Ethereum, the first blockchain with smart contract functionality, launched in 2015.

Since then, several other EVM chains have launched like Fantom, BNB & Avalanche, along with scaling networks like Polygon, Optimism and Arbitrum. 

There’s no doubt that by early 2023, the multichain world has truly arrived. With a whole universe of L1s, bridges, L2s, and cross-chain Dapps - users now have more choices than ever before!

So, which choices are they making? Are users embracing the possibilities, or largely sticking to their favorite chains? 

In this article we’re going to examine these questions by analyzing the cross-chain journeys and behaviours of EVM Externally Owned Addresses (EOAs) aka user wallets.

We’ll go deep into behavior across chains and user choices.

By the end, you’ll understand the overall preferences and behavioral trends of millions of crypto users. Let’s get started! 

Most Popular Chains

Firstly, which chains have the most users in general? 

The chart below compares daily active users across EVM chains, and shows us some interesting trends unfolding over the past 9 months…..

Query

We can see that BNB chain has consistently seen the most daily users, with Ethereum and Polycon continually vying for second place. 

Meanwhile Optimism and Arbitrum have gradually gained users as L2s found their market fit throughout 2022, both hitting >50k DAU by January. 

In fact, Optimism and Arbitrum peaked at  ~110k and ~610k users respectively in recent months. Impressive!

So that’s the high level. Now we’re going to take a look at the following questions:

• What insights can we gain when we start analyzing EOAs that interact across chains?
• What proportion of total EOA’s go multichain? 
• Do clusters of certain EOAs tend to interact on certain chains? 
• Are certain chains more popular with EOAs that use multiple chains or are they more isolated (meaning the addresses remain just on a single chain)?
• Do we see similar or different user behavior across chains? 

First of all, we uncovered an interesting fact......

Only ~6% of all EVM addresses have gone multichain

We started out by looking at addresses that overlap between chains. The criteria: they initiated at least one transaction on at least two chains since 30th July 2015, the date of the first transaction on Ethereum. 

The results were interesting. 

In absolute terms, a large number of addresses went multichain. 11.6m have used two chains , over 2.3m have used three, and over 1.3m have used four or more.

At the most extreme end, a group of over 63 thousand multichain degens have used all seven EVM chains. 

Query

These figures may seem impressive. But relative to the total users of these chains, the multichainers are a small minority. According to our analysis, a large majority of EOAs (~94%) have only used one chain.

Less than 6% have made the jump:

Query

For our purposes, we aren’t particularly interested in the “one chain” cohort. We’re going to be focusing on those who have traversed multiple chains.  

We’re going to go deep into the behavior and preferences of these addresses, let’s dive in. 

Ethereum and BNB have the most active addresses, but the lowest proportion of cross-chain addresses 

Remember - less than 6% of all addresses have gone multichain.

It turns out that’s largely thanks to the two biggest players, Ethereum and BNB. They have by far the most active users, but also the smallest percentage of multichain addresses. 

Though BNB has had the most active users recently, it has only been around for a couple of years, seeing ~84.4 million addresses throughout its history. Ethereum launched way back in 2015, and has had the most total addresses - over 158.2 million.  

Out of all these, 7.21% on Ethereum and 13.83% on BNB have gone multichain.

Next up is Polygon at 23.9%, and Fantom at 33.4%.

With Avalanche, we see the first case where the majority of users are multichain, at 59.61%. Finally we have Optimism and Arbitrum, both with 70%+ multichain. 

Query 

An interesting metric in the above chart is “recency”, which measures the median age of each EOA’s last transaction.

Ethereum, BNB & Polygon have the greatest portion of EOAs that are inactive. 

In fact, we see a strong correlation between the % of multichainers and “recency” of a chain.

Optimism & Arbitrum have the highest % of multichain users and by far the most recency, which makes sense when you consider their relatively recent launches. 

Avalanche alone bucks this trend, having both high recency and ~59% of users cross-chain. 

Actually, the above data goes all the way back to genesis and skews things a little. Let’s drill down to the past 6 months to see what contemporary users are doing.

The chart above shows us that user behavior has been quite different in recent history. 

We can see that BNB, Polygon, Fantom, Optimism & Arbitrum have seen a drop in the percentage of multi-chain users. On the other hand, Ethereum & Avalanche have seen an increase.

BNB has also been the chain with the smallest multichain user group (10.44%), overtaken by Ethereum at 14.25%!

Now let’s dive deeper into the specific journeys that the multichain EOAs undertake……

Top Combinations

So which are the most common routes from chain to chain? 

Query

The above chart shows combinations of chains used by ‘cross-chain’ travellers (sorted by Active Addresses). It's clear that the most popular combinations are:

•  BNB<> Ethereum
•  BNB <> Polygon
•  Ethereum <> Polygon

Another interesting observation is that there is a group of chains ‘Avalanche , BNB, Ethereum, Polygon’ which is a 4 chain combination within the top 20 combinations. 

This alone doesn't tell us much, so lets look more closely at the specific routes and journeys taken by users.

Popular Journeys 

So we've seen the most popular cross-chain combos, but how about the specific journeys taken by EOAs?

In this chart below we analyze this, taking into account the specific routes taken:

Query

It’s clear that the most popular ‘first chains’ (defined as the chain where an EOA does their first ever tx across all EVM chains) in terms of absolute numbers are: 

• Ethereum ‍
• BNB
• Polygon

Hardly surprising, since those are the most popular chains in general. So after these starting points, the five most popular journeys are:

• Ethereum->BNB 
• BNB->Ethereum
• BNB->Polygon
• Ethereum->Polygon
• Polygon->BNB

When measuring the popularity of these routes in terms of percentages, we see relatively low % of users from each “original chain" actually venturing away to a new one.

Only 2% of Ethereum addresses head to BNB, while 3.4% of BNB addresses end up on Ethereum. Ethereum is almost twice as popular with BNB users as Polygon, with just 1.85% of them venturing in that direction. 

Another way to look at it beyond raw absolute numbers: which are the most popular journeys, as a % of specific chains' overall active users?

The top 3 are:

• Arbitrum > Ethereum
• Optimism > Arbitrum
• Ethereum > BNB

These stats give us a good overview of which chains are popular starting points and which chains are popular among cross-chain users. What if we want a more concise way to understand overall EOA cross-chain behavior though? 

Let's break out the trusty pairwise matrix to understand this better.

Common overlaps

Generally, BNB Chain is a common part of a user's cross-chain journey, while Arbitrum and Optimism share large overlaps with each other. 

Query

^{EOAs who used more than 2 chains will be "double counted" to all chains on their cross-chain journeys, ie a user who uses Ethereum -> BNB-> Polygon will be ‘counted’ in the Ethereum<> BNB , BNB <> Ethereum , Ethereum <> Polygon ,Polygon <> Ethereum, BNB<> Polygon, Polygon <> BNB} 

Let’s take a step back and think about all user journeys.  

The 2D pairwise matrix above shows the exact number of EOAs that have initiated at least 1 tx and paid gas fees on any given 2 chains.

It gives us an intuitive view of the overlap between pairs of chains from a different perspective. For example, we can quickly see that there are at least 8.68 million addresses who have at least used BNB Chain and Ethereum.

Since Ethereum's launch in 2015, we can see that the chains with most overlapping users are:

• Ethereum <> BNB (~8.5m)
• Polygon <> BNB (~4.7m)
• Ethereum <> Polygon (~3.5m)

So that's the absolute figures, but what about in relative terms?  

If we want to know how this overlap translates into how ‘popular’ a chain is another, taking the % of each chain’s EOAs that have interacted with a given second chain is an interesting question:

Query

^{The  % overlap between any 2 chains is calculated by dividing the no. of active addresses per pairwise overlap by each chain’s total EOA count (the denominator is the chain in the 1st column, numerator is the overlap)}

There are a few interesting things to note here.

Predictably, Arbitrum & Optimism have very high overlaps. 47.16% of all EOAs on Optimism have used Arbitrum, and 33.3% of Arbitrum's have used Optimism.

For Ethereum, Polygon, Fantom, Avalanche (1st column), BNB Chain is the most popular second chain with their multichain user base. When it comes to Aribtrum and Optimism though, BNB is less popular.

The EOAs on these L2 rollups are more likely to overlap with Polygon, Ethereum, or each other. BNB Chain is not the top ranked chain in terms of % conversion rate, but instead is 3rd ranked.

These kinds of datapoints could guide competitive strategy. For example, BNB could seek to develop more DAPPs and features that typically appeal to Optimism and Arbitrum users.

The Multichain World is still niche

Let's summarize what we have learned:

1. For the 7 EVM chains in this study, a large bulk of EOA’s stay on one chain (~94%), but the remaining 6% have gone crosschain
2. Ethereum and BNB have the largest number of EOA’s in total (‘base’ of EOA’s), but they have the smallest proportion of users going cross-chain at 7.21% and 13.83% respectively.
3. Most popular 2-chain journeys in absolute terms are ETH -> BNB, BNB -> ETH, BNB ->Polygon‍
4. As a % of each 1st chain’s total user count though, the ‘conversion rate’ for each of these specific journeys is relatively low at ~2-4%  
5. The journeys with the highest "conversion rates" are between the newer L2 rollups; Optimism->Arbitrum, Arbiturm -> Optimism‍
6. For any 2 chain combination, the 2d pairwise matrix shows that BNB Chain has the highest overlap % with the following chains: Ethereum, Polygon, Fantom and Avalanche, However, BNB is not the top ranked chain by overlap % with the Arbitrum and Optimism

It's clear that despite the multichain DeFi and NFT buffet on offer, a large proportion of EOAs remain on one chain.

This could be because they do indeed stick with the one chain, or that actual users tend to create multiple wallets for different chains. One of the struggles of on-chain analysis is that we can typically only track the wallet itself, not the human behind it.

However, with improvements in wallet UX and "smart accounts", we should start to see an increase the portion of wallets that go cross-chain as the need for multiple wallets diminishes.

By analyzing these metrics ecosystems, protocols and organizations can better evaluate opportunities and threats.

Ecosystems can make educated guesses on the segments that may be lacking, and DAPPs can plan their multichain expansion strategically. For example, a lending DAPP on Ethereum considering a multichain launch can evaluate the total addressable market on other chains - and which ones they should prioritise. A low overlap could be an argument in favour of expansion.

The next part of this series will fill in the missing pieces. We're going to analyze cross-chain EOA behaviors in terms of which DAPPs they interact with on different chains, and also user behavior while bridging!

Stay tuned.

David C is a Senior Data Analyst at BNB, give him a follow on Twitter