Analyzing suitability of cryptocurrencies for real life adoption in Europe in 2022
April 28th, 2022

TL;DR Cashless payment systems like cryptocurrencies are on the rise, therefore the payment networks have to handle a certain load corresponding of how many people are using it, studied here in transactions per second (tx/s). For cryptocurrencies: — an adoption share of 0.2 % of all cashless payments in Europe, would result in an average x̅ of 23 tx/s and a peak of 33 tx/s — an adoption share of 2.0 % of all cashless payments in Europe, would result in an average x̅ of 227 tx/s and a peak of 332 tx/s — an adoption share of 1.0 % of all cashless payments in Germany would presumably lead to 3.3 million people using cryptocurrencies for 1.4 transactions per week per person at a rate of 0.8 transactions per day per person — A 5.0 % transaction share in Germany with people using for all cashless payments 25 % cryptocurrencies at a rate of 0.8 transactions per day per person would result in an average of x̅ 69 tx/s and a peak of 110 tx/s — adoption will take place first with online merchants, already possible to day at low adoption shares for some unique cryptocurrencies + incentives for merchants are already higher than drawbacks

You may hear people talking about scalability and transactions per second (tx/s)? Ever wondered why this matters for a payment-system, especially for cryptocurrencies and why this parameter is often (lightly) used as a metric of a network-performance overall? Let me introduce you about some mechanics behind and why a high transaction throughput (among other parameter) is crucial for the adoption of a currency-coin (often called peer2peer-coin or p2p in crypto-world).

Elon Musk recently twittered and talked about in Friedman’s podcast, that digital money needs to have a high scalability and low latency to serve as a medium of exchange. But what exactly does it mean for scalability talking in numbers? E.g. Bitcoin’s network has a transaction throughput of max. 7 tx/s and is since inception the most prominent currency-coin. Followed by Ethereum’s ledger with a bit higher number of roughly 17 tx/s. Up to today both projects experienced partially severe clogging issues (network reached maximum tx/s capacity) resulting in people all over the world paying absurdly high fees to boost their transactions to get processed as fast as possible. Conclusion: A (globally) distributed ledger technology (DLT — cryptocurrencies are part of) needs higher numbers in transaction throughput as exmples above, to fulfill the purpose of being a medium of exchange. Looking at traditional finance e.g. VISA’s transaction network handles roughly 1500–2000 tx/s *1 and is outstanding in contrast to well known cryptocurrencies. The fact that VISA is able to perform like this is given by its highly centralized nature, whereas cryptocurrencies in general (theoretically!) strives for a high degree of decentralization (measured in Nakamoto coefficient).

In 2019 VISA and Mastercard combined processed nearly 97 billion transactions in Europe 2. By comparison in 2017 a total of 134 billion transactions were done cashless in Europe 3, not including VISA and Mastercard. (Europe’s population in 2017: 513,000,000 inhabitants, resulting in 0.72 tx per person per day while assuming an uniform distribution of all cashless spendings over the week) Makes a total of 231 billion cashless transactions alone with VISA, Mastercard, EC, transferals, debit entries and checks, when adding numbers from 2017 and 2019. Seeing recent trends the share of electronic payments is on the rise, replacing payments in cash even more.

But what does it means in terms of numbers for needed tx/s? This study will research a region and will show on the basis of actual data, given distributions and some (conservatives) assumptions, how the needed throughput for a payment provider roughly behaves dependent on market adoption rate. Note: a certain region is a start of understanding, globally seen things getting way more complex.

Euro-zone (EU + nearby countries) serves as perfect example for this study. Dense population, many nations, home of a lot of different industries and with this a lot of people with a high purchasing power. Further Europe is divided by four time zones (UTC 0 up to UTC +3) which adds some realistic behavior on fluctuating network load — while people in Ukraine (UTC +3) are mostly done with daily spendings at 19:00, people in UK (UTC 0) doing their after work groceries at 16:00. Resulting in different countries giving different load to the payments network at the same time while living in different time zones. Picture below *4 depicts the time zones and countries in it, used for this study.

This study includes 44 countries (UK, welcome back to EU!). Following table *5 lists the countries and their populations (given in thousands) while depicted in corresponding color to the time zone they are in:

Time zone UTC +1 is over represented in populations count, roughly by a factor of four in contrast to the other time zones. Resulting in most load is given by the countries within this time zone and determine the shape of the load distribution the most.

In total this study combines 678,877,000 people. People who all need to buy daily goods and other stuff. The derived distribution for the transactions is obtained by the distribution of payments in Germany in assumption, that other countries more or less share this behavior, depicted in following chart *6. (Please note: The sum is slightly >100%, maybe the statistic from Statista includes rounding errors at given time frame. This fact resulting in an error in calculations of roughly 1.1 % on the conservative side)

The main times transactions are done is between 8:00 and 20:00.

As above mentioned 134 billion transactions (not including VISA and Mastercard) are made within the EU in 2017 *3 by 513,000,000 people resulting in 0.72 transactions per person per day. Since recent data shows a rising trend to cashless payments and also the Covid-19 pandemic encouraging people using cashless systems, the parameter of transactions per person per day is assumed with 0.80 tx for 2022. Transferred to 678,877,000 people living in the Euro-zone this results in a total of 198.2 billion transactions per year given as network load (200.4 billion transactions are used in calculations due to a small error in distributions of payments by Statista). (Please note: 0.80 tx per person per day is a synthetic value, breaking down the numbers of all transaction done. Its of course not representing a distribution of actual payments, including babies, childs and very old people which presumably don’t make any transactions at all.)

Assumption: this 200.4 billion transactions are equally distributed (Mo-Sun) assuming that people will order food on Sat/Sun and do a lot of online shopping as well. Peak times like Christmas-shopping is also not considered — let‘s stay simple!

The parameter transaction-share is used as a metric for market adoption of a payment network. The resulting transactions per second (tx/s) are derived from numbers above and are a function of the transaction-share. The calculated steps within this study for the transaction-share starts at 0.2 % up to 1.0 % in steps of 0.2 % and from 1.0 % to 2.0 % in steps of 0.5 %.

Results:

Seven cases are calculated and depicted below. The parameter transaction-share varies from 0.2 % up to 2.0 % and describes the share of a payment-system in Europe’s total cashless transactions. The parameter of the number of transactions per person and per day is constant at 0.8 tx/d. In total 200.4 billion transactions per year by 678,877,000 people in Euro-zone, living in four time zones, were distributed according to given distribution of daily payments in Germany.

The time is plotted on the x-axis in 30 minutes intervals. The blue columns are the transactions-total and plotted on the left y-axis. The dashed red line is the corresponding number of transactions per second (tx/s) needed to achieve the throughput and are plotted on the right y-axis.

The following table depicts for each transaction-share (tx-share) the corresponding transactions per second (tx/s).

Average total describes the tx/s from 0:00 to 23:00. This period contradicts the distribution within a day and is hereby not suited to derive a realistic number of a load given to a payment network suitable for p2p-payments.Average x̅ 08:00–20:00 describes the main business times. While making the assumption of an uniform distribution within this period on given daily distribution, this number describes the average load a payment-network has to handle within this time period.Peak describes the maximum number occurring while deploying the given distribution for transactions within a day. This can be seen as a maximum tx/s a payment-network has to handle.

The following chart depicts all seven cases combined as well shows the average load a payment network has to handle, while assuming a uniform distribution within this time frame.

The time is plotted on the y-axis in hourly intervals. The corresponding transactions per second (tx/s) needed for the networks-throughput is plotted on the x-axis. For each case a vertical bar is plotted and indicates the average of tx/s needed for business hours from 08:00 to 20:00 as well as the occurring peaks.

A transaction-share of 0.2 % with a 0.8 transaction per day and per person needs an average x̅ throughput of 23 tx/s whereas a transaction-share of 2.0 % with a 0.8 transaction per day and per person needs an average x̅ throughput of 227 tx/s in average.

During peak times the network load will exceed the average x̅ of roughly 31 % and thus resulting in an even higher throughput needed, a network needs to handle for at least some hours.

Conclusions:

A) Load on a network, represented as transactions per second (tx/s)A transaction-share of 0.2 % resulting in an average of x̅ 23 tx/s and a peak of 33 tx/s. This is the smallest transaction-share case studied. Setting numbers into perspective: A transaction-share of 0.2 % means 2 people out of 1000 using this payment network for 100 % for all cashless transactions. A tiny number of people are already exceeding the capacities of Bitcoin’s network with a factor of roughly 3.3 times and Ethereum’s network with a factor of roughly 1.4 times when running outside of peak-times. Sadly this is true for most of the DLTs, especially for the ones building on a so-called „Blockchain“ (or better to say in a traditional sense like Satoshi Nakamoto did with Bitcoin). Nowhere near of gaining a minimal share in todays payments systems.

Looking at a transaction-share of 2.0 %, resulting by the linear nature of these calculations, in an average x̅ of 227 tx/s and a peak of 332 tx/s. These numbers are in perspective also quite low, but already mean 2 people out of 100 people using this payment network for 100 % for all cashless transactions. While assuming that no one uses one cashless payment method 100 % for all transactions and instead using cryptocurrencies for 25 % of all transactions, this would already result in 8 people out of 100 people using cryptocurrencies frequently for cashless payments (to be precise: for each person this would be 1.4 transactions per week).

B) Usage of cryptocurrencies within populationAssuming no one would use one single cashless payment-system 100 % for all transactions would result in following scenarios. The x-axis shows the transaction-share, the y-axis shows the people per country, using crypto as payment method for 25 % off all cashless transactions.

In this conservative plotted scenarios its remarkably, that already a transaction-share of 1.0 % would result, e.g. for Germany, in more than 3.3 million people using cryptocurrencies frequently as p2p-payment method.

Remaining in Germany, a transaction-share of 5.0 % results in an average x̅ of 69 tx/s and a peak of 110 tx/s, as depicted in the chart below.

C) Thoughts about merchant adoption (Note: For following conclusions, its assumed that merchants only utilize cryptocurrency-networks which are fee-less and transactions are near instant processed, matching a merchant‘s need.)

Bearing the derived numbers in mind, the question arises, if merchants are interested into offering payments with cryptocurrencies? This is not easy to answer and presumably highly dependent on the form of a company. Offering cryptocurrencies as payment method could potentially result in positive and negative consequences like: I) resulting in some up to huge savings in fees and settlement costs (compared to traditional finance) II) resulting in direct receiving of payments (no middleman involved who withhold payments to check if e.g. valid or not. Worst case scenario the middleman don’t cashs out the merchant because a customer raised issues — valid or not) III) resulting in a hassle integrating the payment-system in existing IT-systems or set-up/maintain an own node E.g. a bakery nearby wouldn’t see as much of an incentive as an online-merchant. The bakery maybe needs an additional device/terminal and needs to train their staff for only some people using it with potentially lower amounts to pay and would potentially needs a middleman, frequently exchanging the received cryptocurrencies back into fiat-currencies. Given the number of customers per day, the hurdles are maybe to big to overcome for potential benefits some people would bring paying with cryptocurrencies (at current very low transaction-share) who could easily pay with traditional currencies. Or in other words the incentives in fee-savings aren’t that big for low transaction-shares for local retailer like bakeries, kiosk, butcheries and so on, offering goods for a low number of casual customers per day. Whereas an online-merchant can easily automates payments via QR-codes for potentially higher amounts of money paid by their customers, possibly having more customers per day than a local retailer. Also offering goods online is anyhow a technical way and lowers the hurdle for integrating payments in cryptocurrencies in existing IT-systems. At the same time presumably people buying goods online are more used to utilize cashless payments anyhow and not using cash at all. As a big benefit the potentially saved fees over using a traditional payment-system like VISA, could be (partially) refunded to the customers, adding an incentive for the customer to use cryptocurrencies and especially online merchants offering this payment method. Further accepting cryptocurrencies means once received no chargeback can occurre. Thus, no hassle with unpaid goods already left the store and additional chargeback-fees.

In my opinion adoption will take place first within online-merchants. The incentives are already given at lower transactions-shares cryptocurrencies already reached today. Depending on the cryptocurrency (network/p2p-coin) the incentive is using the network itself — some (very unique) cryptocurrencies offers no fees at all (ever!) and very low settlement times of less than a second. No fees rapidly adds up to high savings over time offering new possibilites traditional finance don’t provide for merchants!

In an upcoming study I will investigate how these potential savings will impact merchants and I will set some numbers into perspective. A further interesting study would be a comparison of cryptocurrency-networks in regards of which network can handle the load up to which adoption share.

I hope you liked this study. Please feel free to write comments/questions in comment section and don’t hesitate to contact me at @FabsinIOTA (Twitter). Thank you for reading.

Special thanks to @mira_hurley for being an inspiration and providing some valuable inputs.

Sources:

Subscribe to dreadfulLizard4
Receive the latest updates directly to your inbox.
Verification
This entry has been permanently stored onchain and signed by its creator.
More from dreadfulLizard4

Skeleton

Skeleton

Skeleton