A Look at On-Chain Data of 25 Tokens on the Ethereum Blockchain
In the following post, I’ll take a look at tokens on the Ethereum network. Tokens using the ERC-20 protocol on the Ethereum blockchain have been the driver of the ICO hype-cycle in 2017. So I find it quite interesting to have a look how these tokens “are doing” and what interesting insights may lie in the on-chain data. To do so, I used data from Bloxy, a wonderful site that brings a lot of transparency and resources to the Ethereum ecosystem. The data is collected at the beginning of October.
In January 2017, 25 different ERC-20 tokens were operative on the Ethereum blockchain. Until October 2018, this number grew to over 3,800.
I picked 25 tokens (more or less randomly) to compare them with another. I picked projects based on transaction volume, ETH collected, subjective relevance and other statistics. I did not include The DAO, as the token isn’t relevant anymore. The same could be said for the ERC-20 tokens of EOS or Tron, but I still included them. In total, these tokens have been transferred 12.6 million times. ETH Gas Station shows a standard transaction fee of $0.014, what would amount to transaction costs of $177k for all these transactions. If counted in that token transfers usually cost more than standard ETH transactions, the sum will be much higher.
Yet, I want to solely look at actual on-chain data, not on secondary market stats, social media or development activity. Therefore, I won’t include any of these metrics in the following post. They are definitely very interesting but also very easy to distort. On-chain data cannot be tampered with, which is why I only include it in this post.
Transactions statistics
Based on the total transaction count, EOS leads the 50 tokens with 3.1 million transactions. Yet, EOS tokens have been swapped to the EOS blockchain and weren’t transferred since 2nd of June. Tron (TRX) is the number two project with 2.4 million transactions and OMG ranks number three with 1.4 million transactions. Atonomi (ATMI) and Envion (EVN) are placed on the last two spots with 52.9k and 36.9k total transactions.
By looking at transactions per day, time-based effects can be taken into account. While Golem (GNT) is the number 4 token in the total transactions graph, it falls to rank 9 due to the token being transferable since November 2016. EOS and TRX remain the top two projects with 6.6k and 5.1k daily transactions. Yet, EOS tokens aren’t transferable anymore since June, which is why this average will remain the same in the future. TRX swapped their tokens to the Tron main net but ERC-20 tokens are still transferable. Still, the fact that the main net has launched suggests that the average daily transfers on the Ethereum network will likely decline over time. Both Basic Attention Token (3,351) and Pundi X (2,142) jump forward in the ranks. These two tokens are not only means of financing to be swapped into other tokens but do propose a real utility, or rather a service, on Ethereum.
Holders, Receivers & Senders
The graphic may look a little complex at first sight but contains quite relevant data for the assessment of real usage of the tokens. The holders count represents an indicator how many users or investors are in possession of the respective tokens. Yet, as account creation is free and easy to do on Ethereum, any single person or entity is able to create as many addresses as the want and send tokens there. The unique receivers and senders count hold a little more insights, as they allow to account for time-based changes. It doesn’t make a lot of sense to compare current holders count with unique senders or receivers, as the latter two involve data that is gathered over the full life-cycle of a token and holders count only represents a current snapshot. TRX (1m+), OMG (644k), BNB (297k) and EOS (250k) are the four tokens with more than 100k current holders. QTUM (13k) and ATMI (7.5k) are the tokens with the least amount of holders. Yet, QTUM is also a token that was swapped to another chain. Therefore, the token holders aren’t that relevant by now.
A high amount of receiver and a respective low amount of senders can be an indicator that (1) the token isn’t used, (2) the token isn’t used on the blockchain but inside a centralized platform or (3) the tokens use isn’t enforced through transactions (but maybe by proving ownership of it). By looking at the 25 tokens, three projects stand out: BNB (13.48%) has the highest comparative difference between receivers (high) and senders (low). BNB falls into category (2), as the main use of the token is to pay (reduced) transaction fees on the Binance crypto-exchange. Tokens are stored centrally in the Ethereum wallet of Binance, which is why (1) tokens aren’t often transferred out of the exchange and (2) in most cases tokens end up in the single deposit address of Binance. FDZ (13.88%) theoretically falls in a similar category: On the Friendz website the company states various utilities of the token, such as discounted purchases or reimbursement for engagement within the app. This would mean that tokens will in most cases also be stored centrally by the app and be distributed through an opt-out process. I personally didn’t test the app but somehow doubt the real utility and usage of the token at this time. BDG (16.09%) is a token that can be used to access the BItDegree online education platform. Instructors are paid though the tokens and users pay with tokens. At the time of writing this article, BDG shows a 24h trading volume of 121 USD and the token has been transferred 26 times over the last 24h. This may be an indicator that there is no real demand for the token and/or the product that BitDegree offers.
The upper graph shows the relation between full and circulating supply of the 25 tokens. It can be seen that 16 tokens have 100% circulating supply, while 9 tokens have a circulating supply between 97% and 36%.
Token Sale Statistics
Unfortunately, I was only able to collect data for 18 of the projects, as the others seem to have used a different approach of their token sale that didn’t allow me to directly pull the data from the token address. With 7.2m ETH, EOS clearly ranks the number one project. This can be explained due to their very smart approach of a token sale. EOS sold part of its tokens in a single sale and then auctioned the same small amount of tokens for about a year against ETH. By listing EOS on exchanges after the initial sale, the daily auctions more or less represented arbitrage contracts that economically made sense to bid for or not based on the current price of tokens on secondary markets. As the price of EOS tokens rose quite high in secondary markets, the amount of ETH that was collected through these daily auctions also rose significantly.
Golem (GNT) is the token sale that received the second biggest allocation of ETH. This can be explained through the age of the project. ETH was valued significantly less than today at the time of the sale. Older projects like GNT were able to collect big amounts of ETH that, due to the rise in price, provides them way more financial capital than they initially asked for (or need). This may be an indicator that these projects are too big to fail but could also be a sign that the incentives and demands are lacking. If projects know that they own enough capital to finance their venture for the next 50 years, development effort may lack efficiency and processes may slow down.
The number of transactions shows how often a token sale contract was called and the number of unique buyers adjusts this data for wallets that contributed multiple times. The biggest difference between unique buyers and total transactions is EOS, while projects like REQ, EVN, MANA, SNT and ZRX are also ranked comparatively lower. This is illustrated in the following graphic.
By looking at the ratio between the number of transactions and the number of unique token buyers, ATMI, BAT, WPR and KNC stand out with a score close to 100%. This can be explained due to the fact that these projects (1) implemented capped contribution whitelists, where contributors were only able to participate up to a certain amount of fiat or ETH in time-based intervals (i.e. KNC) or (2) the token sale was immediate sold out and so-called “GAS wars” happened, where contributors are paying a huge amount of transaction fees to become one of the lucky ones to get access to the tokens (i.e. BAT). On the other side of the statistic, EOS has a ratio of around 27.8%, which shows that lots of participants participated various times in the token sale. This can be explained by the economic model of the EOS sale that incentivized contributors to participate in regular intervals for arbitrage efforts.
It can be seen that the projects that achieved a score of close to 100% in the last statistic were able to sell their tokens in a very short period of time, while the sale of projects with a higher score took a longer time, which is logical.
By adjusting transactions and unique token buyers per day, Status (SNT) takes the number one spot of token sale transactions per day but not for unique buyers. The SNT token sale had a so-called hidden cap that was only communicated once a certain threshold of the sale was reached. Therefore, various contributors participated more than once, as the cap wasn’t shown or reached yet. WPR is the project with the most unique buyers on its day of token sale. This can be explained by the individual, personal caps that interested contributors were assigned during the first five hours of the token sale. This lead many people to participate at comparatively small sizes, as is shown in the next statistic.
WPR and ATMI both show very small individual contribution sized for their token sale that can be explained by the capped structure of the sale and the many participants. GNT has an average contribution size of close to 1,600 ETH. ETH was priced at around $10.5 at the time of the sale, which amounts to an average contribution of $16,800.
When the full amount of ETH raised is adjusted for the number of days the token sale took place, EOS falls onto rank number 10 with over 21k ETH daily.
Distribution of Tokens
The distribution of tokens can be an indicator for the state of a project regarding market adoption, (de-)centralization and other factors. In the following, I’ll compare the distribution of tokens for the number 1, top 10 and top 100 addresses of the 25 projects. I excluded tokens that were sent to the 0-address, as they are burned and taken out of circulation. In the case of QTUM, lots of wallets were more or less other 0-addresses. I didn’t exclude them in this post. I will not assess individual addresses in this post. Please keep in mind that various factors, like exchange pooling addresses, additional burn addresses, vesting contracts etc. play a role and likely distort the validity of the data to some point.
BNB leads the 25 projects, as the number one wallet hold over 50% of all existing tokens, being followed by NPXS (43.9%) and ZIL (38.6%). REP (4.4%), EVN (4.6%) and MANA (7.4%) are the three projects with the comparatively smallest number one token address.
In the ranking of the top 10 token wallets BNB falls down to rank 10 with “only” 59%. QTUM (after its token swap with many burn addresses) takes the number one spot with 91.9%. SNT (76.7%), NPXS (70%) and WPR (65.7%) rank number two to four. EVN (9.5%) has the smallest share of tokens in the top 10 wallets.
The top three spots remain the same for the top 100 wallets with QTUM (99.8), SNT (89.6%) and NPXS (87.8%). ZIL jumps from spot six in the top ten wallets to number four for the top 100 wallets with a share of 86.6% and Bancor (BNT) jumps from rank number 10 to number five (84.7%). EVN remains at the last spot, as the top 100 wallets hold 19.5% of the full token supply. The wallets 11 to 100 of TRX only hold 0.31% of the token supply, which is why the token is on the second last spot with 35.4% of tokens being held by the top 100 wallets.
A comparatively high number of tokens in the top 100 accounts may be an indicator that the major supply of the token is being held by single entities that could theoretically market sell these tokens, resulting in a huge drop in the underlying price on secondary exchanges.
Conclusion
I believe that I found some interesting insights by looking at some simple static data of ERC-20 tokens on the Ethereum blockchain. Tokens without a “real” use case on the Ethereum blockchain apart from being a financing and swapping tool, such as EOS, TRX, QTUM or ZIL, pocket the highest transaction count (EOS), holders count (TRX) or fundraising sum (EOS). In the future it might make sense to do a more fitting preselection of tokens in regards to the respective insights that an evaluation is looking for. This time I included various types of tokens that act in terms of specific use cases that may not be suitable for a comprehensive comparison. While the statistics are interesting, QTUM, EOS and TRX tokens are more or less not relevant on the Ethereum blockchain anymore. Next time it could make sense to focus on a specific sector of ERC-20 tokens, like DApp tokens. Additionally, the data only consists of static snapshot data. It may be quite interesting to look at time-series data to compare activity of tokens over time.
Here’s a brief overview that shows how time-dependent the transaction volume of tokens on Ethereum is:
Big thanks to Bloxy for their efforts to provide open access to Ethereum data. All data that I used can be found there. Make sure to check our their analytics and API.
Feel free to contact me at ante@sicos.io or via LinkedIn. I’m a co-founder of STOKR, a crowd-investment platform powered by Ethereum that will allow for EU-compliant security token offerings (STOs). Everyday investors will directly be able to fund innovative start-ups and in return receive a share of the venture’s future profits.
