cetera.
And therefore your CAC actually, I believe, could increase even though your customer base is also increasing, but the ratio at which you have to grow… I think thinking about CAC, not just as a static point, but seeing how will the CAC evolve as you go along, totally is dependent on the scalability of the channel that you’re using and the readiness of the customer to pay for it. Those are a couple of thoughts. Any thoughts you have?
Shripati Acharya 14:25
No, just a comment there. I think the folks who actually spend a lot on performance marketing will probably identify with this, is that other reason CAC increases is because the keyword which you’re bidding on start moving from the long tail to the more torso of those keywords, meaning that where there’s a lot more competition.
If you’re a FinTech, you might initially be saying that, “Hey, look, I’m a savings product for people who are students studying engineering.” I’m just making this up. The set of keywords to attract that is different.
But as that particular customer cohort gets exhausted, you might have to start advertising on just the savings product and then the click-through required and the bid required on Google to actually get those clicks. Now you’re competing against HDFC Bank, for instance.
Amit Somani 15:15
Absolutely. By definition, whenever you optimize your marketing spend, you’ll optimize to the most efficient frontier. So every time you need to scale it 10X, if you’re at 100,000 B2C customers and you go to 1 million, exactly your point comes into play, because the low-hanging fruit is gone. Now you’re, like you said, you’re competing for a savings product with HDFC Bank and ICICI Bank and 20 other FinTechs there.
So I think that is one. The other is, which is I mentioned earlier, is some of these customers may not be, especially for category-defining or creating products, they may not already be out there searching and all that. So those are very expensive to acquire, because there you’re trying to convince them that you need this new mousetrap.
Shripati Acharya 16:01
The intent is not really there.
Amit Somani 16:03
Intent is not expressed for sure. And therefore, for example, in CAC, attribution is important. I may have seen an Instagram reel one month ago and then saw, clicked an ad on TripAdvisor or make my trip two weeks ago, and now I come on some third channel and I convert.
Now where are you going to attribute that cost of what you spend on Insta, TripAdvisor, whatever, and now the last click? It gets complicated, but I would just say at least don’t do blended CAC. Do paid versus this, and then have segmentation as you go along, and ability to express the scalability of the channels.
Shripati Acharya 16:40
So maybe one more thing before we move on from the whole CAC equation, which is that, how do we answer when entrepreneurs say that, “Hey, look, I’m doing a lot of advertising on the SEM side, but that is leading to organic ultimately, right? That’s the reason why I’m blending the two.”? How do you think about that?
Amit Somani 17:00
Great question. I think the timeliness and literally this first click versus last click attribution will lead into that. The simplest measure of organic, certainly on the search side, now that is being disrupted through GenAI and other stuff but let’s stick to the old world for a moment, is that the organic keyword searches for your brand should be increasing.
People are not searching for your brand or your unique products, USP or whatever. You know that all the other investment you made in brand marketing or performance marketing or SEM is not yielding results. The simplest way I would say is the timeline is very relevant of when you did the spend. If you did an IPL marketing campaign three months ago, it’s going to be hard to attribute that in the month of June 2023 or July, maybe, 2023.
So you can keep some rolling window. I would say a quarter, maybe. Most people, I would say, should be more conservative and do it in the last 30 days or maybe 60 days. But you can use a time window to do it. By the way, over time it’ll be a rolling average, because what you’re going to spend now on SEM or brand brand should have an impact on organic later. It’ll wash out beyond a certain timeframe.
Shripati Acharya 18:13
Amit Somani 18:15
Yeah. Great. Let’s switch gears from a bit to the top of the funnel to really just going to the other end at the bottom of the funnel, which is, let’s say, you’ve got the customer, they’re paying something, now they are beginning to churn. I know that’s one of your pet peeves, and you’ve even written a lovely blog on it. What do you look for when you look for churn, and what annoys you when you see the layman’s or the vanity churn metrics that people usually throw at us?
Shripati Acharya 18:45
Churn really is… Whenever we look at metrics, we have to ask ourselves what is the objective of that metric. In the case of churn, the objective is for us to figure out, as an entrepreneur, whether the value we are delivering to the customer is sticking. Because the customer churns when they’re expecting a certain value and they don’t see that value. And they say, “Okay, I’m not using the service.” That’s the objective, so I understand how well the value proposition is sticking.
And so, I would say it’s true for all metrics. You need to be clear about what the objective is, because taking a particular definition and then just cookie-cutter pasting it into your business usually results in just a wrong set of numbers. And the garbage in, garbage out you’ll make… If you’re actually not using those metrics to influence or inform your decisions, it’s really not a very useful metric.
So in the case of churn, the first thing I would say is just imagine a SaaS business where you know have usually monthly subscriptions, you have quarterly subscriptions, you have annual subscriptions. Let’s say you’re selling annual subscriptions. You sell an annual subscription in January. And a month later, in January of this year, you sold 100 subscriptions, and they’re all annual. What is the churn in February, March, April, May, and June?
Amit Somani 20:03
Shripati Acharya 20:05
Absolutely right. Now you could come and say, “Look, I have zero churn.” That is the first thing which I feel is probably giving the wrong inference there. The way to think about it is that the customer has no choice, in one sense, but they have already prepaid for it and they’re not requesting a refund. There are ways to actually account for that while looking at things like what is the engagement and so on and so forth of the customer, but let’s ignore that for a second.
One clean and simple way to look at it is to calculate churn based on up-for-renewal basis. What that means is that at any particular month, you look at all the customers who are up for renewal, who needed to renew their subscription.
Let’s say you’re in February and now a whole bunch of annual plans, which you sold previous February, and a whole bunch of quarterly plans, which you sold previous November, and a whole bunch of monthly plans, which you sold in January, they’re all coming up for renewal.
You take that as the denominator, and in the numerator is all of those folks who did not renew. That is a very active decision being made by the customer to renew or not to renew. They’re actually paying and voting with their wallets. That is a right measure of both retention and churn.
That, I would say, is probably the most common mistake I see. You see it consistently in SaaS companies specifically, much more so than, say, B2C or any other companies, because the subscriptions keep varying the time periods on that.
Amit Somani 21:40
Completely. Would you, Shripati, ask people to segment it? Let’s say I only have two plans, an annual plan and a quarterly plan. Would you suggest that they do a quarterly plan churn metric differently than the annual plan churn metric? Or would you still say, “Look, a renewal is a renewal. It doesn’t matter when it came up for renewal.” How would you think about that? You can make it more complicated with the monthly as well.
Shripati Acharya 22:05
Yeah. I mean it’s similar to the segments and CAC, which you mentioned earlier, the channels. Actually, it is useful to do it by the… by the duration of the contract, because consistently you’ll find that… This is a very fairly common experience of startups, is that moving customers from monthly to quarterly, for instance, immediately sees a reduction in churn and also follow-on churn. It’s a higher retention.
Because in one sense, the cognitive load on the customer is reduced from a monthly decision to a quarterly decision. So I would say that doing that is useful because that will help you decide if you want to, for instance, completely scrap your monthly plan. It’ll mean that a lesser number of folks might sign up. But by scrapping, you might just be getting much better retention.
Amit Somani 22:50
Great. Another one that we often talk about is this notion of, let’s say, in particular in SMB SaaS, which is not necessarily always the vogue, where the end customer itself goes out of business. Therefore, it’s not like they renewed. Do you double-click a little bit on that and try to see did they go to another competitor or another offering, or perhaps they went out of business? And so, would you count that in churn or would you count that as no longer using, and therefore I get a little credit?
Shripati Acharya 23:25
Yeah, I would actually say it’s important to distinguish between that. That is specifically true for SMB businesses, because that is, in one sense, involuntary churn, for lack of a better word, right?
I would actually remove that. But if you have a large proportion of it, which frequently happens in SMB, you might have to ask yourself whether you’re targeting the right segment in the first place. In SMBs, some things don’t have a high churn, but most things have a high churn precisely for the reason which you’re talking about.
I think it affects the business, it tells you whether you’re on the right segment or not, if most of your customers actually have very high mortality rate, so to speak. But it doesn’t help because I answered the question whether your value is sticking or not, because your value might be sticking and the customer is going out of business. It helps to actually just split them into two.
Amit Somani 24:20
Got it. Moving along, there’s one thing that is very closely tied with churn, which is the customer lifetime value, or the LTV as people call it. I think this will also have a bearing on the B2C side, which we didn’t cover in churn, which is in a B2C setting, I’m not buying every day or every week.
I mean, maybe some businesses are like that where there is an opportunity to buy every day. I may be buying three times a year, four times a year, et cetera. How do you think about LTV and your challenges and myths around LTV in general? And then we’ll dig into B2C a little bit as well, but-
Shripati Acharya 24:50
Probably the biggest, I would say, confusion in an entrepreneur’s mind is that what actually is LTV. It might seem like a very simple thing, lifetime value. Let me give an example. Suppose I’m in the business of selling this mug here, and this costs me hundred dollars, a very expensive gold mug. I decide to set up a site in which I’m going to sell it for $50. So now I might actually get a lot of demand for it.
Now if you buy it, what is the lifetime value of this customer? The answer would be, well, I’m getting $50 from this customer. The frequent and common mistake here is to use that $50 as the lifetime value. Let’s, for the moment, say the customer only transacts once. The customer is actually a -50 lifetime value. Because for every cup which you’re selling, you’re losing money. It’s not uncommon in the early days of e-commerce you had a negative gross margin in a lot of cases, right?
You have promotions also in a number of businesses. So the lifetime value is the value piece of it. We’ll come to the lifetime in a second, but the value piece of it is actually the contribution which you, as a business, is getting from that customer. It must remove your costs to both provide the product and service that customer.
The way I think about, to simplify things, is they should remove all variable costs associated with that customer. In many cases, we’ll find that by the time we acquire that customer, by the time… No, the CAC is not part of LTV because we take it as a ratio, usually LTV to CAC. But you have to remove all COGS, service costs, customer service costs… all variable costs associated with that customer, and then the remaining value is the value. And so frequently when you do that, the so-called LTV-to-CAC ratio will look dramatically different than otherwise.
Amit Somani 26:50
Let me come back to you and say, let’s say that $100 mug, I was selling day zero for $110, and the cost of all the other variable stuff is not more than $10. We just make an assumption COGS is 100 and then the rest of it. How do you still think about LTV in that context?
Because if I may just buy this mug once in life and never buy it again, what durations do you use? Are there any good practical… Everyone will say, “Oh, my goodness, I sold Shripati a mug. He’s going to buy 1000 other mugs from me over the next five years.
Shripati will never, ever even come back to the site because he just saw something, clicked it, somebody’s birthday coming up or whatever, and bought it.” How do you think about the timeliness of the LTV, assuming that the basic condition of value is met?
Shripati Acharya 27:35
Yeah. I think that’s a difficult question for entrepreneurs to answer because they have usually not been in business for too long to actually know that. Let’s actually take the case of both a B2C and a B2B. In the case of B2C, you’ll have to make a educated guess on why on earth will that customer come again for this month.
If it’s an consumable, you can say, “Hey, look, when I come, what is my repeat rate? Will they actually come often enough?” If they’re actually buying toothpaste, they’ll probably come often enough. But if they buy something which lasts them a year, then actually you’ll have to reacquire that customer. This frequently happens, for instance, in fashion.
Right? Even in travel. If you’re actually doing international travel or something like that, you’ll have to end up reacquiring that customer. The repeat rate here becomes really, really important. The entrepreneurs have to focus on proving and validating that the repeat rate they’re presuming and the lifetime they’re presuming actually is correct. And so for B2B, it’s a slightly easier answer by actually changing the question itself. Usually, my suggestion to entrepreneurs is to change the question from a lifetime value to payback (for B2B).
How soon is my cost of acquisition paid back by the customer’s contribution? If that is anything less than 12 months, it’s a very good number. 18 to 24 months, you actually start wondering, because now you need to retain that customer for so long. If you’re getting a payback in one month or in the first transaction, you’re actually in very, very good wicket.
Amit Somani 29:10
Absolutely. Absolutely. If I may add on B2C, I think again, like what we talked about earlier, seeing the proxy for that demand, so whether it is travel versus a food startup versus a mobility startup, like Booking an Ola or an Uber or some kind of taxi service. You say, “Look, on average a person going to office 250 days a year, 300 days a year, books a mobility or shared mobility twice a day.
What’s the share of that wallet am I getting?” I think there are ways to model it, but at least my peeve is when people say, “Look, I’ve got them for literally five years or something.” You don’t really know because the customer behavior and the innovation and the competitive landscape changes so quickly that you have to be a lot more reasonable about the actual lifetime. B2C certainly, customers are a lot more flippant and a lot more demanding.
Shripati Acharya 30:05
Absolutely. And I would say that, for entrepreneurs, they need to be conservative in these assumptions because you’re running a business business. And if you’re not conservative, you’ll quickly run out of money on this one. So making optimistic assumptions on that is really very risky.
Amit Somani 30:20
Absolutely. Shripati, let’s pick it up from the average order value and how do you kind of look at that? And also, this notion of aggregate revenue or aggregate metrics as opposed to cohort-based metrics, whether it’s for B2B companies or for B2C. Any of your thoughts on that in terms of what’s the right way to go about it and what is a bit annoying to look at?
Shripati Acharya 30:45
Yeah, so the thing is that when we look at revenues, let’s for the moment say you’re a gaming company and you have a 300k or 500k ARR, right? And you’re saying, “Look, I got a 500k ARR next month, I’m 550k ARR, and so forth.”
But the challenge is looking at these numbers is that it doesn’t provide any intelligence on how this number is going to trend in the future. Because what we are looking for as investors and what really you are looking for as an entrepreneur is what is my future revenue and future profitability going to be like?
And if you don’t tease apart what are the constituents of this MRR, you will not have any success with it. So what I mean by this is that let’s say you have $50,000 of revenue this month. You see my MMR is 50K.
How is that distributed between customers you have acquired this month between customers who are in their second month, those who are in the third month, those who could be a year ago customers, and how they are doing?
So that for example, is a cohort-based analysis. You’re looking at cohorts, which are acquired by time period, and you’re looking at how they are performing for their first, second, third, fourth, and fifth month.
And this is really important, not only because now you’re looking at what is the contribution for my new users, but also how well are my older users continuing to contribute to my game,to my revenue this month, and hence to the LTV, which is something which we talked about in the last episode, what is my lifetime value? Without a cohort analysis, just understanding lifetime value itself is going to be very difficult.
Now, cohorts themselves can come in various other shapes, right? Usually, it is a time-based cohort that the most common and usually most useful. So a person or a cohort acquired in January versus February versus March and so forth. But you could also look at how am I doing with respect to things which happened?
For example, I did a particular promotion in a particular month or from a particular channel. What happened or how is that cohort behaving? For instance, you might also look at how a segment, how Android users versus iOS users behaving, and so forth.
And then you can split that cohort again time-wise cohort, but you’re not looking at just the iOS users or just the Android users and so forth. And this might give you a lot of insight into what happened. What you’ll usually find is that not all cohorts behave similarly.
You’ll find that suddenly one or two cohorts have a much higher retention, much higher lifetime value for instance. And that gives insights into what is it that the company did at that time? You might have done a particular channel, which actually did a lot of acquisition from there, or you might have changed some of the gameplay mechanics, which made sure that you had a very high retention and a very high monetization from that segment.
So the long and short of it is that looking at cohorts, not only just how is my revenue divided this month? How is the first month of my various cohorts doing? How is my retention doing over that and how are my longer-serving cohorts doing across the time period is really the only way you can get meaningful insights into the business.
Amit Somani 34:05
Absolutely. And whatever you said, I think almost all of it applies to B2B companies as well. I think gaming was just an example. So I think I would, and it’s not just from an investor point of view that you should do this.
I think even to run your own business to figure out what are the most attractive cohorts or the most attractive segments or the most attractive channels from where you’re acquiring these cohorts, that’ll be very indicative of where you should do your capital allocation or your intellectual bandwidth allocation to say, “Oh my God, users acquired on iPhone seem to kill it or vice versa, or user acquired from a particular channel they seem to do much better and so forth.”
Shripati Acharya 34:45
Absolutely. And one of the things we talked about last time for a SaaS business was paybacks. So one of the things we look for in when evaluating companies is if you look at various cohorts, how is the payback trending over time? So if suppose you actually had a payback of 12 months, is that actually becoming shorter?
And the way to figure that out is look at the revenue realization from each of your cohorts over a period of time. And if these cohorts are becoming steeper, meaning that if the y-axis is revenue and you’re actually going to a higher revenue earlier in the cycle, that’s indicative of a good business trend.
Amit Somani 35:20
Absolutely. And one of the metrics that I know both you and I are big fans of the dollar-based net retention. That will also come very easily from the cohort data because yes, there’s logo retention. I had 100 customers last year, I have 110 now, two churned, and so on and so forth.
But I could also have two churned. So I only have 98 customers, but the cumulative value of the dollars I’m getting now is 115% of what I was getting last year. So I’m actually more valuable to my customers either because of my product offering or my pricing power or increased number of users or licenses or whatever.
Shripati Acharya 35:55
Absolutely. I mean, we talked about churn last time, and it’s related to that, which is that if you look at your revenue cohort versus user cohort, you might actually get other insights.
So for instance, if you find that your revenue retention is increasing and you have 110%, 120% revenue retention, but if you look at your user retention and you see that it’s actually going down, what that means is that your customer mix is changing and it’s going towards a higher value customer.
Now, it is a great thing if that is what you intended to do, but not so great. If that is not something, which you plan to do and you want it to actually have a large number of customers.
Now what’s going on is that your average order value, which we’ll talk about in a bit, is increasing and your product probably is now moving up market and it appeals more to a certain segment than more than the previous, what you were doing earlier. So cohorts actually give you a very good indication of how things are moving in your business over time, which averages almost universally.
Amit Somani 37:00
Yes, we’ll talk about averages and medians later. There is one other scenario from what you just described, the last example. That could be the case that your dollar-based net retention is increasing, but your actual user or business retention is decreasing.
It could be that your aggressive marketing strategy is not working, so you’re just randomly acquiring customers because you’re trying to grow your way into your next kind of business milestone, but you’re not acquiring effectively.
And therefore, sometimes I find that that behavior tells you of the customers you don’t want to acquire because they’ll be rapidly churning or they’ll not be high retaining, or they’ll certainly not be high increasing in terms of their contribution to your business.
So I think the cohorts are absolutely the way to go and you should literally run your business on it, not just from a reporting perspective. Let’s use this as a segue to talk about the average order value. So there’s a lot of this confusion on ARPU first transaction value, et cetera, et cetera. Any thoughts on when you look at average order value or ARPU for whether it’s a SaaS company or anything else, how do you think about it? What are some of the challenges in terms of how people report these metrics?
Shripati Acharya 38:00
So when you’re thinking about for a SaaS company, the ACV, the annual contract value is the way you look at it. And annual contract value presumes first and foremost is that the contract is annual, which might or might not be the case.
And the second thing is that it becomes a little bit harder because in a lot of SaaS companies you have subscription, and you have transactions. So the total value is actually a blend of that really what you’re going to realize from that customer over the year.
So it is very important to both calculate and delineate very clearly what your average contract value, annual contract value is going to be for that customer.
Especially, since more and more businesses are now moving from subscription into a usage-based model because being driven by the customer side because they are not using, they don’t want to pay for it. It’s probably the first thing which comes to mind from a SaaS thing. Perhaps you can talk about from a more B2C perspective, how you think about AOVs.
Amit Somani 39:30
Yeah, no, I think I had just one more thing to kind of add a rejoinder to the B2B side, which is that sometimes for a certain class of B2B companies could be on mid-market or enterprise, there is a one-time setup fee or an onboarding fee or a services fee or whatever, which is not going to be a recurring revenue.
So certainly, when we are looking at it and again everything is aggregated and blended into one whole thing saying we are a $50,000 ACV, but when you double click on it, you’re like, “Well, $15,000 is onboarding and services for setup, which will probably never happen again.” Then really the annual contracted value is really $35,000. I’m not saying that the $15,000 is not relevant, but that revenue is a little blue dollars versus green dollars. So that’s one thing that…
Shripati Acharya 40:20
Yeah, I think that’s an excellent point, Amit. And what I would say where I found that useful is when entrepreneurs are able to say, “Wait, my setup or initial fee covers my CAC.
Amit Somani 40:30
Yes. Including the variable cost of the people providing that onboarding or whatever service plus, the variable cost.
Shripati Acharya 40:40
Correct. So I think that’s a very clean way of thinking. To me, that shows good understanding of running a business in a capital-efficient fashion because you go and acquire a customer for $1,000, but as you correctly say, all loaded in cost of getting that customer onboarded, the fee you’re charging the $1,000.
You can say, “Hey, look, I’m not losing money, so my payback is there as soon as the customer comes on board.” But as you correctly said, for calculating LTV for that customer, now you’ll have to understand the recurring nature of the revenues.
But Amit, when you’re looking at B2C businesses, the almost always companies will be providing promotions, they’ll be doing… They’ll be refunds, returns, and going on. So what is a clean and fair way of accounting that? Because almost any business with stars has to, we will face these kinds of things as the product matures and the value proposition kind of firms up.
Amit Somani 41:35
So typically, though, I mean there are different ways to treat it. I’m not an accountant, but certainly evaluating it from a business point of view or from an operating point of view, I would say I would cleanly separate all variable costs.
Just like you got revenue at the top of the kind of top line, you eliminate all the variable costs with respect to CAC, onboarding, setup, whatever it might be, payment gateway charges, et cetera. And then I would have another line item for refunds or promotions or whatever. So I included in the variable cost because it is a variable cost from a P&L perspective.
Rather than saying, let’s say you just to take a degenerate example, which is not that far off in certain fashion kind of verticals, you might have 30, 40, 50% return rates. So now I’m booking crazy revenue on the top line saying I’m at five million GMV, I’m at 50 million GMV, but $25 million of that comes back every month or whatever unit of frequency it is.
So therefore, your revenue is really not $50 million. I don’t think we covered the gross margin and so forth, but maybe we can talk about gross margin versus net margin versus contribution margin and the kind of-
Shripati Acharya 42:47
Yeah, this is a good time to perhaps talk about that.
Amit Somani 42:52
Yeah, so I would say that these days, especially since we are recording this in June 2023, people are not looking at GMV and top line and all that stuff. People are looking at, in fact, there was a little quirky quote going around that the gross margin is the new revenue.
So it’s not even your cost of goods and all that, which technically even by accounting standards is revenue. So I would say in the modern definition of a capital efficient business, I think you have to eliminate the cost of goods and obviously refunds and any of those top line kind of variable costs in your actual revenue.
Now, that may be a little bit harsh, but that is the reality because that revenue is never going to really come to you, right? There has no ability to hit your bottom line because it’s right off the top. So I would eliminate that and really look at it from that point of view.
In terms of just gross margin and contribution margin. Gross margin is just less cost of goods. Very simple. Whatever product or service you’re providing, that’s what it is. And then I get very annoyed with CM1, CM2, CM2.5, CM3, look, it’s very simple. Contribution margin is basically the contribution at a unit economics level per unit sold less all variable costs.
So like I said earlier, payment gateway goes from that, amortized refunds go from that, whatever a CAC goes from that, onboarding goes from that. And that is your real ability to figure out the economic viability of this business. Is there “juice” in this business or not? And I think all this notion of, “Well, CM1 doesn’t take that out, this one doesn’t.” I mean to the extreme level that I would even say there are certain fixed costs that might be amortizable on a variable basis.
We talked about it in the earlier episode SEO. So you might have a massive investment in SEO or doing video creation for your YouTube channel or your Insta Reels channel, which is amortized over time because you’re using that to build brand and that’s going to affect your CAC or hopefully reduce your CAC.
But that needs to be amortized. You can’t just say, “Well, I just have 10 people doing content. And that’s just like, that’s different. That’s a fixed cost.” No, it’s not fixed cost because if you’re trying to grow the business, you might have 50 people doing content, or if that’s not an effective strategy, you might have two people doing content and therefore there should be amortization of that.
Shripati Acharya 45:17
So would you rather that we actually stop using CM1, CM2, CM3s, and so forth?
Amit Somani 45:24
100%. I would just say, look, there is your top-line revenue, there’s your gross margin, there’s your contribution margin, which is less of all variable costs. And if you’re a higher bar startup, I would say even the fixed costs that are naturally amortisable, of course, like office space and software engineers and so on, may be harder to amortize over variable transactions ‘cause those are really assets you’re building for the long term.
So just keep it simple, right? Real revenue, real variable, I mean revenue less variable cost, and then eventually, hopefully someday you have EBITDA and PAT, right? So it’s like profit after everything.
Shripati Acharya 45:48
And for SaaS companies, that would almost definitely include the cloud costs.
Amit Somani 46:03
Of course. Absolutely. Significant variable cost is cloud cost, right? So absolutely has to include that for sure.
Shripati Acharya 46:10
And I would add that initially in a lot of companies have credits and things like that, so those costs are hidden initially. But the right way to think about it is that what happens when those promotions and credits and everything else expire, and then what is your margin?
And actually, you would be surprised that many SaaS companies we see do not have that much of a very high gross margin because the cloud costs and now increasingly the generative AI, ChatGPT, and all these costs are going to be non-trivial part of delivering the value and they all need to be properly accounted for.
So let’s move to a related topic, which is averages and medians. Yes. So I know that that’s something which you really have a lot of thoughts on, so let’s hear it.
Amit Somani 46:58
So I think, again, we’ve covered this in both the episodes that don’t look at aggregates, look at cohorts, don’t look at total revenue, look at revenue by segment, don’t look at total CAC, et cetera. This should be leading the witness to, or the jury in this case to what the outcome is. Averages are extremely misleading for any metric.
We looked at a few in the past and give a few more examples. And the reason they’re misleading for any metric, it could be engagement time, it could be session time, it could be whatever other metric, which are very relevant, may not be purely from a P&L perspective, but from a product success perspective.
Because they hide the behavior of the power users versus the naive or new onboarded users versus your typical user. Because average and medians are synonymous when the distribution is uniform or a more normal distribution. But in early-stage startups, probably even later-stage or mid-stage startups, distribution is hardly but uniform. There are a certain set of customers that will absolutely love your product and double down on it and spend hours.
We have a couple of gaming companies in our portfolio where two, three, 5% of users will dominate… in terms of usage, number of games, times, in-app purchases, et cetera. But if you did an average of somebody who spent a dollar per week or per month on a game and somebody else who spent $500, real example from real company. But you do the average of those two users, $250.50 is extremely misleading.
So I would say you should, and medians of course will work better with skewed distributions, which is the case. Also, I would basically appeal to the founders that it’s also a good way to run your business. So you need to know where the money is coming from, where the engagement is coming from, where the retention is coming from.
So even beyond the median, which is the highest frequency sort of recurring data point. I would say look at the 90th percentile user or the business that you’re serving and the 10th percentile user that gives you a true picture.
So it’s almost like “three medians” or if I drew a graph, that at the 90th percentile, which is the highest engaged users or businesses or whatever, this is my retention, this is my revenue, this is my AOV or ACV, et cetera at the newest, and I’m talking newest in the sense of either could be onboarding, but like you said, on a time basis or could be by how much they’re producing for me.
So I could have a gaming user who’s been with me for a year and has only ever spent $5 or 50 cents, still a very loyal user. But there in that 10th percentile if you do the cohort, not by time. So I would definitely say that those are a couple of things that I would really encourage people to do is look at medians, look at 90th percentile, look at 10 percentile. And in fact, run your business on it ‘cause that lead you to the ones that you want.
Shripati Acharya 50:00
Actually, one of the terms which is used in SaaS metrics is this thing called the smile curve. Which if you draw a curve wherein on the… I suppose there’s a product which is being used on a daily, weekly basis, and on the x-axis, you plot the number of the users who are coming once, twice, three times, four times, so many times. Just think of a game, if people who come daily would be clocking at the 30 times a month, they’re coming.
And the y-axis is of course the number of people and really interesting companies, SaaS companies have this smile shape to this curve wherein you have a large number of folks who are coming for just one or two or three days in a month so that the infrequent users in your distribution. And then it goes and as the number of days goes increases, it goes down.
But there are a small fraction of power users who are coming at the 28, 29, 30 days a month and it again trends up. So that trending up is a very healthy sign because what it indicates is that the focus on acquiring that kind of customer becomes the focus of the company, because understanding what is the behavior of that customer, which channel are they coming from? What are the things which can be done in the product to incentivize that kind of behavior becomes like the… Gives us an insight so that it really leads to a high-value business.
Amit Somani 51:28
Absolutely. And I think the exact same thing applies to gaming startups, content startups, any kind of B2C engagement startups, because you’ll acquire a lot more at the top of the funnel and then you’ll have some PMF for the ones that really liking it. And then really the ones on the other end of the smile curve will be the ones who are the highly retained, highly engaged, highly paying customers. So I think that behavior is quite common to B2C as well.
Shripati Acharya 51:54
Exactly. And how that kind of curve trends over time gives a very good insight into how the business health is.
Amit Somani 52:00
Yeah, that’s another very good point. Oftentimes entrepreneurs will critique this notion of, “Well, all this cohort and all this is fine, but really my product was very shabby six months ago or 12 months ago. Now it’s getting much better, but you will see the trending of the entire curve going up.
So when we are really looking at it as well, we are not just looking at an absolute data point saying, “Oh, month three must be this, and month eight must be the or six must be this.” We’re saying, “Is the overall curve trending up?
So is your onboarding getting better? Is your acquisition strategy getting better? Is your day seven, day 30 retention or for a B2B SaaS company it might be quarterly revenue plan or whatever.” Beautiful point that the directionality of the curve is what we are looking at, not often, not the absolute number for any given month or quarter.
Shripati Acharya 52:50
Well, I think we kind of covered, we covered cohorts, we covered aggregates, we covered medians and averages, and AOVs. So which is where we would like to wrap up this episode, but one thing I will leave our listeners with and especially entrepreneurs. Is that it’s very important to visualize the data because if you don’t visualize the data, and by that, I mean actually drawing the trend lines, drawing the right kind of curves, so that you can draw meaningful inferences in a quick and efficient fashion.
It is very difficult because you can get all kinds of data and it’s just noise if you cannot actually make meaningful inferences out of it.
Amit Somani 53:35
And if I may add my closing comments before we wrap this up, Shripati, is that in addition to the visualizing look, we talked about a lot of different data, a lot of bad math about different metrics, et cetera. I think you also want to quickly zone in on the two to three or maybe three to five metrics that really matter because you don’t have the ability as a young team to focus on 18 different things.
You should measure everything. You should visualize everything. You should capture everything. Because you may come back with an insight later and like, “Oh, we never measured that. No, I don’t have ability to figure that out.” But then really pick the three to five that are defining to your business, this kind of customer, this kind of cohort, this kind of ACV leads to this kind of retention or this kind of churn.
Everything else is noise and you’re seeking that highly repeatable, exciting customer cohort or behavior that you’re going to try to tap into. So I would say measure everything, but focus on a handful of few in terms of driving the business.
All right. With that, we’d like to call it a wrap for this episode of the Prime Venture Partners Podcast. If you like the episode, please write into us. If you’d like to hear more about stuff like this or any other metrics that are your peeves, please do send it to us on our LinkedIn or our Twitter handle. Thank you.
Enjoyed the podcast? Please consider leaving a review and subscribe so you don’t miss future episodes.
Follow Prime Venture Partners:
If you believe you are building the next big thing, let’s make it happen.