Outlier Ventures

Transcript

[Music] ready to go [Music] hello everyone today I am going to talk about data-driven token design and optimization my name is ahim and I'm married I have two children I'm coming from North Germany and I have a traditional engineering background in mechanical engineering civil engineering working on Innovative support structures for wind turbines and I have just a

passion for sustainability in every sense and sustainability is also a very important topic for token ecosystems and I am very fascinated by yeah the aspect of what we can do with cryptocurrencies like it's programming programmable money and yeah the opportunities are infinite so that's why I'm quite excited about the space and happy to also share some thoughts some contribution contributions I made so far to the space um and I also want to point out that I'm really really impressed by all the speakers here that were here before me so really in depth and really

fascinating all the different angles on which on how we can approach token ecosystems so my angle will be a little bit more from the early stage startup web 3 business perspective so I think a lot about how we can use essentially token engineering to help early startups to make them also sustainable and also to educate them on token ecosystems and their own ecosystems in especially so and I would like to talk start by showing you like some token design tools we are using and I developed so the first token design tool

is not a tool because it's essentially about people we have a very very great team of different experts representing like different aspects that are important for early web3 startups so we have yeah the token designers I would say the best in the world and yeah we have me as the token engineer we have dimi he already spoke here today very smart mind Dao strategy we have nft people product engineer marketing we have the legal side represented people taking care about fundraising aspects and so it's just like a very very great team and we have

like all these Specialists and teams working with us us have essentially like access to all of these people so that is in my opinion one of the most important resource and tools tool that we have like individual partners for all questions related to web3 startups oh why does no one tell me anything about this let me share again no one is telling me anything

so now you can see all these beautiful faces of our team okay great let's move on um so we have like this a little bit differently interpretation of the 3DS like um the discovery phase the design phase and the deployment phase that how we apply it like in our context working with young startups so we have like different programs focusing on these different stages so like the base camp program is more focusing on the more qualitative side so looking at ecosystems from a higher level perspective

getting like the problem statement down defining the different stakeholders like value flows between them and like just we are thinking about and talking about and drafting if the whole ecosystem makes sense if the token makes sense and really drafted out from a higher level perspective and that is all happening like in this base camp program and then we have like the more um in-depth program I would say where it is about actually launching the token the ascent program where we also start parameterizing the ecosystem and this is essentially where we also apply more more quantitative tools such as I mean

you can always start with analytical calculations you can use spreadsheets you can use machinations you can use like more sophisticated environments like cadcare token spice and so on and you can get like optimized validated models out of this can do risk analysis and yeah then forecast essentially how your ecosystem is going in terms of token Supply but also in terms of valuation and then we have like the deployment phase in terms of how this like theoretically analyzed and drafted ecosystem then will actually be deployed on chain like there is we use like

several different tools like the different programming languages like solidity rust the environments for deploying those ecosystems like Hardware truffle brownie and so on and the output is essentially then the live token or the product and so from what I see there is really like the spectrum of tools that you need to have in all the different phases and there are also different emphasis necessary for different types or sectors so a lot of talks today were more around the D5 sector but there are also many application projects infrastructure and

so on gaming sector and it also depends like in which sector or about which sector we are talking when we are considering the level of detail and there is like the one side who says okay we don't need any simulations we are just looking on our ecosystem from a qualitative perspective we just use like the market standards and that is good enough for us and then there's the other side who says okay we need the one-to-one representation a digital twin of the whole ecosystem and we want to be very accurate and maybe there is a lot of money at stake and

in the end it is like the resource intensity increases the more you come to the to the right side and the coding knowledge increases um also so you need to have like more coding knowledge when you want to build more detailed um yeah models and ecosystems in general and this has like one main downside the user friendliness because you really need to be an engineer who is capable of programming to essentially um yeah work on the right side and it's difficult to work with sophisticated models when you're not like too much into math for example

yeah um so what are the tools on the left or on the right side so like on the left side I said already like spreadsheet models like we have qualitative tools like problem statement stake problem statement and stakeholder mapping and certain value flow we can even leverage AI powered reasoning so leveraging llms to draft like a first token design for example and then we have in the middle and I would say for example our qtm the quantitative token model is also a spreadsheet model but it it's like a macro perspective on the whole ecosystem it can cover like a variety of different

sectors I would say so it's not just specialized on defy which obviously comes with a trade-off that it's not really accurate but it can really help young startups to get the first insight to get a first idea First Forecast and interview duration for their parameters so and this is really really helpful and I decided to do it in the spreadsheet because this is accessible right everyone can use a spreadsheet and everyone can in this way learn about their own token ecosystems currently I'm working together with other students we are working like on a cadcat integration

so that we can also extend it to more sophisticated analyzers where we can then do parameter sweeps optimizations Monte Carlo's simulations and or runs and Markov chain probabilistic agent behaviors and so on so this is what we are currently working on for teams that want to build more sophisticated models yes yes so this is like like a little overview of how this qtm quantitative token model looks like it has a 10-year fixed simulation time one time step is one month so it's really really

rough it's really like a macro simulation model and we are coming at the beginning of each time step there is essentially the emission of tokens into the ecosystem so we have like the incentive incentivization module uh token investing air drops and so on and then those tokens will be fed into several meta buckets and from this meter buckets they will be then redistributed in finer more granular generalized utilities and then from these utilities the reward payouts will be defined and so on and then there's also like the off-chain business aspect so which takes account for the rent General Runway of

the business so work from which we can also perform burnings or BuyBacks and which also measures the user adoption or where we can Define the user adoption so the quality of output the quality of results of this model really really comes down to the quality of input and you need to do like own market research on Research in general to yeah to get these inputs and really understand what is going on if you want to generate generate like more realistic outcomes but as I said we consider this and treat this more like an educational tool for early startups to get a first idea of

their ecosystem so you should not derive any financial advisors out of it or rely only on the results in it and yeah then we have like the data analytics perspective what kind of data can we extract in general so obviously there's the macro Market perspective um we can just look at what is happening in the overall Market thread 5 perspective then also crypto market then we can look at metrics from fundraising rounds like how much did they raise um at which value Asian and how much Supply did they sell in these individual

rounds then you can look at vesting schedules so everything what others are doing which gives you again like another data point another perspective that you can use or we can use to put it into like this qtm for example or another or in your own model you have like a starting point just to see okay what is the market are doing and then there is one main and major difference to like tradify that is like the on-chain reality because on chain we have a lot more information than in tradify unchain data is accessible for everyone and we can literally see every single transaction that is going on in

the ecosystem so we can literally see um yeah I mean we can measure also obvious kpis like onboarding of users or tvl or volume and all this but I think what is maybe even more interesting is to see how different incentives are work or work out like we saw in the previous speech from Gauntlet very impressive how they analyzed it for uniswap and we can generalize in general just see how agents behave and this is really crucial and fundamentally different to tradify and when we are just looking at shares in the stock market for example and we have socials we all know that

like Twitter Reddit Discord Telegram and so on all these social media platforms also play a crucial role for um yeah for token economies and how they perform and all that is available to us and I think it's very valuable Veil data that we should use we should leverage to first get a first idea for our own parameters of our own ecosystem but also to validate our models yeah here's like one little example of what kind of data we can look at like vesting Creations I know it's very tiny

here from but in general so we can look at the vesting durations for different stakeholder groups and in this particular diagram you can see like the vesting duration the minimum the mean and the median and the maximum for vesting durations across all different sectors but we also yeah do the same analysis then differentiate it with respect to the sectors and from this you can already see like there is a wide spread um and of of values and I will later on I will also talk a little bit about um why these values are not necessarily the

best but it is again it is the starting point I just have to look on the time and then we have like for example token bucket balance histories here is an example from Maple Finance so I looked at the native token of maple finance and then looked at all transactions in the whole ecosystem and categorize them into certain like meter buckets so for example all Maple related addresses centralized exchange related addresses decentralized and so on we see addresses and then we can see

essentially or we can create representative balances for each of these stakeholders and then we can essentially see what is going on in the ecosystem so here in this example the balances of all Maple addresses are going down continuously and then here in the middle July 22 they introduced their stacking contract and suddenly a lot of Supply got allocated to the staking contract and we also saw like the VCS participated in this um yeah staking program and this is what we can directly see out of this diagram as an example we can also see how like the balances of exchanges exchanges change over time and

it's pretty useful to get like a first overview of what is going on and then I also looked at individual stake or individual address Behavior so I know it's very small in general I looked like what happens with token that got sent by a specific address so for example from the staking contract tokens got sent to certain addresses who were the receivers how much did they got from this address and and then as the next question what did they do with these tokens did they reinvest for example did they send it

back or did they just sell or deployed it somewhere else and this is what I looked at in this little analyzers here and it really helps to understand like the behavior and from from the individual stakeholders and we can use this data essentially and feed it back in into our models here in this diagram I also analyzed the token receiver Behavior but this time not for individual addresses but again for represented a representative aggregated stakeholder groups so um here for example not just for Maple but also goldfinch trufi and trufi and you can essentially see like this big

blue bar here represents that around 38 of all the tokens that have been received by a state by the staking contract then have been sent back to the staking contract as their first transaction after receiving it and same for the centralized exchanges it was like eight percent for decentralized I think exchanges what light was like 14 percent and yeah you really can see when we are going back to the to the qtm where we have like this meter bucket allocations where we can essentially see what are the what is the free Supply floating Supply are doing we can then essentially apply these values to our parameters

just to get a first idea of um yeah what the behavior in the ecosystem would look like foreign yeah as I said we can use the values here this is just an example um what would be the output of the qtm the quantitative token model so we have the balances of um or the supply development in different buckets of a ecosystem so for example from the foundation the team from the staking allocation the overall circulating Supply and how many tokens the liquidity pool and so on and you can essentially

forecast this for the whole for like 10 years in this case and we also have yes we have also a price simulation or forecast again this is not meant to use meant to be used for any speculation or any like Financial advice but it helps to understand what is the relationship between Supply emissions and also demand for the token so that we get like a feeling of um yeah of these two forces how they are balanced out in the ecosystem over time as I said it's more like on the macro perspective and yeah not meant for speculation

and we can again also analyze other things like the allocations to the different utilities if you have different ones so like how many tokens are staked how many are used for liquidity mining incentive program or how many are burned if you have a burning mechanism or if you have like a store or something else where tokens were sent to and on the bottom right we see the monthly utility rewards where we essentially see how much of these incentives or are valued in dollar denomination because like the speaker before already gave like you already mentioned it is really important to understand also that tokens always have

a value and if we incentivize our ecosystem it comes at a cost and yeah so another topic I was working on is a new way of thinking about like vesting schedules because often I see this question what are good vesting schedules and sometimes people come up and say okay we just need very long vesting schedules but this is not always like good because this means we have a very low Supply at the beginning

in the in circulation and essentially means or opens up the room for a lot of speculation and potential like pump and dump behavior for the market so this is important to keep in mind and that's why I came up with this idea to introduce like an adoption adjusted vesting which is also demand agnostic so in other words we don't even need to forecast any demand for the ecosystem because the emission for from investing will be controlled by a controller which again will be controlled by certain kpis that we can Define so it can all all kinds of

kpis can be tvl can be volume can be user adoption can be profitability of the of the web 3 business whatever in this example I used simply the price of the token so here this is an example from the real token how the price just goes down because of the vesting because in the beginning in the first years there is a lot of vesting a lot of Supply hitting the market but there is no demand for it because the product is maybe not mature enough there is not enough demand not enough adoption and this is essentially what I also um modeled here with the qtm where you

see exactly the same behavior there's like this dumping the price goes down in the beginning because yeah just the Supply hits the market and then after a certain amount of time when the adoption comes when the users are coming the revenues coming BuyBacks can be done then price will eventually recover so but this is like the state of how it is at the moment and with adoption adjust investing um I think I have five more minutes right oh well someone else um so here I just mod like three different demand scenarios so like with logistic function linear function exponential

growth and then the controller essentially managed like the different emissions to all the different points of time and you see like for each of these different growth or demand scenarios the controller will manage the um the emissions of tokens that is what you see on the right side the emission rate for the vesting participants and on the bottom right side you see essentially like how the price of the token would look like that it will be essentially stabilized so when the price goes up more Supply will be released to the ecosystem which means obviously the early investors they will sell which means the price will then go down and

when the um when the when the price is below the implied price that I defined here then the emissions will be cut down but they will never be zero because we need to make sure that all the yeah all the initial investors will get their allocation eventually and yeah and then this will essentially Drive the price up again and yeah you can see through this controller approach we can in the end reduce the volatility and we can stabilize the ecosystem and I think in my opinion price is a very very important metric for an ecosystem we

already heated this morning then um if the price just dumps down minus 90 in the beginning of Earth year um it is very very bad outlook for an ecosystem and if we don't consider this even if we know we are not able to to forecast anything accurately but at least we we think about this demand side and we try to model it and to forecast it and it's never like about one specific outcome one specific value that is why we are doing Monte Carlo simulations and parameter sweeps where we can essentially sweep the whole Space of yeah solutions that are available so yeah this is just a few more Graphics

about what happens to the individual stakeholders or the individual receivers of these vested tokens because we can also apply different weightings to these best things so that for example in the beginning the ecosystem incentivization will get more allocations of the vested tokens and the team will get less and maybe vice versa in the end because in the end we don't want to incentivize the ecosystem just through vesting of tokens we want to incentivize it through a sustainable flywheel yeah and just come to coming to an end we are an accelerator for web 3

businesses we are number one in terms of numbers how many where three startups we are helping we have a portfolio of 250 plus companies and 500 plus investors in our Network and feel free if you are considering to become a Founder to reach out or get also access to all this knowledge and all these experts in our Network okay that's it from my site okay yeah [Applause] we have one question can we have a mic here just in your quantitative models how

exactly are you calculating the probability of someone dumping the token yeah um in the model we this is essentially an input where we Define essentially what is the demand for the token and um clearly we can't if you can't Define it but we can give it a range and we can play around with values and then we can see when we can essentially simulate different scenarios so what happens if there is essentially no demand we put in zero or what is there is huge demand and we put in a high number and then we can see like what will happen to our ecosystem to all the other metrics when we change it can I ask a follow-up question just

like have you guys tried a classified based on how often people dump after their vesting period if there's some sort of clusters that are being created based on that yeah good question I mean it's something we are working on so analyzing like what is going on in the overall Market what is like VCS are doing in general and like I said or I showed you before we are also looking in specific sectors and see like um what is the percentage when they receive the token how what is the percentage essentially of these receivers allocating it to utilities or reinvesting it and or using it for staking or whatever or selling it so that is essentially what I showed you

before like the the distribution of behavior in the ecosystem yeah unfortunate hi um great presentation thank you so much for covering a wide array of topics my question is more around is this model also applicable for digital assets where you're looking at forward-facing contracts where the number of tokens are limited and kind of similar model used be used for you know simulation simulating some of those factors that you spoke about um so we can simulate like fixed Supply or also inflationary Supply it doesn't matter both as possible yeah well brilliant thank you awesome thanks again

from outside Adventures [Applause] and we keep on going with more token verification we'll