Finding the right KOLs and measuring their impact

Transcript

Okay, start now. All right, so hi everybody. Always great to be back at HCC in this year in the south of France. Couldn't be a a nicer time of the year. Um so today um I'm Sasha and this is Jackson with me. uh and we'll talk about you know be after the previous talk you look at you know how to engage with PR strategy how to engage with a PR PR coms and what to communicate about we'll look into now how do you do this at scale with data insights um so first a little bit about what we do some of you guys might know us more for the Thai terminal uh so our core offering and the main product that

people know us for mostly is the Thai terminal which is the premier institutional information services um for banks, OTC desk, market makers, uh investors, VCs, uh financial institutions. Um and so what we do is we basically bring all your 50 tabs on your browser into one single consolidated interface to know and get a pulse on crypto what's going on right now. Um from this we built a number of proprietary and advanced data sets to know what was going on. Uh and so we derived a number of other services. So we do APIs that we sell to hedge funds uh quant funds. Uh we do integrations

for example if you're a DEX and you want to add a news feed um that will show all the latest news on the assets that are covered and and and uh have your users trade try to trade a bit more and be more informed or if you want some sentiment data to be displayed on on your decks. Um and then we we also have a range of services trying to use that trove of data and really help the ecosystem um make sense of what's going on, optimize their marketing strategy, optimize their product, engage more people, contribute to the ecosystem, and do this also in a more affordable way.

And that's where um Token Labs comes in. Uh and I have with me Jackson who leads Token Labs at the Thai. Um and so let's get into it. So marketers a long time ago were told the way you engage users, you have this funnel and you come from many brands to fewer and fewer brands and then you have a conversion. Uh and that's how a lot of people thought about marketing. But obviously that changed with the digital the digital world. Um so now it's more of an engagement loop. So you're constantly interacting with your users. you're constantly trying to engage them, build communities, um trying to have them bond together, interact, and at every step of the way

through their lifestyle, through their their journey, you're trying to engage them. And in web 3, it's even more important because in web 3, you not only have your users, but you have who are using your product. Uh you have an economics layer on top of it. Every single thing is economics. every single interaction there is potential value attached to it. Um you have your investors, you have token holders. Um you want to know what they're doing with the tokens. Some of them they're using the tokens to use your products, some of them to invest. Uh and then you have builders and these are the ones building your products. So you really need to know how do you engage them to best to to build the best products um for your users. And so all these three things

come together and with all the different data sets we have access to, we thought, okay, how can we make it a lot more effective for project builders to engage their users? And so I'll turn it over to Jackson. And this is going to be a series of real use, you know, cases, case studies, fictional case studies of things we've done to really attack it from the data perspective and see how we can improve everything. Awesome. Like Sasha said, we're going to go through a couple different big data categories that we work with on the ecosystems and token projects that we partner with. Starting here with some social data. Some big questions that we get. One is like how can I grow my

social presence? How can I engage key voices? And how can I actually understand what's going on and track the narratives that matter? The first case study that we're going to look at is about identifying relevant KS. This is a question that probably every single project has faced at this point. And in this specific example, the ecosystem client didn't really want to do a large budget here. So anything that we shared with them needed to be like really really targeted. We didn't just look at social data to identify KLS here. We actually identified onchain behaviors on the Ethereum blockchain that were most connected to the project use cases and then were were able to use ENS data to find that like web 2, web 3 connection point. This resulted in like a really highly curated list of uh accounts that

they could then do web 2 marketing on. any this would work for any chain that has you know some sort of ENS name service but we find that like one of the biggest challenges in in web3 marketing is like getting that connection point between web two social profiles and uh web 3 wallets. Another big pain point that we hear about from projects is that their marketing their content their social team is spending all day on Twitter just to realize or just just to feel like they're even keeping up with what's going on. What we did is use our proprietary social data from Twitter, from Reddit, basically anywhere where people talk about crypto and then use AI tools to summarize the high impact uh tweets and content. This has been really efficient for a lot of our ecosystem

clients and a lot of the tools are actually used weekly across marketing organizations across the board. Sasha is going to have a couple, you know, real real world examples of of how this plays out and and some of their workflows later on in the presentation, too. All right, switching gears a little bit. Social data is of course extremely important. Um, but a lot of projects also need to understand like what their actual onchain users are doing and we get questions like how can I better understand my onchain users and and do initiatives to drive growth. In this case study, uh, an ecosystem client engaged us to maximize the, uh, reach of a new ecosystem project launch and we took an onchain user segmentation

approach. Here did quantitative analysis to identify a cohort of users that were most likely to engage with this specific new project. In the figure that you see here on the right, we sort of broke down their targeted users into diff different like risk profiles. And those that were most comfortable with um engaging with this new project were the ones that they actually uh focused on this was incredibly successful for them. They didn't have to spend a lot of time on users that probably weren't going to try out this this new project anyway. And it's, you know, it's not hard to see how work like this can inform other onchain initiatives as well, like really targeted airdrops, which we're going to get into in a second. Uh whitelisted app access um for targeted users, more

informed protocol incentives. It's it's been a really interesting way for us to um help projects like optimize new launches and and better understand their users. And now what's amazing with crypto is not only can you access the data about your users, right? If you want to engage retail people, if you want to engage institutional wallets, if you want to engage proumers, but you can also have data on your competitors and beyond that you have direct access to those users. So you can also try and experiment and try and see how can you go further. And so here is a full-fledged experiment that where we've gone even a step further um all the way to building engagement. And the big

thing when you join an ecosystem is okay, how do you reward the real users, the re the really active users? How do you then allow them to figure out what to do? What else they can do? How do you help them grow and have their journey across the ecosystem and then propagate throughout the ecosystem? And so this is uh a partnership we've done uh with Avalanche. Uh and so this is for every single onchain address, every single wallet that's been active on Avalanche. We built a score that gives an idea of how often they use they have a participation score, a consistency score. We're gifying the experience and then that really combines into a chill factor that tells you how big of a user

they are, how consistent they are, how much they've used the ecosystem. If you run out of ideas, you can go put plug in your onchain address and it will tell you there's a recommendation engine that will tell you, oh, maybe you should try this um this application, this this DAP. Um, and at the end of the day, you get your score. So you can use these cores in airdrops and really incentivize um surgically who you want and who is really participating in the livelihood of your ecosystem. Cool. Sticking with some more uh onchain stuff here. A huge question that we get is when projects are faced with decisions like should we do this partnership?

Should we invest this uh BD or strategy spend? How can I actually estimate the potential ROI of the opportunity? So, a really cool case study that we did, one of our big ecosystem clients had the opportunity to do a potential partnership with a major F1 team. This was going to be like a multi-million dollar spend for them. They smartly engaged us and asked us questions like, "What could we actually expect in terms of social and brand growth? What could we actually expect in terms of actual onchain users with any like NFT or other onchain stuff connected with this?" So we took like a case study approach and looked at a similar partnership from another ecosystem and we found some pretty interesting results. First chart in the top right is uh visualizing the social impact of the partnership over

time. We can see here the red bars uh show the various races throughout the season and then we can look at the number of uh discussers about the partnership. They did see that thousands of new accounts were talking about the partnership and we even went in and looked at where these tweets were coming from and we found that tweets in the certain countries and regions of the races throughout the season actually increased by uh you know more than 50% during those race weeks. So it was really successful from a social perspective. On the onchain side we uh looked and did you know various types of wallet analysis to understand retention and and user behavior over time. we found that, you know, fewer than 1% of the wallets actually stuck around. So, while the social impact was pretty

solid, they didn't see a lot of actual onchain growth and this this analysis was instrumental for the client, you know, at the end of the day opting not to do the partnership because they already had a a very strong social presence and we're we're more curious about onchain stuff for this one. We also have done some pretty cool work uh looking at targeted airdrops. In this case, the client had uh used Ethereum data uh and identified wallets that had done certain behaviors on that blockchain and then uh airdrop them uh you know a certain amount of their native token. The goal here was to incentivize activity from wallets that would likely become actual long-term users. and they engaged us to not only

look at airdrop design and how people responded to it, but also do an outcome study of of how it uh influenced long-term user behavior. I'm glad that they asked us to look at airdrop design since we were able to identify a big cohort of wallets that were were gamifying the airdrop and and not uh participating in the way that the the project had hoped. So, it was cool for us to be able to filter out these like inbadstanding wallets and not include them in the analysis. And we also worked with the ecosystem on on future airdrop design to ao avoid pitfalls like this uh going forward. Looking at the actual user data and uh and uh you know end results from the outcome study here we can see in the chart that there's a big jump in

transaction activity and and user activity in in the week following the airdrop. But after that things weren't really too much different than before. We even went in and understood you know what types of contracts users were interacting with. Were they going and doing DeFi stuff? Were they doing NFT stuff? Were they playing games? It was very useful for the client to understand um what happened here. And we we used a lot of this work to help structure better structure future initiatives um down the road. And the biggest takeaway point here is that we need to do things that actually are connected to longerterm user behavior. It cannot just be like one airdrop. It needs to have things down the road and and on and on.

Awesome. Um, we'll I think we're running a little quick on time, so we'll go through some of these a little bit more quickly here. Um, Discord data we found is really helpful for our clients as well. Most of the feedback that we've gotten from native Discord analytics is that it doesn't really give projects a full view of their communities. Those who are at tokens here probably feel the same way. So, we built a custom Discord bot that in fully ingests all the server messages. We found a couple different outcomes that have been really helpful for projects here. One is identifying potential ambassador positions. Basically, use like NLP and find people who are leading the discussion in the community that they might not be aware of and then can structure partnerships, ambassadorships, and even, you know, hire them on as as full-time employees

as well. Um, another uh uh cool outcome here is for like developer focused subreddits. We've been able to use some of our AI summarization tools to better summarize like common pain points in developer experience. It was a huge hit for one of our clients just in just in recent weeks. Awesome. Um, nice segue into developer data here. We also have tons of developer data here at the TAI. We run a number of scrapers looking at um, GitHub commits for for various ecosystems. Um, you know, there are reports like uh that are published every six months.

Sometimes these are not frequent enough for projects to actually make decisions. In this case, uh, one of our ecosystem clients had released a new SDK and done a bunch of other cool Devril stuff in recent months and they wanted to understand the actual impact that this had around the ecosystem. We saw that in a span of about two or three months, developer activity from existing uh, builders on chain jumped by almost 67%. Which was huge and like far outpaced competitors. Um, a lot of what we do here, by the way, it's really helpful for like internal communication around decision-m and getting leadership and everyone else on board. Of course, people on the Devril team want to do more Devril initiatives, but in order to convince leadership, a lot of times we

need to see these like datadriven insights. So now let's pause for a second and think what's my how can I help or how can how can AI help? Is there something we can do? And now when I wake up in the morning, is there a way for me to just say, "Okay, tell me what to write." Um, but for this you need context. And that's what we've brought, right? We've bring all um we'll say, "Okay, who am I? What type of post do I want? And let let's imagine I'm uh working for Poly Market and I want to bring a few I'm have a busy day. I want to bring a few few tweets on the table." So let's say I'm the CMO of Poly Market. I want to

write cool X posts. uh drive viral engagement um and then we'll rely on all the data sets that we have on the TI terminal, the tweets, the news uh and we can add more context and let's see what happens. And here we go. You run it through uh chat GPT type of augmented um LLM and you have a first set of uh of tweets you can post. Not necessarily perfect but usable straight away. Now, in combination with um some of the summarization algorithms that Jackson showed earlier, now you can add even more context. Let's say the big thing I want to focus on today is that Poly

Market is doing a partnership with X and I think this is going to change the game for what we're doing. Use that context. Give me a set of more tailored tweets right for that. And boom. And you get like the full thread of everything like beautifully worded, right? The world's heartbeat. Every second X generates a title wave of conversations. It's me debate. Might not be the type of tone you want for your brand. Might be something different. But the beauty of it is then we can help you put it in the context. Add the tone that you want. add the right data sets, the right information, and you've gone from nothing to a few ideas of content you can put for the day in literally 20

seconds. Um, and that's the beauty of bringing all those data sets together and then having the ability to track the impact you have, build dashboards that are real time, and gives you the feedback to iterate faster and faster and really engage your users. Um, with this I don't know if we have a minute or two for questions. Perfect. So, and if you have more questions, feel free to reach out to us directly. All right, we got at least one.

Um, I mean, kind of awkward, but how does pricing work? So like obviously don't give me the the number but like yeah ballpark or whatever. So the the the long uh short answer a lot cheaper than getting any of those data sets. Uh we try to make it work. The idea is really to be contributing to the space and help the space grow. So we have a range of possibilities but the idea is to make it as affordable and work with with projects based on the stage they're at. So I mean it looks really interesting to us. So we would uh contact you guys for a quote as I as you know. So yeah. Yeah for sure.

Hey guys, thanks for the talk. Um you mentioned developerdriven insights and I saw number of commits is what other data points you use besides number of commits for developer insights? Yeah, that is a good question because not all commits are created equally and there's so many things going on behind the scenes. I mean, daily, weekly, monthly active developers is usually a good one. We can further stratify the definition of a developer. Are they a full-time developer? Are they a part-time developer? Are they a one-time developer? What does it mean when a developer has churned? Uh, another cool thing that we can do is let's say you onboard a new developer, you've onboarded a cohort of new developers, what is their history? Are they

completely new to crypto? Did they come from you know XY ecosystem? What is like their typical development path? What are like the flows that we're seeing? So it's a really interesting data set and like lots of cool ways we can answer that question. So like you can look at programming languages for example like if you're a project you're like we want to grow our uh grow the number of builders that we have. There's so many cool ways you can approach it. You can say here's some target ecosystems that have overlap in terms of developer history and you know language contributions and other other various things like that. Very interesting. Thank you. Thank you guys all very much for listening.

[Applause]