This website has 50 million users. Our app has been downloaded by 20 million people and 6 million of them have activated it. My YouTube channel has 1 million subscribers but only 500 thousand watch the videos.
We are fixated on user numbers.
Somewhat rightly so.
Number of users show us where we stand in our sector. It matters to say that 20% of the users are here. Users and usage help us to have healthy comparisons with the competitors. These analysis when placed in a timeline help us to construct time based analysis that can evaluate cause and effect over time. They also excite and motivate us. Thus we tend to say only 1% of the potential user base knows of the product, it has a huge potential.
Did you say total number of wallets?
On September 11, in the Medium publication The Capital Mark Helfman published an article on Bitcoin use cases which was enjoyable to read. In that article a sub-heading have been the catalyst for this article: “Mark, bitcoin is going mainstream…”
It was followed by the graph that showed the number of Bitcoin users were increasing:
It seems that fact checking has become a reflex and especially so if there is table of figures gently trending towards one side. First of all I wondered how did they find the 50 million Bitcoin user figure and if such a neat increase was plausible. I have great respect for Statista. I follow their publications and believe that they do public good. So I was even more intrigued to look more closely.
When I checked the source Statista quoted in their website, I found out that it was Blockchain.com’s Blockchain Wallets data they referenced. In reality, only Blockchain.com can verify their own figure but it seems that Statista’s title “Number of Blockchain wallet users worldwide” is somewhat unclear. If they mean the Blockchain.com’s wallets are increasing, then they are right depending on the data source from Blockchain.com. However if the usage is, as per Mark Helfman’s article, to refer to Bitcoin users in general, then there is an inadvertent misunderstanding because the data is company specific even though the company has “the” generic name such as blockchain.
In order to avoid future confusion and to aid the general knowledge level, in this article I will try to clarify the concepts of users, wallets and addresses with numerical data where possible.
An address; on the Bitcoin blockchain, is 26–35 character long reference properly cryptographic for receiving and sending the currency of the realm. Its balance could be 0 or more. It might never be used or be used more than once. It can be created online or offline through various channels.
An address is not a user. A user can have multiple addresses. An address does not care who the user is. Anyone with the proper keys to the address can use the address and its available balance.
As per Glassnode data, on September 10 2020, there are 704 million 581 thousand 259 Bitcoin addresses. The corresponding figure was 464 million at the end of 2018 and 595 million at the end of 2019. The constant increase in the number of addresses is not fully linked to the increase in the number of users. The primary reason lies in the single use of some addresses for privacy and security preferences.
Below is the Glassnode data for active addresses that have been used at least once per day and it shows 600 thousand to 1 million average daily active users from 2017 to 2020. The only meaningful relationship of this data to total addresses is the 2012–2017 period directional parallelism that corresponds to increasing interest from a low base.
Now let’s check the addresses with a non-zero balance from Glassnode. As of September 2020, there are approximately 31 million addresses with a positive balance.
I would like you to pay attention to this. This figure shows the maximum current users on-chain at the moment. I will come to exchange accounts later on as it is a different additive point. Back here, these 31 million addresses belong to maximum 31 million and most probably to less than 31 million users. This is the first instance so far where we can infer a connection between addresses and user numbers.
The graph shows a decline of 6.5 million in the number of positive balance addresses in early 2018. This is properly the linked to the users leaving this market due to the price correction. If the same logic is extrapolated further out, assuming no major change in addresses per user ratio, then the user numbers of late 2017 have been breached up by December 2019.
First of all, Bitcoin wallets are not real life wallets per se. In real life, one puts money in a wallet. In the Bitcoin blockchain, one puts addresses and its cryptographic keys in the wallets. Wallet is a software and through it one accesses and manages the addresses regarding sending, receiving and encrypting.
A wallet can hold multiple addresses. The same address (depending on its specifications) can be managed by more than one wallet. An address does not necessarily need a wallet.
Do we know how many wallets are out there?
This is most enjoyable question to pose because its answer does not hold any structural value to the community in general. Wallet data is fundamentally a company level information. Naturally it matters to the company and to the developer how many wallets of theirs is in use.
Analytically speaking, wallet data would have natural correlations with other blockchain data, it will have an indirect use case but that pales in comparison to the importance of address level data.
Then why do we look the wallet data? In broad terms, it is so because it is a source of great publicity for the companies that create and market those wallets. When Blockchain.com says it has 50 million wallets and the number is constantly rising, then it has a lot more about the company than about the Bitcoin blockchain. By the way, to their credit, they are valuable member of the community.
Can the big exchanges remain silent? For example according to Binance Blog one can learn the benefits of the Trust wallet which Binance itself has bought earlier. Great marketing opportunity combined with real technological information.
Now I really want you to pay attention: The critical importance of the wallet is that the wallet is a major gateway to the world of crypto currency. It is highly likely that the wallet user has a meaningful tendency to use the services of, in essence become a customer of, the company who provides the wallet (usually at no extra visible charge). Therefore, secure and easy to use wallets are a significant customer acquisition platform for the exchanges and wallet developer companies alike.
Getting back to the question, we can only find out the number of wallets by adding the up the individual number of wallets per related company. These figures are neither transparent nor readily available. Moreover, we cannot ascertain if each wallet corresponds to a single unique user.
The silver lining is such that we do not need this data point anyhow.
We started from end and traced our way back to the user. We noted that the wallet data is not available and also not sufficient to figure out the number of users in general. We observed that we cannot strictly determine the number of users from the address data as well. Our best proxy was the 31 million Bitcoin addresses with non-zero balance. Now let’s get in to further detail.
How much does the positive balance addresses help us to explain the current situation?
Consider for example that you used all your Bitcoin to purchase books and your addresses all hold zero balance. You have been active recently but now you are not a user from this perspective. I will take this opportunity to call out to Glassnode — it is time for a new metric: “Number of unique addresses with positive balance in the last 12 months”. Consequently, active users in the prior time period that are not counted can be included in the user calculations as well.
Going back to the main theme, users who have been active before and can be re-initiated easily should be added and multiple accounts of the same user should be subtracted from the grand total of active Bitcoin users for the given period. I have no practical method in my sleeve but if anyone has a suggestion about the mechanics of this can reply to this article.
For the readers who have made it so far, some might be saying “what is this address and my Bitcoins are in the exchange”. At this points, let us revisit Glassnode data. According to the graph below, Glassnode has calculated the Bitcoin in the custody of the following exchanges: Coinbase, Huobi, Binance, Bitmex, Okex, Bitstamp, Bitfinex, Kraken, Bittrex, Gemini, Hitbtc, Poloniex, Coincheck and Luno. This figure is approximately 2,590,000 Bitcoins.
Another question: How many users might be holding the exchange deposited Bitcoins?
We learn that Coinbase has 35 million users according to their own disclosure. If 20% of these are active, that makes 7 million. If Bitcoin users are slightly below the pro-rata to the Bitcoin Dominance, then Coinbase might have 3.5 million Bitcoin users. If numerically insignificant institutional users hold 10% of the Bitcoins in Coinbase’s 936 thousand level custody, then we can make the following calculation: (936000 x 0.9) / 3500000 = 0.24. Thus each individual customer (user) will be holding around 0.24 Bitcoin. Even though this appears a bit too rich, current information points out to Coinbase having 3.5 million or more users.
Assuming this number of users is correct for Coinbase, can we make an inference about the number of users on other exchanges?
Not easily because trading behavior is not consistent with custody preference. For example, Binance has more than 5 times trading volume in Bitcoins than Coinbase which has more than 3 times Bitcoin in custody than Binance. With parallel extrapolation, Binance should have 15x users of Bitcoin than Coinbase if their trading behavior is similar. However this seems unlikely to me. A more plausible explanation is that different user cohorts have different trading behavior such that the data from one exchange cannot implemented on another without substantial analysis.
Yet again if Coinbase has around 3 million users, then other exchanges could have at least 10x or more Bitcoin users. Simply put, there could be more exchange level Bitcoin users than on-chain users.
To make matter even more complicated, we should take in to consideration that (i) a user can both on-chain and exchange level accounts, (ii) institutional funds could be holding Bitcoin for groups of users without creating direct individual Blockchain presences and (iii) some of the fund level users might also have on-chain or exchange level presence as well.
Finally, there probably is a huge chunk of zero balance addresses (or accounts) both on-chain and at the exchanges. While these are not counted towards the current active user count, these users with previous experience represent a different ready to join back in category altogether.
In my opinion, current user number is around 50% of the on-chain user with a 30 million addition from the exchanges thus making 45 million in total. As per world population, this is 0.7% with a potential to reach close to if inactive users with previous balances join back in. This figure is compatible with Paribu’s Turkey-wide Bitcoin survey conducted by Akademetre which identified 0.7% of the population to have invested or traded in crypto currencies.
I tried to examine the Address, wallet and user concepts with some explanation, some conjecturing and a lot of data. There will be technical shortcomings in the explanations but I hope that either the data or the information will contribute to your understanding of the Bitcoin ecosystem.
My personal summary is that Bitcoin blockchain does not have users or wallets, only addresses. These addresses are controlled by their related keys. The blockchain broadcasts all of its data and shares it with all interested parties. Address data is transparent and publicly open. The blockchain is impartial and code is the law.
Wallet and user data are useful for commercial or analytical purposes. Analytical usage is only possible with a high degree of uncertainty in relation the insufficient raw data available as exchange level information is their proprietary data.
What this analysis drives home for me is that Bitcoin has managed to get over the 2017 rise and 2018 fall cycle. Since then new service providers have flourished in the ecosystem, quality of data has improved and technological innovation has taken place. Users have come back, exchanges have made progress and institutional users have joined albeit in small numbers. A volatile but interesting future beckons. This journey requires solid information and robust analysis. I hope this article has helped to assess the related concepts and data a bit more extensively.
This article has been prepared to be published originally in Turkish at BTCHaber.com and also on Medium in English. Turkish version of this article has been published first in BTCHaber.com and then on Medium.
Also Read: The Best Crypto Trading Bots
Also Read: The Best Ethereum Hardrawe wallets
Also Read: The Best Crypto Tax Software
Also Read: Best Crypto Trading Platforms
Bitcoin: Getting Lost in Users’ Wallets, Wallets’ Addresses and Addresses’ Balances was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.