The Art of Scaling Games: When to scale your game?

essentially once you plug all of that core data in we'll just tell you like where are you in terms

of fundability and then we'll break it down in terms of like why why what what key indicators tell us give us confidence this is one that's unreleased um this is this is a debut for the two and a half gamers uh podcast but our benchmarking our benchmarking tool yeah so like I said, we've got a bunch of, I think

right now, maybe like over

like 200 over 200 or 300 cohorts for 2024 that we've looked at like across all these different games it's 4 a.m and we're rolling the dice but day drops knowledge made of gold and ice felix with ads making those coins rise jackup designs worlds chasing the sky we're The two and a half gamers, the midnight crew, talking UA adverts and game design too. Matej, Felix, Shakug bringing the insight, we're rocking those vibes till the early daylight.
Matej, UA, master eyes on the prize, tracking data through the cyberspace skies. Felix acts dollars like a wizard in disguise, Jackups crafting realms, lift us to the highs.
Two and a half gamers talking smack. Slow hockey sick, got your back.
Ads are beautiful, they light the way. Click it fast, don't delay.
Uh-huh. Uh-huh.
Uh-huh. Uh-huh.
Uh-huh. Uh-huh, uh-huh Ritsky from PVX Partners.
Welcome to the show. Can you give us a little bit of introduction about yourself and the company? And then we're going to talk about what is the topic? Yeah, sure.
Hi, Matej, Felix. It's really great to be on the show.
As Matej mentioned before, my name is Ritsky. I recently founded, along with my two co-founders, a company called PVX Partners.
It's a platform trying to help gaming and consumer companies scale. Our flagship product is a credit line that helps developers fund their marketing in a way that is very flexible to help them scale and is also risk participative in the downside case where everything goes wrong and the campaign fails.
So I've spent the last 12 years working in a mixture of marketing and finance roles, most recently at the Indian real money gaming company called Mobile Perimeter League or NPL, where I focused on international growth and corporate development. Actually, Matye, I think we met for the first time when I was working there and we were doing diligence on a potential target together.
So it's good to have it come full circle exactly nice okay good so uh before we go into the topic okay let's let's talk about the topic is when to scale your game or when is the right time to scale your game because this topic and this question comes up a lot in the different discussions we talk about it on the podcast when we're reviewing different games. And I have an answer, which is a little bit vague, but kind of works because I am in the gaming industry for 11 years.
When I see the data, I know what to expect, let's say.

And I'm looking at different KPIs, but what are the KPIs you are looking at, Ritzke? And how are you kind of working with different companies? Because, I mean, I assume it's a very long and interesting process. Yeah, for sure.
No, I'm happy to apply on this. And I know that you and everyone has like, you know, the key sanity metrics that they focus on to kind of give them conviction to scale up.
I think the way that I will be presenting here is more, you can call it like high level picture. But I think it's a really important view for any decision maker or capital allocator of the business

to understand very intimately

before they go and ramp up

and spend millions of dollars on UA.

So I can get to how we work

in a little bit later.

Yeah, of course.

I actually did put together

a few simple slides

if you don't mind going through them.

Yes, that'll be awesome. Because we like visuals.
And visuals, that's important. And people love it as well.
Yeah. Let's get you some visuals.
Nice. Let me know if you can see it.
Yes. All good.
We see. We see.
We see. Perfect.
This is awesome. Yeah.
So um like i said i preface this this is a high level high level picture and um and uh i hope i hope it's not too basic but um we'll ask questions we'll ask questions okay okay so um i thought it'd be good to bring up a familiar case study that you guys covered recently in hexasort by magic tab episode by the way um so i'm using app magic's you know revenue trend chart here to illustrate the ramp up uh in daily iap revenues but on the episode felix you discovered or you you kind of like brought brought up that calculated yeah you calculated you calculated that it was actually like only 30 of the total revenue was being generated by the IAP. So that was kind of like the mind blowing moment where I was like, okay, these guys are killing it.
And so I think I'm taking a rough estimate of 30% IAP across the timeline, probably not accurate, but let's just for simplicity sake, they started off like the only way to see this is that it's a banger, right? So in January 2024, they started with like $10,000 of revenue a day, and then ramping up for the next three quarters to half a million dollars a day. And I think, looking back, like looking back, it's pretty obvious to highlight that this is what everyone wants.
But in order to produce this outcome, you really need to be in a ready state all the way back in January 2024 to have built the conviction and be financially prepared to scale it in this way. So I think another way to frame the question of the day is really how can you build conviction and be financially prepared to scale your game? Yeah, nobody's ready.
Well, not nobody, but a lot of companies, they're not ready. And they say, oh, let's scale.
I had this situation when we discussed, let's scale because it looks great in terms of the ROI, in terms of the numbers, CPIs, whatever else. And then we start scaling.
And then after six months, like, hey, well, we need to scale down because our cash flow is almost dead. We don't have money for salaries afterwards.
I mean, the question is, right? Like everyone can scale, but the issue is scaling profitably, right? Like that's the problem, right? Like I could scale anything, but you know, if I had a billion dollars.

Actually, you'd be surprised at how many companies we see that actually are scaling, but then like really shouldn't be scaling at that point in time. Or maybe they got a little bit too excited by the early metrics and then they've like rolled that back.
But it's really important to kind of look at it like a common set of metrics can kind of universally accept as, okay, this is us doing the right thing versus not. And obviously, it starts with the soft launch, right? When you're in that soft launch phase, you know, in mobile gaming in particular, you're fortunate enough to have the ability to track, test, and iterate the game quite quickly.

So it makes complete sense to take advantage of that and soft launch and then test your acquisition and monetization metrics well before you go out and burn a bunch of money. And the goal here, I think, I would argue, is to be able to track your cohort performance and achieve profitability and consistency.
So during that soft launch period,

like your operational team should have quite a long list of like checklists

that they should be validating. I'm sure you could do a whole podcast on this entire topic, but essentially you've got all the standard sanity checks around economy balancing, retention and monetization benchmarking, like UA marketability testing.
And I'm not an ad mona expert, Felix, but I assume you'd look at viewer rates and completion rates and impressions per DAO. Impressions per DAO and ad viewer rates.
Those are the ones. Yeah.
So all of these are super critical and you should definitely do the honorable thing in making sure you go through this in like a meticulous manner. But like the outcome we're looking for, if all of this goes swimmingly, is cohort profitability and consistency, like I mentioned.
And here most companies fail because there's like, oh, wait, we don't need to be in soft launch for six to nine months. Everything is working.
It's like, yeah, well, okay. But I would like to see how the courts improve over time and progress over time.
And, and to have that, I'm then afterwards, I'm able to build a global launch strategy or global launch plan, how much money we want to spend because then everybody's like, well, tell me what's, what should, what's the payback period we should aim for. I was like, I mean, come on, guys.
This is a discussion you should have. And then we look at the data and then we say, oh, look, okay, my bank account looks like this.
We need to get the money back sooner. Okay, based on that, I'm able to build a global launch strategy and how much money we should spend.
Because if that LTV evolution tells a lot, and if you don't have those cohorts at least like you are I mean five to ten that's that's ideal scenario you know companies are not that patient in soft launch to have this so yeah but but yeah for for the soft launch that's that's the kind of where should companies are aimed for, I think there's a couple of things I would also highlight in that soft launch period, which are like super critical. Like one is getting your instrumentation right.
Like ensuring that when you're looking, when every like your marketing team, your product team, the business and management team are looking at the metrics, they're looking at the same metrics and that's actually the source of truth.

You don't want to be looking at your mediation dashboard or your Facebook ads manager dashboard and then it's actually the wrong thing and you're optimizing towards different metrics. I think that's super critical.
The other thing is that I think if you have enough budget and the luxury of testing, you should test quite broad in like different geos,

different channels and the luxury of testing, you should test quite broad in different geos, different channels, and a breadth of creatives as well. Because that'll give you a sense of how scalable this can be in the immediate next steps.
And then try to get some benchmarks, I think, is also another one. Right.
Like back at MPL, I would I would be calling my relationship manager at Meta and like Google and be like, look, trying to get a sense of where the CACs are at. And then, and then sensor tower for some of the monetization retention metrics, right? Yeah, absolutely.
Yeah. So I wanted to dive into a little bit of basics if, if I may, um, and clarify what I mean by core profitability.
So like if you take an arbitrary period of time, like a month or week or day and tagged all of the new users during that period, and then measure the lifetime value generated over those users, um, over time, you'll have a bunch of cohorts that you can track, right? And compare. And the way that we would recommend recording the value that's generated by them is essentially trying to get as close to cash as you can.
So whatever that your users pay or what you're getting through your IA revenue minus the processing fees, like platform fees on Google, Apple, Stripe, RevenueCat, wherever you're getting your revenue. And then you take this as your net revenue and then blend across organic and paid users and then divide it by your marketing dollars.
I would even argue throw in like the creative costs to the relevant campaigns on there as well, just to make sure you have like a really like fair, conservative number here. But essentially, you know, once you do this over time, if your cohort is good, it'll reach 100% net ROAS.
And that effectively means you've paid back your marketing investment. And then beyond that...
Sorry, just to stop here for a question, like, Mati, since I work on the Admon side, like, out of the clients you've helped in the, what is it, 11 years, how many people actually do this? Well, not a lot, but then they quickly realize you can't look at gross data because then, oh, wait a second, this is not exactly the money that comes to my bank account. So the net ROAS, and it's easy to calculate this

based on the platform fees,

but then there is, like you said,

like the cost of creative production,

then there is a team,

and then there is like a lot of different things.

And suddenly you are very far from profitability.

And then suddenly you're very surprised that money that comes into the bank account is not enough. Yeah.
Yeah. And you want to have the clearest picture of whether or not you're profitable or not.
So when you put a dollar in, you're going to get more than a dollar out, right? So that's what this curve represents. Once you hit 100% ROAS, that, that's your profit margin.
Right. And I think the fundamental thing to understand here is that if you knew for a fact that your marketing dollars would return.
Sorry, one other last aspect here is that if you track this cohort over time, and then you track your user, your payer retention, and that basically goes to zero at some point in time, your users churn out completely.

This asymptote. If you track this cohort over time, and then you track your user, your payer retention, and that basically goes to zero at some point in time, your users churn out completely.

This asymptotes, like horizontally asymptotes towards a figure.

And we would call that a predicted terminal ROAS figure.

So it's important to, like, there are some tools that can help you kind of predict this.

I know, Matias, you maybe use Cohort.

They're really good at this.

We have a tool as well.

But it's important to kind of estimate where this kind of lands up.

But the fundamental thing to understand here is that

if you knew for a fact that your marketing dollars would return

a terminal ROAS number quite consistently,

then you're essentially acquiring an asset. It's similar to buying a government bond.
It pays you an annuity. You're acquiring a group of users that will pay you over time and quite predictably.
So the metagame that all of the biggest consumer businesses play in corporate finance is stacking as many of these cohorts as possible for as long as possible. That's the name of the game.
And I think that's quite important to know. That sounds amazing.
What's the error margin on the predicted terminal ROAS? It really depends on how much data you have on it, right? So the larger your sample size more the more you can do um i'll take you through kind of how we do it at at pvx later on but essentially if you if you if you have enough data that correlates to your game like a certain version of your game um and also category like other games in your category or genre, you can kind of build a data science

model that does this quite accurately. And if you can do it very accurately, it's very, very like useful information.
So if I asked it a different way, per $100,000 of UA spent, how much does the error margin of the predicted terminal ROAS decrease? that's a great question um it's it i don't think it's based on how much you spend. It's based on how many sample sets of cohorts you have.
And there's a trick with this, right? The standard cohort kind of duration is a month, but you can technically slice this down to a week, weekly cohort or a daily cohort, right? And suddenly if you have three, if you've been doing it for a year and you're doing this on a daily basis, you have 365 daily cohorts. Yeah.
And you can measure the variance and the statistical variance between them and get to a pretty accurate predicted terminal ROAS number, I would say. Okay.
Yeah. Any thoughts, Matias? As complicated as it gets.
No, like this is the volume play, right? So you need to have a certain amount of data to be able to say statistically, significantly, like, okay, this is what we're looking at and this is basically what you can expect, which is pretty normal. But PVX partner puts his money where its mouth is, right? So I guess there's a user number, a cohort number that you guys are looking for before you're like, hey, I'm quite confident that we're not going to have an error margin over X, basically.
Yeah. So typically we like to see businesses with nine to 12 months of data, and we can be quite confident in that.
But to give a slight correction, we are not in the business of predicting the predicted terminal ROAS because our business, we don't take the upside up all the way here. We're betting on you getting to 100% net ROAS and enough money beyond that to pay our interest.
So there's a margin of safety kind of element here. Whether you get 180% ROAS versus 140% ROAS makes very little difference from our upside as a business.
But it makes a lot of difference for the company. And we want to make sure that the company is trying to spend optimally so that they can make as much profit as possible.
Yeah. So moving on, like the second thing, as I mentioned, is core profitability, but the other thing is consistency.
And this is a really kind of like distilled way to look at it that you can eyeball, but we're not looking. So the point here is that we're not looking for the latest and greatest cohort um we're trying to get comfort on how volatile the future cohort outcomes could be by looking back historically so um take for example on this left chart this is one game one and then the right chart is game two let's say this top curve is the latest cohort of game one.
It's the highest profitable cohort amongst the two games. But I would much prefer, I would have more conviction in investing if on the right chart, this top curve was the latest cohort.
Because there's a lot more consistency in this right chart versus on the left right so a lot of gaming companies that we deal with they're like i've got a banger cohort i'm ready to scale yeah but i would say hold on one second let's let's let's see whether or not that plays out in a few more cohorts and then whether we can rely on that otherwise it's just going to go wonky and you might lose a lot of money. Yeah, because this on the right-hand side is stability, basically.
And this graph feels like the company or the game knows what they're doing. On the left-hand side, it's like, oh, well, we try a million different things and see what sticks.
It's like, it could be random. Also, one good cohort doesn't mean anything.
You need to have, yeah, like multiple good cohorts to be able to say, oh, wait a second, we have something here. How do you link this basically? I guess it's a question for you, Matje.
Like when you're doing UA and you're always like in our chat group saying like, oh yeah, look at this creative. We just got CPIs down by 50%.
Well, if you have a banger creative, or if you start doing, you know, lying in the creatives, which helps a lot, right? How do you get that into this? Because that could drastically reduce it, but it could also be fake information or not repeatable. It's misleading.
It's good for, let's say, five or six months at the best. And then when the winner creative dies, then, well, you're the square one, basically.
So if you have a killer creative, that's a great start. But then you need to be able to produce multiple of these.
Insistent. Yeah.
Which then is like, that's the pressure on the creative and the UA team to come up with all these ideas and test them. And before we scale, and that's what I'm always talking about in the reviews, you need to have the creative depth as well.
Multiple different concepts that work across different channels. I was in the position a few weeks ago where we discussed scaling on Uploadwin or starting Uploadwin campaigns.
For this, you can't have only one creative winner because then you start scaling. And then after three weeks, you will see the drop and the decrease in performance and even the spend.
So I was pushing for more testing on the creative side to be able to get at least three or four. So one works, and then you start refreshing to creatives and just putting this into the campaign so you prolong the life cycle of the campaign.
That's really important. Yeah.
And I would argue also that that winner or hero kind of campaign or creative, you should always be testing them like in parallel with other new concepts, because then you'll identify better like the next one that's going to take over. Right.
Yeah. And I think a lot of folks are just super happy with their hero creative.
And they're like, I'm going to ride this out until it essentially dies. And then you have no backup or lead into the next one.

It's like wishful thinking, right?

Yeah.

No, because people get excited and they're like, oh, wow, this is great. Now we should push.

It's like, yeah, we can push, but not like 100% more or 10x more or whatever, slowly,

because you don't have the next winner and we are not testing enough. We're not producing enough.
So it's not the right time. Yeah.
Another way to look at the consistency one is like, if you have enough cohorts, you can take a look at this statistically. And essentially, if you take the predicted terminal ROAS of each of your cohorts and plot them across a normal distribution curve, if the curve fits pretty well within a given band of spends, then you might be able to build conviction by taking a look at what the net ROAS would be at negative two standard deviation.
So back to statistics 101, essentially at negative two standard deviations from the mean, essentially 98%, if you were to produce another cohort, 98% of the time, it's going to be above the negative two standard deviation. So if your negative two standard deviation is already profitable, then the next one is probably profitable, right? And this is if you have the luxury of sample size and oftentimes at the very beginning, you don't.
But but yeah, you can you can you can take a look at this like a year or a year after and be like, yeah, I did a great job. So I guess the thing is, like, I would I would urge people to go and figure out how profitable and consistent your cohorts are and see how much value you would expect to accrue over the lifetime.

And if you haven't reached 100% ROAS on any of your cohorts,

I think you should be heads down in trying to figure out how to get there before scaling up.

Yeah.

I mean, if your cohorts are not profitable,

then I'm not sure if you should spend that much money or money at all. That's the most important part.
Yeah. But going back to hexasort, at the inflection point prior to scaling up, essentially this is what they probably like, they're probably thinking, I've got good cohort metrics.

They look consistent.

I should try to put more money into this cash printing machine before, you know, a competitor builds something that will eat into my market share.

Or CPMs inflate, maybe because of competition or maybe because the platform itself is just inflating and making it harder for me to defend my profit margins, right?

So my argument would be that if you if you build conviction and you you are able to try to scale, you should like now is better than later to do that. And yeah, you never know what else will happen.
So I think at this point, a team should suddenly be hyper focused at, you know, how can I scale further? And the first thing I think they're going to ask themselves is, how are they going to fund this growth if they don't have it, right? Yeah. It's usually the UA managers are not there.
They're like, oh, let's scale. And then the C-level people always, oh, you know what? Let's check if we have enough money to actually scale.
Sometimes you have people are not business oriented enough to understand there is not unlimited amount of money on your bank account. Yeah.
And I think it's worthwhile going in and trying to understand what are the actual sources of capital that are reasonable to draw from. So I wanted to just do a very simple walkthrough of the common growth strategies to finance growth finance growth right so if you're a self-publishing game and you are showing awesome metrics i would argue that you have three or four options the first option is like you mentioned how much bank how much cash is in my bank and is in your balance sheet right yeah and this is technically the lowest direct cost of capital.
But the problem is, as you pointed out, as you scale up, your payback periods are inevitably also going to be pushed back, right? If you're spending $100,000 a month today and you're getting three months payback, you can anticipate that if you're spending a million dollars, it's going to be more than three months payback. And what this means is that more and more of your equity or your balance sheet is going to be tied up in the marketing payback cycle.
And the best case scenario here is that you do scale up to a million, $10 million a month, but you'll eventually reach a level of spending where it's completely depleted your bank account. And by the way, this also means less money for new product development, less money for hiring, less money for strategic initiatives.
And that's not great for the long-term options for the company as well. So yeah, people still forget that you spent, the money comes into your account, let's say in three or six months.

But then there is also another 45 days from getting the money from Apple and Google, which very few people count with.

And that's the problematic part.

Add revenue, net 30, net 60, depending on some networks. Exactly.
Yeah. There's definitely a working capital kind of like dragged over there, but it's exacerbated when you're trying to spend double or triple the marketing dollars as well.
And so a lot of people run to venture capital, right? That's the next one over here. And raising equity capital from VCs has a big pro in that the money is not guaranteed back to the investors.
You take money from VC, they're not like asking for it back right away. There's no fixed repayment schedule.
It's highly flexible capital. And if you have good investors, they might be able to help you in more ways

than just capital, which they'll happily

do because they're aligned with you

on the equity side.

The issue with VC capital

is that it's dilutive.

It's equity dilutive, which means that you

now have less ownership in your company

and you have new investors to report

to.

AKA selling your soul a little bit yeah i've heard you refer to that way but it's okay selling your soul to nice people is yeah that's fine yeah i mean i mean joking yeah i mean there's also a risk to it right like if you look at for example uh grand games is a good example now they spent six months raising their last round right and i know from speaking to a lot of ceos that basically people in the industry took notice because basically their metrics were so good that was like shooting off a flare gun and i know a lot of companies that we've reviewed have taken a lot of creative liberties and inspiration from that game right yeah so those six months probably cost it quite much in terms of competitors i guess yeah yeah it takes time to raise equity capital like yeah and if you really are gung-ho on doing that my advice would be to preempt your investors and get them excited about your early metrics earlier on right get them get them get them hooked before you really need to ramp up. Yeah, because that would be the right.
Yeah. That's already too late to start around, to start that discussion.
Yeah. The next one is debt, right? So there are a lot of debt instruments and I think it's, and you, that you can use to explore to fund growth.
And the biggest, I know it's kind of like a bad word sometimes because a lot of people think of it as like kind of a bad thing. But I think the biggest benefit here is that debt is a capped instrument.
Let's be clear. You don't have to give the lender such a significant amount of your upside like you would for an equity investor.
And the price is agreed upon up front. It's usually transparent.
Those are all great things, right? But the main issue with debt and why people don't really consider it, especially as I've seen in the gaming sector, is it has recourse. And what that means is that if you miss a payment, just like your mortgage, and you can go into default, which means they can take your company, like they can take your house or whatever you've decided to secure.
And that's not like, is that a risk worth taking for you? Like, do you want to go there? But yeah, what are your thoughts on that? Well, probably also issues or debt don't really know how the gaming industry works. Like having someone explain to them that's sitting in a office like, oh yeah, we just had a bad cohort and we changed our payback period from 90 to 180.
They don't care. No chance they'll understand.
I don't think they understand what a cohort is, right? Exactly. Yeah, no way.
episode for like we'll go down the conversation yeah but i think there's one thing with debt that a lot of people don't anticipate a lot of founders don't anticipate the covenants and so if you're not familiar with what covenants are they're basically conditions or restrictions on the borrower that are designed to help the lender feel good about your financial standing. And the most common covenant is like an EBITDA covenant, which is like, hey, you can borrow this money, but your EBITDA cannot fall below a certain amount, which is completely paradoxical to the point of growth, right? Like going back to my first first cohort profitability slide like if you knew for a fact that your cohort would produce 200 percent ROAS in two years like who cares if you're negative negative EBITDA this year if you're going to be 200 percent the next year right like it's absolutely the right thing to do for the company because you're going to make more money than if you didn't spend that money.

Sounds so European. Yeah, but the bank is not going to see it that way.
They're going to see it as you're breaching EBITDA governments and that means that you're not financially stable and I'm going to pull the plug. right

and finally we

arrive at

cohort financing

which is kind of

definitely the

one that people

understand the

least and I'm going to pull the plug. And finally, we arrive at cohort financing, which is kind of definitely the one that people understand the least and is the least common, I would say.
So basically, cohort financing is a form of credit that's generally only secured on your cohorts, the cohorts that are being financed. So at PVX, we offer cohort financing for user acquisition.
So the way that it works is we would lend, say like 50% of your marketing budget, your monthly marketing budget. And then we would take a 50% net revenue share from the ROAS that it produces until we get our money back and plus a little bit of interest.
And because it's structured, so, okay. So the, the way to think about it is it's about like 10, 10 to 14% annualized cash costs.
So let's say if it was 12% for easy math, annualized cash cost, and your paybacks were three months, it would be a 3% cost.

If it was six month payback, it would be 6% cost.

It depends on how long the cohort actually lasts.

Yeah.

So if, yeah, if you make money in one month, it's 1% for you.

Exactly.

Okay.

Exactly.

And I can lend you $10 million and it'll be like that 1% for that. Yeah.

But then as you scale, and we discussed this, that your payback prolongs

Thank you. and it'll be like that 1% for that.
Yeah. But then as you scale, and we discussed this, that your payback prolongs, but you're scaling, you're spending more money.
So, I mean, then you go from three months to six months and then you make 6%. I mean, that's kind of fair.
Yeah, I mean, I guess what you're monetizing here is risk, right? uh for me to understand better like what's the average interest actual cost in percentage terms of your current clients because i guess it's different that's more i can give you an average but it's not yeah it's not a really indicative number because yeah it's i've got clients that pay back in two months and i've got clients that pay back in 14 months so if i average that out it out, and especially if you weight the amount of spends, it's going to come up to a

not a very...

I'm just curious if you want to shout it out.

Like maybe 5%

to 7%, something like that.

Yeah.

Because look, you work with different gaming

companies and different games have very different

payback periods. There you go.
Yeah. And what this buys you, well, okay, one other important thing to say is because it's structured this way, it's a non-recourse instrument.
Which means if the cohort fails, like the cohort never reaches 100%, let's say it only reaches ever 80%. I'm only going to get 50% get 50 of 80 yeah right and and the company doesn't owe me that remainder 20 they just owe me whatever the cohort produces so in that way we're taking risk alongside the company because we're we're we're um we're in the cohorts with them and that's why you don't work with everyone actually, you know, like you want to make money.
I mean, you still need to pay bills and you still need to also pay yourself and not everybody. It's just, that's what people don't really understand.
Sounds like a really good deal because, you know, the US interest rate is what, four and a half right now so if your average is six that's quite good yeah and yeah so like again um if you were comparing us to a bank it's going to be at a premium the interest yeah but that bank is is a recourse product right this is an unsecured product that's that's why you pay that premium and the reason why we frame this as a scaling up product like a product for company who's ready to scale is like a bank is not going to touch you when you're trying to look for to double or triple your spends right they think that that's like too unstable but we're we understand the cohort mass and we want to take that risk with you do you mind me just asking a basic question here because you still understand like so basically if the risk-free rate is four and a half and you're offering six like why not use invested in treasury bonds take all the risk away um well to be to be to be honest the the fair that that four and a half percent is uh annualized apr right so if you were to annualize, you would compare that to a 12% or the 10 to 14% that I mentioned to you. So it does come at a premium.
Devil's in the details. There we go.
Yeah. No, I'm not trying to market myself to be the cheapest instrument, but that's the fair comparison.
Yeah, yeah, yeah. But for a founder...
Yeah, there's also, I'm pretty sure, yeah, there's also not only pros, but cons. And I think what you mentioned is that you need to have a lot of data to be able to actually provide this type of financing.
Yeah, yeah. So typically, like I mentioned, 9 to 12 months of data is what our sweet spot is yeah however we are getting better uh at modeling early stage cohort cohort uh chord so basically as our database scales are underwriting database scales we understand specific genres that like for example we have about like 30 or 40 merge games in our database and i've and i and i've taken a look at all of their historical core data yeah and um you get a little bit better at at predicting kind of what each one is and we do like retrospective testing on to see whether or not our predictions are correct we're not like perfect there yet but like i mean i wouldn't say that we're at a at a level where we're there's the confidence level is is high is high enough.
It should definitely be higher than 90% if I can't afford a loss. But we will get there as we get more and more data.
And because of that, we are launching a new product. It's called PBX Early, which basically can underwrite you with three months of data.
Yeah. With only three months of data, yeah.

What's the ceiling usually in terms of financing

that you can go up to or minimum?

Because I guess you need to make it worth your time as well, right?

Minimum and maximum?

Yeah, we need to make it worth our time.

And we need the developer or the publisher

to have tried enough kind of different kind of campaigns

and channels for us to understand the stability better.

So typically it's around $150,000 per month in spends. The PDX early product is lower.
It'll be like $50,000, but our flagship product is about $150,000 minimum spends. We have a very close relationship with our investor, which is General Catalyst.
General Catalyst is a San Francisco-based VC fund. And we focus on, they invest in all sorts of businesses and they typically go for relatively large scale deals because it's a multi-billion dollar fund.
But we focus, our sweet spot is really like anyone who's spending $2 million a month or less. So that's the starting range.
But as they scale to $5 million a month, $10 million, $50 million a month, we collaborate with general catalyst to, to, to, to facilitate them. So basically sky's the limit.
Yeah. They, they, they fund companies that are spending, you know, you know, over 20 million, 20 million a month easily.
Yeah. Nice.
Yeah. That's, that's kind of all I had with about like the the uh you know when when i think it's it's appropriate to scale or when it's a good time to scale um but i think you you had also asked like what does the process look like or sorry what was your question yes like what does the the process look like and how do you actually go from working with companies to to fund them and yeah like what are you looking at because that's the meat yeah yeah like me and matthew come to you we're two 20 year olds just out of university and we have a we have a merge game and uh we're excited yeah but yeah but we are spending 200k per day.
Per day. Per month.
Yeah, per month. Per day, that would be interesting.
Yeah, no. So we have like a proprietary intelligence like database called Lambda.
And essentially what this does is it basically what we just discussed in terms of cohorts. If you plug in your MMP, like if you use like AppSlyer or Singular or Adjust or whatever, or even if you want to plug in your like historical transaction ledgers from like an S3 bucket or a GCP bucket, we can import that data onto here and basically visualize all of the cohort information that you would ever want to see.
So the cohort analysis is what we typically look at. So we would plot your data and we would, by the way, this is just composite data from a couple of companies that we've taken a look at, like maybe four or five.
So the numbers might not make a lot of sense, but the curves essentially... Blurred out for the European users or watchers.
okay okay okay so so so this is this

composition but the curves essentially... Blurred out for the European users or watchers.

Okay, okay.

All good.

Okay.

So this composite data shows that all of these cohorts are kind of projecting towards profitability

at different payback rates,

like between five to eight months.

And this is validated.

This is NetROAS. This is validated on their actual transactions and their marketing spends.
You can see kind of the shorter lines over here are kind of degraded compared to like the longer lines, which means there's some core degradation and the payback periods are probably going to be longer. And another way to kind of look at this is on a month-on-month change basis.
So this is basically incremental net ROAS per month. So at month six, this is around 10% incremental versus month five.
And what this means is basically this thing is printing money. Every month it's returning 10% even after it pays back, which is a good sign for your profit margins.
And you'll see that this kind of asymptotes at a certain level. I mean, this is outstanding, actually, but usually we would say it's good if it's above 5% per month, at least for the several months after they've paid back.
Yeah. I'm not going to go into all of these details, but basically this is the marketing spends against the payer CACs.
Pretty simple graph. Yeah.
You know, these are the payback periods and this is the statistical kind of deviate, like the standard deviations that I was just mentioning before. What this means is basically like at two standard deviations, it's like 9 9.5 months payback which means if you were to do spend at a similar rate and this was had enough data in it um you can expect something below a 9.5 uh 9.5 payback period on your next one um we take a look at retention metrics i'm not going to go through all of them but it's basically like net revenue payer retention rpu retention transaction retention we take a look at the cac data um we look at correlations between like if you spend more what is your m1 m3 m6 roas look like if this is relatively flat that means that there's a lot of inelasticity in how much you can spend, right? So like the more you spend, it doesn't really change this number is a great sign.
But you should be looking at this, you know, ideally at a month six level. We look at it on a channel basis.
So these are all of the channels on a row, like row as per channel, a huge grain of salt here, knowing that half of this shit is like iOS and like, we don't know how they're attributing. And so like a huge grain of salt there, I'll admit, but like other than channel data, this is interesting.
If you trust, if I looked at the Android data on this and these were like quartile performance ranges. So like, for example, Adjo gets you like an M1 ROAS at 75% or the median at like 14%.
And somehow cash kick is like way better. I don't know if this is actually real, really reliable, but it tells you like how the ranges of profitability for each of these channels.
On each channel. And I will stop you here.
Both of these are different channels than, let's say, UploadMintegral and Facebook and Google. They are rewarded UA channels.
And obviously, those are always higher ROAS on the first months,

but then depends on how it actually evolves over time

because you have different rewards in these UA channels

and then people churn

and then suddenly your LTV projection

doesn't really go anywhere.

Yeah, exactly.

Makes sense as well.

So yeah, that's why I think what you said on the month six, Roz, and how everything evolves over time, that's really important. We see here clearly we're still waiting for Unity to roll out its new ML models.
Yeah, yeah, yeah. This looks terrible.
I don't know. I think this is just dummy data, to be honest.
Don't worry. They've announced that they're rolling it out soon, so everything's going to be fun.
Yeah, but still, this is good because then

you see the trends on the marketing channel, which is important.

You need to get the context on that as well.

OK.

OK, nice.

And then the last, you get the revenue split per channel,

which is pretty standard.

But the revenue split by cohort is also quite interesting.

So what you want is what I said.

You want to stack these cohorts as much as possible.

So these tiny, thin layers, over time, if you have great payer retention,

they actually stick around for a long time.

And then you're able to stack these really great cohort towers.

And obviously, by platform and by monetization type and by geo. So you get all of this rich information.
And like I said, what this at least can help you do is try to build that conviction on the cohort analysis before you try to scale up. This is free, by the way.
So anyone can go and do this and use this. Okay, another...
You build a calculator as well, like how you can... I think it's on your website, right? Can you actually calculate if you are eligible for funding or how your cohorts evolve over time.
And you can do a lot of different activities on the website that you have there. Thank you.
I appreciate you might be the one of five people who has ever tried to try to do that. I really appreciate that.
I think my CPO also really appreciates that. But yeah, this, this is another way to kind of look at it yeah essentially once you plug all

of that core data and we'll just tell you like where are you in terms of fundability and then we'll break it down in terms of like why why what what key indicators tell us give us confidence this is one that's unreleased um this is this is a debut for the two and a half gamers uh podcast but our benchmarking

our benchmarking tool

so like I said we've got a bunch of, I think right now maybe like over 200 or 300 cohorts for 2024 that we've looked at, like across all these different games. And so it's going to narrow down if you narrowed it down by category and genre.
But like, for example, this, again, this is demo data, but like, if you were to take this as true, your M1, this game, compared to the game's genre, peer set, and category peer set, is performing really badly in month one ROAS.

But for some reason, over M3, M6, and M12, it gets better in terms of the percentile range. The percentile is 61% at M12, but 15% at M1.
So they must have some kind of retentive capability, which pushes them this way. But again, if I was the CEO of this business, I wouldn't worry too much about the M1.
It's a leading metric, but what you really care about is that latter kind of ROAS. Yeah, people always ask me, oh, these are the numbers.
Is it good? Is it bad? How do we compare against the other companies?

And I always say,

look,

you need to look at like here,

like month one,

three, six, 12.

I had,

we had a game,

we worked with Jakob on it and it was great.

The first seven days

with ROAS was like 70%.

Wow.

Everybody was like,

oh my God,

this is so good.

But then the next 30%, it took us 11 months to

actually get that back so it's not good yeah it's like it's not really that great so that's why you need more more data points to be able to compare like where you stand actually it sounds like you more interstitial ads,

mate.

Yeah.

Absolutely.

Yeah.

So,

so I think like right now we're still a very young company. The database is modest.
But as we build out and take a look at more companies, this will be a much more rich kind of like source for you to compare against. It's not perfect.
I'm not purporting that, but it's somewhere. I don't have to call my

Facebook relationship manager as often if I had access to this data. You can also see it on a

channel basis. Again, with a grain of salt, attribution is last click, seven-day attribution,

most of them. I don't know how it's being done on the SCAD network to be honest it's wild west it's wild west of war yeah and um and then some other metrics like payer retention revenue retention payback period so you can really compare like a lot of people are like hey what is a good you just asked me like what's the payback period that that that you really Like, I don't know, this one says six point, you know, median is six.
Yeah. Right.
So it's one data point to take a look at. So this hasn't been rolled out.
We will roll it out soon. Definitely let you know.
Nice. And the last one, which is going to be rolled out, is basically if you plug your data in and you have limited data, we will pull together our data science model based on the other companies we've seen in your category and genre and your historical data, most importantly, and forecast what the next 12 months on each of these cohorts look like.
Eventually, this can help you find that asymptote, that terminal ROAS, predicted ROAS asymptote you mentioned Felix it takes you know more the more data you have the more accurate this will be and this translates directly into your financials and essentially for a finance person like at the end of the year you know December they're like okay what's your budget, what's our marketing budget, everyone? If you put in your marketing budget, this will plan out the future cohorts, what they'll look like based on your historical performance. And you can tell that finance person, this is what, this is what the revenue every month is going to look like.
Yeah. And they can do that planning.
So that's the tool that we're, again, this is all for free. We hope that'll be useful.
And yeah, hopefully we'll be able to see more fundable cohorts. Nice.
Okay. Oh my God, this is great.
So how long does it take from, again, to go through all of these? Like Felix mentioned, we have a great game. We are spending some money.
We plugged in the data. Now we want to scale.
Yeah. So if you wanted to get cohort financing from us, this process takes 24 hours, like doing the underwriting and actually spitting out whether you're fundable or not.
it will don't see over here is what's the price, how much we can extend to you in the facility, what's the financing period and all of that. That will be out in 24, max 48 hours.
And typically after that, and let's say you were fine with those terms, usually an agreement will take a minimum two weeks. I'll give myself two weeks.
And then that month itself, we can fund. So if you wanted cohort funding, you could technically do it within the same month.
Yeah, in 30 days. Okay.
And basically you send the money in the bank or does it have to be spent in some way or it's just money in the bank? Money in the bank. Money in the bank.
You tell me, look, I'm spending a million dollars. Give me $500,000.
You get 50% of this. I transfer that directly into your bank.
The next month, you ask me for another million dollars, but your first cohort owes me $100,000. Then I just net that off.
So i send you nine hundred thousand dollars instead of a million so that's that's how it'll work until yeah oh okay so yeah nice i mean it's great so how how people can can contact you and say say what we just said like we we have a great game, but we need money.

Yeah, I mean,

first of all, if you're just more curious

about the broad topics and I'm happy to

talk to you, you can just add me on LinkedIn.

I'm Ritsky from PBX.

But if you're

really keen on the financing, just go to our website

www.pbxpartners.com

and there should be a

Get Terms button and it'll take you through the process real quickly. Oh man.
Okay. So we'll also add some notes into the description with the link to the pvxpartners.com as we usually do and also link to your email and LinkedIn profile as well.

And anyone who wants to email you,

use the subject heading asymptote because we've messed it.

It's the word of the day.

We mentioned it the most.

Well, I don't know.

Like the last episode I watched from you guys was like,

why hexagons because of their bestagons or whatever.

So yeah, everyone gets a new word for each episode exactly okay thanks Ritsky for coming again thank you very much guys for listening if you have any questions please comment on the YouTube video it's gonna be the visuals. Join the Slack channel and see you next time.