Ambre Soubiran: The Business of Analyzing Blockchain Data (from EthCC 3)

Transcript

so many different types of people when it comes to breakfast so no it's not my usual breakfast I usually have yogurt fruits and nuts but I didn't have time this morning I'm here with Alma should be home she's the CEO of Keiko and so Keiko is a French company based here in Paris and specializes in data and analytics around crypto and so I want to get a bit of a yeah but a background on yourself and also what you guys are doing so yeah you came into Keiko the story of how you you came to CEO of

Keiko is kind of interesting one so yeah tell us a bit about your background and how that came about yeah of course thanks thanks for having me so yeah indeed my background is in traditional finance I was working for HSBC for about a decade structuring equity derivatives so using a lot of data interested in data interested in financial markets and money in general but I was not kind of from a more start-up intrapreneurial background I got introduced to blockchain back in 2013 started looking at it investing it personally getting more and more interested making quite a bit of noise even within the scope of my professional word around blockchain and I met a guy called Pascal he's the current CEO at

Leger Pascal gucci is a former Crito executive he had just IPO the company on the NASDAQ and had discovered blockchain we got in touch back then it was 2014 and he was starting Keiko and so Keiko is a market data company in the space of blockchain I met him back then we had a great conversation we discussed on how we could one day sell market data around cryptocurrencies to financial institutions but it was way too early there was no market back then this was like 2016 2020 for 2014 ok guys been around for longer than people know and so press Califano the company in 2014 I was like having just regular interactions and

we had a great like discussions around financial markets and market data and blockchain but nothing formalized at that stage he started the company hired an entire engineering team and they built an awesome product which we still use bits of today however there was no market so two years later late 2016 the initial team I think started being a bit discouraged which I can understand knowing how hard it is to run a company and must be even harder when you don't have a market yet yeah so everybody kind of got a bit tired and left and Pascal was also one of the board members of ledger and ledger had grown significantly in those years because it was a more retailed focused market and so he told me I'm going to join ledger

and I'm kind of pretty much dropping keiko because there's no market and it's there's no team anymore I need to move on hmm at that point I left HSBC and I invested in Kaikohe acquired a controlling stake of the company and relaunched keiko so we were still leveraging back then on the historical data that Pascal had been collected over the past years and we had the brand we had the reputation which was great and also when people started needing data in 2017 we already had some inventory and we had a product and we were ready to ship so it was a double-edged sword being too early and having a product when people needed the product I think he told me once that basically when when pescado wanted to leave he brought you on to sort of get the company into in

mistake where he could sell it or something like that yeah and then you started just improving you know the business and and generating revenue and yeah so the original reason kind of I started working with keiko was we got an offer on the table which was more kind of an you know verbal offer that Pascal received one day talking to a big company in the crypto space won't say the name and Pascal told me if you can get kind of Keiko back an order you'll get a cuts on the sale price so that's kind of how I started getting involved more on a consulting basis and then I realized that there is something bigger than just this sale that could happen and I told Pascal like let's not sell it let's restart it and I'm in and

I'm keen to reinvest in the company and so the so it's hard that's how it happened cool yeah that's cool story so uh what do you guys do like what is what is the blockchain data analytics company so we're basically connected to every single crypto exchange in the world we're connected to over 100 exchanges and for every exchange that we cover we track all the trading related data for every listed instrument so we cover I think something over 30,000 instruments if you ask me a simple questions such as what is the price of Bitcoin it sounds like an easy question it's actually not an easy question if you what is the price of Bitcoin well right now I hope it's higher than I think but but

basically it's not an easy question price discovery in the crypto markets is purely based on offer and demand and so what we do is were connected to all these trading venues and we track or what's called like order book data or double data is all the buy orders and all the sell orders the combination of all these orders create what we call an order book and represents for every exchange the depth and the structure of the markets like what how much volume is there at what different price levels when you look at the intersection of the price curve and the upper curve you get trading happening and all those transaction those consecutive transactions executed trades is how price discovery is made and those to reflect the actual share will not share price token price for those different assets so the price of Bitcoin is

technically nothing less than the last price at which transaction was executed and that's the information we provide to different market participants all the order book data and all the trade data so you're connected to every exchange virtually virtually every exchange and every and then also collecting all the data from all the trading pairs yeah what like how many trading pairs does that represent in difference what we call instruments so for example Bitcoin dollar on coin base is one instrument Bitcoin dollar and crackin is a different instrument so 30,000 but like 30,000 individual unique instruments which is a specific asset pair on a given exchange that's massive it's huge

and I suppose it's growing to like very we well depending on the days we add from like a dozen to hundreds of new instruments on a daily basis we're talking about hundreds of terabytes of data we've been storing that since 2013 so you also because you also obviously have historical data yeah so we're basically kind of storing a historical ledger some why it's not a ledger historical database of every single transaction that has been executed over all the different trading venues in the world where people buy and sell crypto assets and not only executed transactions but also all the buy orders and all the sell orders without them being executed and that's what we call the depth of the market in the structure of the market

you can see sometimes where there's like strong like imbalances on one side on the buy side or the south side this is actually really important information for market participants to know what does the structure of a specific market looks like across different venues what do people use this data for like who are your clients and why are they wanting all this data so we're like solely b2b we only work with companies those are either crypto funds quant funds people that need a lot of historical data to run back testing to trade algos these kind of guys so they use us for like live monitoring and real-time trading we work with researchers we work with exchanges interestingly exchanges of

course they have their own data but they use us to be able to know how they compare to other exchanges because everything is standardized of course everything is kind of available in one place in the same format so we work with exchangers researchers people such as like the bitwise guys are doing research or that are building ETFs people that are building indices benchmarks we work with all top universities around the world that are carrying out research on crypto market makers yeah anybody that really needs a reliable kind of price infrastructure and also people who need valuation data like really reliable valuation data so that they can value assets that they have either under custody or for providing services such as ledger when you plug in your key on

your laptop it tells you they can read the blockchain right they know your address and they know how many bitcoins you have but the app actually tells you how much is it worth in dollars they need to have a reliable kind of third party source for an exchange rate so they can provide to their users and they cannot be themselves figuring out what's the price of Bitcoin right because you have all this data I mean you provide this data to your clients and then you know they'll go and exploit that data but are the things that you extract from the data that like what kind of valuable insights are you extracting from the data that you can provide to your customers in addition to just the raw data do you have sort of like layers of analytics on top of the data itself that you also sell or is it no question it's

kind of strictly the data in the sense that we don't you know we don't have a platform a terminal like a UI with fancy charts or whatever we don't do that at all however we create a lot of data derivatives so for example like when we have all the buy and sell orders of an order book on a given market we calculate things such as slippage slippage for a given size order for example if you ask me what's the slippage on coinbase for a Bitcoin dollar for a 50k dollar order what we're going to tell you is we're not going to throw the order book at you we're going to tell you it actually in order for the order book to absorb a 50k order it will cost you eight bits of the mid-price for example so those are kind of data sets that we provide they're not raw data but

they're not there they're kind of actionable data that you can use directly as a consumer but it's not it's not raw but it's not kind of processed either then we we just were launching technically next week a research fact sheets this is like four pages of data charts and in correlation analysis it's really like fully quantitative only we don't do any analysis we don't do any interpretation we don't try to say okay this is why the price went up or down however it's going to be four pages of like technical analysis on fundamentals for a Bitcoin theorem and litecoin I suppose there's there's probably lots of business opportunities here for you to

produce a lot of derivative data and analysis and this sort of thing is this just not in sort of you like what you want to provide or you just you know depending on how the company will grow we might start like you know hiring researchers for now it's more data scientists it's we're very engineering focused team for now and we partner with people that are doing research we're going to announce a few partnership with known media companies in the space in the coming month and the idea is we said kind of a level before we're really infrastructure focus so we enable people to do research on our data we provide aggregated data data derivatives as you mentioned but we're not like in the analysis game yet okay I wonder like curiously some of the

technical challenges that exist in connecting with you know 100 exchanges and you know pulling 30,000 instruments at this sort of second it's hard it's very interesting because technically the data is there it's available like if you're relatively you know kind of nerdy you can connect to an exchange API and start pulling Bitcoin dollar data just what is really hard is doing it at scale in a highly reliable fashion and none of these api's are standard I suppose so you must have had to build all the infrastructure to connect to any kind of API yeah based on because all the more structured differently absolutely yeah and there's like one more thing that's really horrible is just even taxonomy

like there's no center date there's no standard and how you even once you're connected to an exchange API the like calling the data for a Bitcoin dollar is a different kind of name for the crypto refer to those assets depending on the exchange so actually we're 16 in the company and like two-thirds of the team are engineers it's really a very kind of tech company we have a full-on you know on like rotating organization so that there's all right always somebody who's checking and monitoring that everything is live we is a full on-call organization and and definitely like the the cost of doing that is also why it makes us valuable is that if every

crypto fund and the space had to hire a bunch of Engineers to do what we do pay for the server costs and internalize all of that it would just be really expensive so there's also a kind of economy of scale of we have a full team of like engineers working around the clock to deliver a low latency millisecond level of precision product for data so that you don't have to internalize that and the out like paying our costs is a fraction of what it would cost to any company to build it in a house and or so are you have you been able to sort of set standards at all I like or you know cuz you're talking to all the exchanges you know you have the ability to tell the exchange like hey you the this is the data this is the API

standards that everybody should be using you would make everybody's life easier like you know you guys but also other companies doing the same thing those those clients that might be using the API like we're working we've launched a working group about six months ago with remember of this thing called the object Management Group and they're a big international nonprofit that is defining standards right they have names for like fertilizer they also have this thing called open Figgy which is the financial identifier sorry financial instruments unique identifier we are basically creating a database of unique identifiers for every crypto asset instrument imagine a world where

we start having financial instrument referring to Bitcoin like what is the price of Bitcoin again like it comes back to our question you have to specify in the contract are we leveraging the price of Bitcoin dollar on coinbase are we taking the average daily price of Bitcoin over all places where it's traded like we have to have a way to refer to those financial instruments in order to start leveraging them for financial contracts or applications so we're on a workshop we're hoping to deliver by the summer an infrastructure with like a taxonomy report for how can we identify in a unique way all those instruments I know I know you have to go but before you do you tell me a little bit about what this

conference and your involvement with it so we we love at CC we've been we've been sponsors for the past three editions we've hired a bunch of our team members here we've met partners we've done amazing deals met customers and I'm especially kind of involved in this year's edition I buy I am in Kaikohe we've been running a VC track so last year during see see I was myself fundraising for Keiko for our seed round and I realized well two things one thing that I already knew which is they're a bunch of amazing people building awesome tech and projects and companies that are coming to this conference but second there are investors coming there are investors that are trying to deploy capital in the space and they just there's there is

very hard there's definitely like a bridge that needs to be created between investors and blockchain companies and I suffered from that last year I was trying to find VCS at the coffee area at the swag area I kind of haul way and like we were like finding each other and you're the epicenter booth it was kind of not structured and so I basically proposed that we run a VC track within the conference where there's a real framework for investors to meet startups that are currently fundraising so we had 32 companies pitching this year 27 VCS attending it's really like success hopefully in a couple of months I'll be able to say that you know X millions will deploy tanks to the track yeah that would be great

and so these VCS are coming from where all over the world we have some San Francisco LA New York Berlin Zurich zhuge London there's a few like from Paris of course the event being here in France it's easier for the French VCS to attend I have a couple of companies from Hong Kong and we know that there's this coronavirus situation so we have actually livestream in place where investors have been connecting from all over the world - just like view the pitch and listen to the to the presentations and at the end of the track tomorrow night we'll share everybody's contact details all the pitch decks to all the investors vice-versa and hopefully things will happen magic will happen then cool thanks a lot Thanks