Leveraging On-Chain Data to Assess Web3 Protocol Health

Transcript

three two one all right hello everyone uh we often hear that on-chain data is open and that is transparent and at the same time we hear that there is a lot of unchain analytics that we can do but the question is is that really the case in in real life in in practice so today what I'm going to talk about is the problem of on-chain data and yeah how to deal with it through standardization of metrics I am mihai I lead the web 3 protocol research team at

massari and yeah together with my team we put a lot of yeah data standardization and metrics for over 40 protocols out there so we're covering l1's l2s we're covering also uh web3 mid layer infrastructure and consumer and one example of actually our work is exactly what I have at my back and the typical consumers of these charts are Enterprises they are investors they are protocols themselves but also crypto enthusiasts and we very often get yeah a series of questions

from from them with respect to how are we able to standardize and Source all this data to create these these charts and how are you actually able to these charts all together and and display them so in order for me to actually get you step by step through how we get to to this output I'm going to start with the very very basic things first so essentially what we do is we're analyzing who does what and when on chain and how this looks actually in practice is a little bit like this so on the darker background we have a transaction in in HEX and then next to it we have

ether scan so it's that interpretation of the transaction and there if we look carefully we can see who did what and when and besides that we have some protocols that are actually a bit friendlier and that aggregate different transactions altogether and you can see a series of transactions and also there we have in the columns like who does what and when now the question is how do we put this thing at work so what can we do with with this and essentially it boils down to classifying transactions and classifying who does those transactions and one typical example here is to aggregate revenue of a protocol so here

in this example we see yeah supply side Revenue that is accrued by node operators so by the supply side participants of of protocols but then at the same time I have protocol Revenue that is actually accrued by the transaction fees or like the usage fees that are that are accrued by the protocol so this is one example here of what we can do and whenever I'm showing this to Jimmy the investor the question that I get back from him is actually okay uh mija you give me one metric you give me Revenue about one protocol however I would like to put this in context I'd really like to see how this protocol

performs against other protocols I would like to see to what extent actually this Revenue generation is sustainable or not to what extent the protocol makes enough money who uses the protocol is that uses sustainable over time are the users sticky or yeah are they turning so all these questions are are being asked by by investors and um yeah when when getting back to them and saying okay so what do you actually need besides actually having like metrics in in isolation after like countless conversations we we come up with a list of requirements so a list of yeah Target protocols that are looking

at so yeah normally let's say we take lending protocols in the um yeah looking at which of them have growing revenue and deposits so then I'm asking Jimmy again so why aren't you able to actually do this today so this ball is down to to the problem that we we all have and it's essentially like this so starting with data source a so Jimmy goes on a portal and tries to aggregate um yeah different metrics for for his protocols and you essentially what you see here you have coverage only for one metric and then for all the other metrics yeah the coverage is very very

sparse and then of course like data source a is not enough so you go in and check data source B and here the situation looks a bit differently so you have like better coverage in terms of like different metrics however yeah not all uh the protocols are are being covered and of course whenever you try to merge actually these two data sources you get something like this which is a combination between data source a and data source B and you see that still some metrics are completely missing but then I'd like to point the attention to revenue so in this case you have Revenue coming from like uh yeah both both sources and Jimmy is telling me well

actually the problem that I'm having is I don't know whether I should choose like data source a or I should choose data source B and the problem here is the values are very different from each other and there is also no given methodology in what way these data sources are actually being computed so there isn't there isn't something that one can track over timing can go back and actually check whether those numbers make sense yeah so this is where we actually come into play and uh this is what we are building at massari so uh we are building standardization schemas together with the graph that are called subgraphs where we Define standards for

the industry so we Define standards with respect to what do we understand with respect to usage what do we understand with respect to Value accrual and revenue and what you understand with respect to governance and quantifying governance decisions so this is a snapshot of the open source software that yeah we write together in a partnership with the graph and yeah if you go to subgraphs.massari.io you can see that we've already developed over 100 such standardized subdrafts for quite a lot of protocols out there in the industry and if you want to see like what the output of this is so you see like a list

of protocols here and we essentially have the same standardized methodology for which we compute different types of of Revenue now how do we use this actually in our products how do we use this to uh yeah come up with uh um with charts so I'm going to show you a couple of charts here so one the first one is related to usage so essentially what we did here is my colleague Ali went to um yeah the lens ecosystem and she was able to aggregate all the users of applications that are built in the lens ecosystem and what she was able to see is that Leinster is the application in

the ecosystem that actually attracts most of the users from other applications and this is only possible due to yeah the openness of the data and being able to actually quantify and track yeah who was the user across all of these different different applications if we move into value crawl or into Revenue in terms of comparability my colleague Sammy here has put together yeah total revenue for different deep in or like decentralized physical infrastructure networks and what he was able to see in this case is that two protocols on the decentralized storage and decentralized compute space are actually holding quite well with respect

to revenue although the whole sector is pretty much declining in terms of Revenue at the moment and then if we go to staking and inflation so this is actually the first chart that I've I've shown in in the presentation so here what what we can see is a beautiful mechanism that this protocol is using so essentially the protocol has a dynamic adjustment of the inflation so whenever the staking participation is below 50 percent the inflation is actually growing and whenever staking participation is above the target equilibrium level of 50 actually the inflation goes down so you can see here since of February we have inflation

growing because taking participation suddenly dropped below 50 percent and my interest here over the next quarter is is actually to look whether sticking participation will will increase over time being incentivized by uh yeah this growth in inflation now we put together not only charts but we also put together tables and standardized tables so this is how yeah our standardization looks like for decentralized compute uh protocol in this case it's life peer so yeah we have financials and key performance indicators as well as staking metrics that we put together what we do is we track this on a daily basis and these are snapshots on on quarterly basis we

aggregate all of this into quarterly reports that we publish for almost 40 protocols by now all right so this is only the tip of the iceberg actually at Missouri we do much more than this um so we are a company of uh 100 plus crypto Geeks essentially half of the company is out of Engineers how the other half of the company is researchers and uh yeah what we do is we do a lot of unchain data engineering uh we aggregate a lot of yeah governance decisions Venture deals uh diligence reports and uh yeah of course uh research research reports all right so in case you'd like to find

out more about like our our next product releases we're going to organize mesari mainnet conference in New York in September we're going to have a couple of product launches there around around data in case you want to play with our protocol Explorer please go ahead please do so in case you want to read our research and also our engineering blogs they are like pasted over there and also in case you want to contribute to our open source development efforts um yeah our our developers are happy to help and to include you in the community and thank you so much we have five minutes four questions if

anyone um yeah thanks that's really interesting and um you know all of us have tried looking at chain data know these pain points um one of the things I was wondering about is are you also planning on including uh metrics around the mall standard product Market fit metrics like retention of users maybe cost of acquisition lifetime value because you know stuff like TVO and so on it's kind of like a bit of a vanity metric like if you're actually using if you want to analyze how good a protocol is it's a bit you know you know not not always that that useful yeah I love your question because I've been thinking exactly about including vanity metrics and actually presenting some of them in this slides but yeah to

your point we're looking actively at usage actually so we're putting together several analysis on especially Yahoo sticks with the protocol for longer and is not there only for you know some initial incentives so we're definitely looking at that and yeah in terms of uh client acquisition costs it's essentially how much of the rewards that you give away are actually taken by some of the users that maybe might might flock so might might just like leave the protocol after two weeks so yeah we're definitely looking into that in one of the slides that I've shown on on usage is this actually is actually looking at uh yeah two to what extent actually users leave one application to go into the other one

thank you talking about vanity metrics to to what extent are you planning to integrate data that is not on chain because many of the non-vanity things needs additional insights Beyond on-chained transactions yeah so we work with a lot of off-chain data as a matter of fact and this is unavoidable for us um so in order to have like completeness of metrics you will have something that is out there on chain and then we have to fit the holes with off-chain data and yeah we're speaking with protocols and we're trying to determine them to expose their off-chain data that they have into block explorers into dashboards and then working together with them into yeah

defining apis and then probably also getting them a subgraph and being able to standardize their data yeah great question we're having that problem on a daily basis yeah I'm actually have a question for the online data source I'm a little bit lost I'm going to say there are two different sources giving you uh different values if the on-channed data has its own ground shoes wouldn't just be there like right right so this is one example one hypothetical example of data source a and data source B that a typical user or a typical investor will have a look at so maybe you can think of like you know the major data aggregators out there you can pick your favorite one

and then it's already secondhand that that would be data from third-party providers and this is essentially what investors or uh yeah other yeah crypto lovers will will have a look at um so what we're saying is actually we have to go to the ground source and we have to go to on-chain data to to source that yeah great question thank you very much unfortunately we don't have any more time you know the drill if you want to ask more questions reach outside for me hi thank you