Signals Through the Noise: Pixels Not People

[00:00:04] This is the Crypto Hipster Podcast. This is not a traditional interview show. These are perspective-driven conversations with founders, builders, and independent creators shaping what comes next.

[00:00:26] We go beyond headlines, beyond hype, and beyond price to explore ownership, freedom, and opportunity in the digital economy. Where builders talk freedom, not price.

[00:00:48] Earn, was not an economy. It was a loop. So what happens when the hype disappears? I'm here with Luke Barberkowski from Pixels. Let's get into that. In one sentence, I want to find out what is the economy inside of Pixels? In one sentence, essentially it's AI-driven rewards now.

[00:01:19] And where does that value come from and how does it work? So the value comes from any other game's value as well, too. So essentially, Pixels is now building in the category of rewarded play. That means we'll build games that are fun, that generate revenue just like other games, but will also reward you while you're playing. And that key concept of rewarding you while you're playing is a nuance that we're building here at Pixels.

[00:01:46] It's not so easy to achieve in a way that still maintains, you know, profitability for the game, still rewards players in a fair way. And that's kind of what I mean by AI-driven. Like essentially our whole reward stack is, you know, years of work of data science and AI. So it's external revenue or new players or combination of both?

[00:02:07] So when it comes to the economics of the game, like we're modeling a lot of our economics basically purely like a free-to-play game where we have a customer acquisition cost and an LTV of the user. We also have one more dimension that we need to play with now, rewards essentially. So you kind of can view all three of those things working together to build out the profile of a, you know, per profit per user. It's the LTV minus the customer acquisition cost minus the lifetime rewards, right?

[00:02:37] Then what makes Pixels different from previous play are models who are kind of similar. So really there's nothing in web three that has ever come out that's similar to the tech that we built. Essentially, what we've been doing over the last few years as Pixels was released was we built a game and we had play to earn rewards in the game immediately start, right? Now, the issue with the play to earn rewards that we were first giving out was we had the rewards intertwined with the game mechanics.

[00:03:04] So essentially, if you reach like level five, this is a very gross simplification, you would get these rewards. If you started to progress further through the game, you might get more rewards, right? Now, when you have reward systems that are that simple and when I say reward systems, I really mean any ownership based economy, because even if you're not intentionally giving out rewards, if you have parts of your economy that are tradable with other players, and especially if there's a real money aspect, you're giving out rewards whether you like it or not, right?

[00:03:32] Whether it be an in-game item that's farmed, an in-game currency that you earn as you progress. So, yeah, we had a very, very simple and naive approach to reward distribution at first. Essentially, level gating and rewarding based on like different effort that players put into the game over time. And that's kind of like the default reward distribution of any game, right? Like that's how normal game design works.

[00:03:55] Now, however, if you're doing this, this is a problem that MMOs already came across for the last like 20 to 30 years, right? It's not unique, especially if there's a real money aspect. Your reward currency or any currency inside of the game will be farmed, right? So then a lot of our approach switched to anti-bot and cyber attack resistance at pixels. We also kind of realized that was a never ending battle and not really the right approach.

[00:04:20] The reality of the situation, I think more people are realizing it as AI bots are getting better. But we saw in games two or three years ago is that there's no way to determine at all on the Internet who's a bot and who's not a bot. Like really the only solution that might work is WorldCoin if you're doing the eye eyeball scanning or world like, you know, digital fingerprints, things like that. But yeah, basically anti-bot technology online just doesn't work. So we had to approach the problem in a different perspective.

[00:04:50] We said instead of, you know, trying to avoid giving rewards to bots, what if we instead really made sure that we only gave rewards to the good players that help the ecosystem and a good player helping the ecosystem is spending more than they earn. But there's a lot more depth and nuance there. Like a good player could be somebody who is bringing the game to other people and creating content or just making the experience better.

[00:05:17] And that's where a lot of the tech stack lies that we've been building out. And this is what's different than us versus, you know, even the first iteration of pixels and other things inside of Web3. Essentially, what we built is we built a data driven stack that basically allows us to ingest every single event going on inside of the game. And we're able to make predictions on what users might do when we give them rewards.

[00:05:39] And we're able to also figure out what types of rewards might lead to different types of outcomes, whether it be revenue, engagement, sharing the game and increasing K factor. So pixels is at a great advantage because we've given out hundreds of millions of dollars of rewards at scale to millions of users. We also had a ton of historical data to go and look at to help build out these data models in order to actually build out the best system possible. And that's this iteration, essentially.

[00:06:07] This concept, it's called play to earn rewarded play. It's also something that's kind of working in Web2. There's been rewarded play platforms that are starting to break out inside of Web2. Now, the difference between pixels and these rewarded way play platforms is these rewarded play platforms have understood if I'm going to go and acquire a user, if I'm spending $2 on ads, maybe there's a way to go and acquire a new user for $2 just by giving them it to them directly.

[00:06:35] There's platforms out there like Adjo, Misplay, Justplay, a few others that are starting to break through inside of Web2. Now, really, all of those platforms, what they're doing is they're basically giving you a download incentive. So you download the game, maybe you'll get a dollar, but it actually in order to get that dollar, you might have to hit like hours and hours of progress inside of the game. They're just essentially the most naive versions of rewarded play and play to earn.

[00:07:00] And what's different about pixels is we have a lot of real data on rewarding players through all aspects of the gameplay and with incentivizing all aspects of engagement, not just purely a download or a strict UA thing, but also engaging the game daily. How might that lead to net revenue inside of the game, even if you're not in that spender and other spend mechanics as well? So essentially, we have a really in-depth platform that can kind of figure out who the right person to reward at the right time is for the right amount.

[00:07:28] So you would say what the other older models that failed in play to earn did wrong is they didn't leverage data or the prediction markets? Not so much prediction markets, but like when I say prediction, I mean, like building out data models that can help predict what a user might do inside of your own game or ecosystem, right? So a very good example would be one of the predictions that we're doing is LTV prediction, right?

[00:07:55] So we want to know if a user is likely to spend inside of the ecosystem. If they are, it makes a lot of sense to give them rewards to stay inside of the game like much earlier on inside of their journey. Now, the opposite is why we started building this. If we know that a user has no intention of ever spending inside of the game, they're likely a bot. And if they're not a bot, you know, they're likely an extractive user. They're normally a huge net sink inside of an ecosystem. So that's the type of user who we probably shouldn't be rewarding inside of the game as well, too. Right. So there's two parts of that.

[00:08:24] So, yeah, the uniqueness of pixels is around the data driven insights. It's only possible because we've been literally working on this for years to build this out. We have built it internally in pixels with our own first party title, and we did it with giving out like tons and tons of rewards in order to get the data that we need. Got it. So I just want to find out what decisions in the game the user are making that are driven by these predictions.

[00:08:52] Exactly. Now, the most like basic way that you can frame this to a user is it's almost like a loyalty platform, right? Where we know who to give out rewards to for the right reasons. And for the user, really how it should feel is they're just playing the game like they normally would. And the reward platform rewards them for good behavior. And the rewards are nice because we can also steer behavior in certain areas of the game that we think are beneficial, both having the user have more fun and also leading to better economics for the game as well.

[00:09:21] Right. I want to talk about the return on, you know, the returns. Right. So you showed two dollars per return per one dollar of spend. So walk me through exactly where that two dollars of earn per dollar come from. Yeah. So what we've seen across the board inside of Web3 games is a default Web3 game that ships. We've been talking about this metric called return on reward spend.

[00:09:47] The most simple way to define this metric is essentially a monthly aggregate of how many rewards a game gives out every single month compared to the amount of revenue it does. So a return on reward spend of one would mean for every dollar of revenue that you're getting or every dollar rewards that you're giving out, you get one dollar revenue back. We see most games default at is for every dollar of rewards that they're giving out. They normally get about 20 cents back. And that's where the unsustainability part comes from. Right.

[00:10:16] Pixels started at that metric. I think the first month that we released after our token released, there was about a return on reward spend of about 0.2. So for every dollar rewards we gave out, we got 20 cents of revenue back to the protocol. This is where stacked has come in play. So we immediately identified that as a problem and we tried to figure out how to fix that. We tried a bunch of different things. Again, the whole journey was like, OK, do we do reputation systems?

[00:10:44] Do we do reward tuning? What we found the most effective was basically doing all these data driven insights. So our return on reward spend went from about 0.2. Now over the last 30 days, I think we're averaging 2.5.

[00:10:57] Last seven days has been about four. So that means for every dollar rewards that we're getting out, we're getting two and a half to four dollars of revenue back, which is amazing because that also gives us one, the ability to swap the reward currency and the spend currency from our cryptocurrency into U.S. dollars. So we wanted to write that actually means that we poured in U.S. dollars for rewards, we'd be getting more U.S.

[00:11:22] dollars back. And then two, what that also means is now we have a cushion for actually doing paid user acquisition as well. If our user acquisition cost was generally about one dollar to like three dollars anywhere in that range, then that means we can also profitably acquire new users. I'm interested to know as you were stress testing your as you're stress testing your model and you saw different things break, what broke?

[00:11:52] And why do you think why why do you find like why did it break? Do you think so one of the things that we initially saw when we started optimizing return on reward spend was you did see a big DAU drop off initially. Right. Like the reality is there was a lot of users inside of the game who really only wanted to take as much money out as possible, which, you know, that's obviously something natural. You can't fault any of those users if you build a system that allows them to do that. Right.

[00:12:19] But the reality also is when you create systems like that, you also do attract a lot of users who are not interested in the gameplay, no matter how fun it is. And obviously it's going to upset some people when like a free money faucet dries up. Right. But the issue is you have to do these decisions, you have to make these hard choices in order to actually build something that can really grow and be real. Right.

[00:12:39] So it wasn't easy to make the decisions to like basically upset a lot of our users who really liked how the system worked previously in the pursuit of sustainable growth. But yeah, now we have a smaller user base than we did before, but it's sustainable now and now we know we can reliably grow it. Got it. So how many of those people who left were actually people versus bots?

[00:13:07] Yeah, it's interesting because I think at our peak DAU, our estimates, and this is like after a lot of, you know, data science and looking at things, I think we had a million daily active addresses at the peak of pixels. We estimated at that time it was probably 50% bots. But the interesting thing that we've learned about was we were trying so hard to figure out who's a bot and who's a real user.

[00:13:35] With our mindset, when our mindset shifted from does it matter who's a bot and who's like a real user if they're doing the same thing? It helped us a lot. Right. So we would often find that we would eliminate a large bot farm and there would be real users getting, you know, caught up in that. Sometimes it was a farm of real users, like a real user that had 10, 30 accounts. Right.

[00:14:05] That's also potentially an issue, but they were actually playing. They didn't feel like it was so fair, even though that was against our terms of service to have so many accounts. But yeah, you do piss off some real people when you're going to be doing these things. But again, it's in the pursuit of building something real for the people who actually like the game, who actually are interested in the token ecosystem. One of the hard things is you have to remember, you might have many stakeholders inside of an ecosystem, but not all the stakeholders are keeping the community or your own interest in mind as well.

[00:14:35] So, yeah, that's what building something real looks like. You have to really build for the right people. And that means sometimes moving on from the people who are not serving the ecosystem. Got it. Got it. So I'm wondering if the growth still like right now you're good at a sustainable growth level, but if it slows down further, is your is your are your how's your economic model going to be affected by that? And what's the weakest part that you don't need to address?

[00:15:04] So right now, our return on reward spend is quite comfortable with our current audience. The core game that we have right now has not been growing because we don't have a paid funnel to it. This is why we're also releasing a second game right now. It's an early access. We've been running a lot of paid ad experiments on that we've done probably tens of thousands of spend on like probably 200 different formats at this point on this new game that we're releasing. So we have a very, very good idea of what the customer acquisition costs can be for that game.

[00:15:36] And what the challenge will be is how do we scale the AU while keeping this return on reward spend where it is because it's the same issue as scaling a normal game. You'll have your first cohorts of users who will be, you know, the cheapest to go and acquire because you have the perfect fit for the game. Right. And as you start to get more and more users, you're going to start to probably see higher CAC as you're expanding the audience a little bit more.

[00:16:05] I have a feeling that we might see something similar with return on reward spend if you start to get sufficient scale where your margins will start to get smaller. But also, that might not be true because truthfully, my bet also is that the current crypto audience will probably be the lowest return on reward spend audience. This is what we've seen time and time again. The expectations that have been built by the crypto audience are actually like really weird. Right.

[00:16:34] We've been doing a lot of study on the different rewarded play apps and web to right now. And it's crazy the difference in mentality. You'll look at apps like just play a misplay and you'll see people saying, wow, I only had to play for like two or three hours every day and I earned $10 this month. That's amazing. And that mindset in crypto is like if people aren't earning $50 in a day, your game's a scam. Right. Even if they didn't pay anything.

[00:17:00] So the real challenge will be scaling the UA and getting new users who are not as like crypto familiar and getting them into a system that has like crypto elements behind the scenes. So what we've seen is anytime we mentioned crypto in our ads, the ad performs five times worse. I think the real challenge will be scaling to this audience and saying, hey, there's stuff here that's interesting.

[00:17:27] We know you like the concepts behind crypto and making sure that we can scale while being compliant in the app stores and getting that new users who are not as crypto native into the game and ecosystem. But as you scale, what would you say? What would you think breaks first rewards or retention?

[00:17:50] What we've seen is rewards and retention are pretty easy for us to keep up even core pixels right now. Our day 30 retention is so insanely good compared to like our D1, D7.

[00:18:13] We know that when we have the users inside of the game, it'll work. but are we going to be able to acquire new users for cheap enough or long enough will be the big thing that might break the other thing too will be app store compliance we've even seen that there's web to play their own platforms out there that have been removed from the app store um and the reason why they're being removed is holistic it's not for any one reason the one that we just saw

[00:18:43] removed free cash um they were having pretty misleading advertisements so um and also the app store is still not very crypto friendly so in order to be app store compliant we actually do need to remove almost all of the crypto parts of the ecosystem um but we can still have the crypto parts work in other areas of the ecosystem but not in the rewarded play part if we're going to be ios native um so those are the things that can break basically compliance app store um and like scaling

[00:19:08] the ua i'll get into that a little bit i always want to make sure that your system the system requires constant growth in order to survive right to continue update the game and forget new games and continue to innovate well our system doesn't need constant growth to fuel rewards right because we've had a positive return on reward spend um for a long time with the same amount of audience um like essentially the monetization model is just like a free-to-play game where if you have

[00:19:37] people playing the game um they're going to spend money on it if they like it right now you have to get an audience of people who actually like the game in and disincentivize the audiences that don't like the game from getting rewards um but yeah our dau at pixels has been basically the same for the last three to four months and return on reward spend has been over one averaging about 2.5 for the last three

[00:20:03] to four months as well um and that's not because we're funneling in new users and churning users and requiring like constant growth to fund these rewards the rewards are being funded because people like the game and they're willing to spend inside of the game right got it so let me back up and ask why do people like the game so much well pixels is uniquely interesting in a lot of its aspects

[00:20:27] right it's like stardew valley if it was multiplayer and had a real economy attached to it um and we do a lot of live ops inside of the game that's not to say that like the game is perfect right there's a lot of things we need to fix in order to make a mass market like one of the weaknesses of the game right now is the new user experience is not so good um so before we start advertising and doing paid ua for pixels we do need to change the first time user experience and that's what the team is working

[00:20:53] on right now we also probably need to get it ready for mobile native but when people actually learn how to play pixels they understand there's a lot of depth there and it is very interesting like a lot of players play in a bunch of different ways um some players just like to you know chill and farm resources some people love to play like the marketplace um speculate on like you know what demand for different items do like if you've ever played an mmo that was kind of how i like to play mmos like this is more even this is more interesting because there's you know an actual cryptocurrency

[00:21:22] involved in this right so the decisions you make have like a little bit more weight and stake associated with it um and then we do a lot of live ops too that keeps the game interesting the game keeps getting constant updates so the meta changes quite often uh so the players are are the players here for the game or the money so at this point um like the players that are here here for the game um the earn aspect is like an interesting part also right um but

[00:21:48] it's not like the early days of axie where people are quitting their jobs to play pixels it's like a nice to have right um we did have a bit of that in the beginning stages also i was really trying to actively discourage that um because i just we didn't want to go down that same route um but if you don't have the right systems in place you might encourage that by accident right it's one of those things where if you're not really smart about the systems that you build out um that will be the end result for a

[00:22:18] short time even by accident um and that's why intentional reward design is so important because you really want to make sure that you're giving out the right rewards to the right people at the right time um not encouraging like quitting your job to play a game got it so so so if your rewards drop back to 0.2 tomorrow say they drop tomorrow who stays and who

[00:22:43] leaves um well if aurora's drops to 0.2 more people would probably come because that means we're giving out more rewards than revenue right um so that's always a growth lever that you have if you really want um if you can like sustain um like a bit of unprofitability um to go out and acquire new users but like right now this is probably like the least amount of rewards the pixels given out

[00:23:09] in its history day to day right um so um right now we're in this very sustainable part of the gameplay we've been able to keep a core audience we've been able to monetize better um and players are still showing up right so let's talk about let's talk about the players they come to your game they own a lot of what three games they own something what do they what do they actually own yeah so ownership

[00:23:39] is an area that we've experimented with a ton in pixels at one point in pixels every single thing was on chain now the issue that we found is a lot of people don't really care and also you don't need everything on chain to have quasi ownership right what does ownership actually mean to a user well truthfully an ownership to a user what it means is can they trade things um and is there at least some

[00:24:07] real money component um like a lot of users might want to exit the game at some point and be able to cash out um in their mind that's what ownership means now in order to do that do you need every single piece of the game on chain um what we found is no one is when we had everything on chain one day it broke and no one even noticed for like a week um because they kept all their stuff in the game

[00:24:31] anyways right two if you have one element on chain typically that has flows to the rest of the game economy for example if we just have our hard currency on chain and there's some earned element even if you're the other parts of the game are not on chain you need to use all those other things to you know get the on chain value right and those things have inherent worth because of that um right

[00:24:58] so you also need to think about liquidity markets um like even if you had everything on chain with full ownership if there's a really specific in-game item that has super liquidity inside of the game you will never find a buyer for that item outside of the game right so when you think about where the ownership part matters matters for whether the parts of the gameplay that other people outside of the game might want to speculate or own and when you're in the game economy there's not that much

[00:25:28] right um so we have two pieces on chain right now we have our nft farmlands and we have the token the hard currency on chain right and that's generally enough to have all the other benefits of on-chain ownership inside of the game economy um while still having like you know liquid markets um for those particular pieces um if i was to give advice to like a new game studio building one have

[00:25:57] ownership aspects in their game i would say just have one piece of your game economy on chain that's really all you need to do that way you're not managing liquidity amongst many many markets um and not splitting it up and it's concentrated on one asset so say say pixels for some reason it disappears tomorrow what actually survives then for your users so this is the tricky part right um people

[00:26:25] want like a fully decentralized system inside of web3 gaming and i'm fully a proponent of decentralization to like i in my personal life i'm a very big fan of the free cities concept decentralized governance all of this like i'm a buyer into the system right now the issue is practically getting there right um so here's the case of a game games do have cost right and games have servers to run right so say that

[00:26:53] the pixel foundation completely ran out of funds who's going to pay for those servers um we've seen a few other games have that same issue that have shut down right and this is the key issue with ownership like if somebody else has to pay for something and the community's not willing to pay for it um is there actually really ownership of the assets um eventually maybe you'll get there right but like truthfully for ownership to really work you need a protocol that has revenue that can support

[00:27:21] itself long term right then the ownership aspect makes sense in the long run right um so our priority is building something that works first before we get too idealistic about fully decentralized ownership because if we're giving everything ownership instead of the ecosystem right now um but we're not able to build sustainable revenue then ownership doesn't really matter for any of the users at the end of the day anyways which is the unfortunate truth right now in our situation if we were ever to get into that

[00:27:51] to be honest like we'll open source the game code i really wish that more teams shutting down in web three would do that say okay if you're not using these assets anyways why not at least let the community try what's nice about pixels is you could probably host pixels for cheaply but again we want to get to a situation where you have a protocol that's working sustainably where the community can have true ownership and the way to do that is you have to build something real that works with real revenue um the team is looking like we're on track to do that so we're going to go and push for that first

[00:28:21] guys so right now it's dependent it's more dependency and in his ownership and that will be how any system that's bootstrapping itself will look right um that's kind of the whole point of web three like i i've never seen anything in web three that can be fully decentralized off the bat without bootstrapping in some aspect um if we could figure a way to do that that'd be great um but here's the other issue

[00:28:44] too when it comes to the idea of governance um you know as we've been running pixels you know i've also become more skeptical of community governance in the early stages um you really need somebody that has long-term thinking um in order to build something real and long-term um a lot of the community asks in the beginning stages we're very short-term minded like for example there is a community treasury that

[00:29:13] eventually will have a dao govern but if the dao is governing it in the beginning stages the community likely would have voted for a total airdrop to all of themselves and that treasury would be gone right um and like the treasury would probably be sold off immediately one of the really sad things we found out um was in the beginning stages of pixel we were giving like a lot of rewards out and the whole thought process behind that was like okay we're going to slowly distribute your rewards to

[00:29:41] all the best players inside of the game and the community will have ownership of this network great on paper right it's like oh wow we can actually distribute ownership of this whole protocol network to the most engaged players in the network um but 99 of tokens were sold immediately when given um so really the community didn't want ownership of the network they wanted you know money um and that's fine right like you can't fault that human behavior but that also shows you how much people actually

[00:30:08] truly value ownership of a network um so given that we also need to make decisions for the long term uh building out something real so that eventually these ideals of decentralized ownership can happen but we need to focus on the economics first so we can actually do that um so the long-term thinking is more important in these early stages and to say there is that treasury sell-off okay

[00:30:32] say you remove you remove financial incentives tomorrow you know totally does your game survive and then thrive as we go forward so core pixels right now it is profitable um we have a profitable like reward method right now um so like we could turn we could get rid of the token and it would work

[00:30:58] right now right um so yeah there is something viable there right now um in our other game that we're about to release chubkins we've been testing a ton of ads with it it's about to come out next month or two um the user acquisition cost right now is sitting at like a dollar fifty to two dollars and when you have that that's like a green green light um that game is looking pretty financially viable also right um now the issue will be again like can we scale that to something that is very interesting right

[00:31:28] because you know monthly revenue in pixels is not huge um it's able to support itself if we were to switch it to us dollars but is it going to be interesting to outsiders so much maybe not right the revenue is not insanely big um and that when we're trying to build something real and great not just something real is what i'm shooting for you know i didn't come to build out something that is you know doing all right like i want to build something that really changes this space in the

[00:31:55] industry and that means you know getting revenue numbers that are going to make the web 2 industry say oh look at that right um so yeah in order to do that we need the economics to look good we need to actually scale the user acquisition spend um and that'll be the real test but yeah if we want to scale back everything um say hey we're just going to keep the lights on we can do that right now

[00:32:20] he said earlier systems failed because they didn't reward the right players at the right time and what they built at pixels is clearly more efficient better targeting better retention better economics but the core question did not change

[00:32:46] where does the value actually come from because even with better models better data better incentives if the system still relies on players spending predicting and circulating value inside the system

[00:33:10] then it's not a new economy it's a more refined one then that might just be enough but that's the line and whether holds at scale well that's the real test