Crypto Analytics: Exploring Uniswap v3 Data
This was originally posted on: https://www.msilb7.com/posts/crypto-analytics-exploring-uniswap-v3-data
A few weeks ago, I published my first crypto-related blog post: Bring Your Own Wallet: The Future of Consumer Choice and Personalization on the Internet
Since then, I’ve continued to fall down the rabbit hole. So much so, that this post is probably 100 steps more advanced than my first post. So apologies for that, but the place I’ve gotten to now is too exciting to not share.
[Scroll to ‘Analyzing Uniswap v3 Data’ to skip to the analysis]
The Plunge Into Work
As I progressed down the rabbit hole, I joined crypto project’s Discord channels, read through forums and governance posts, and (my favorite) checked out their analytics dashboards. I have some experience from my day job building dashboards, and a ton of strong opinions about dashboard design, so I figured that this could be a spot for me to try. 99.9% of the dashboards I saw were from Dune Analytics, so I signed-up on the day of an ETH all-time high (which I thought was good timing), and off I went.
A Dashboard is Born — Uniswap v3
A very fortunate coincidence was that this happened to also be a few days before Uniswap’s highly anticipated v3 launch (read more here). Within a few hours (minutes?), the Dune team had parsed the Uniswap v3 contracts and made the data available for people like me (SQL is more my speed).
There were a few burning questions I had going into the launch: How does this impact how much Uniswap is used? Does v3 impact what pairs Uniswap is used for? Where is and isn’t v3 being adopted?
A few hours later, I started getting some answers:
Now, enough of my personal journey. Time to answer some of the questions that you all are here for.
Analyzing Uniswap v3 Data
[Data was analyzed on May 8 2021]
Dashboard Link to follow along: https://duneanalytics.com/msilb7/Uniswap-V2-vs-V3-Ramp-Up
How much is Uniswap v3 being used?
To answer this question, I looked at the total trading volume split between v2 and v3, both by total number of trades and by total USD trades.
For most questions, I’d want to look at USD volume, but since my question is about usage and USD value is kind of volatile, let’s focus on number of trades:
As we can see, somewhat predictably, the volume spikes up as Uniswap v3 launches on May 5th, then steadily progresses towards a greater and greater share. Near the end of May 5th, v3 topped at 10% of all trades, but at the time of writing (May 8th), this number has already reached 25%.
So yeah, people are using this thing. For reference, total USD through v3 topped so far at 37% on May 7th.
Does v3 Impact What Pairs Uniswap is Used For?
This was probably my most burning question coming in to this. One of the major innovations of Uniswap v3 is that liquidity providers could specify the range at which they’d like to support, rather than the whole range of 0 to infinity. If done right, this would help users get prices on Uniswap, especially for like-for-like trades, where the exchange rate is consistent and predictable (e.g., DAI-USDC, ETH-stETH). Like-for-like trades have historically been better served by protocols specifically designed for these pairs, such as Curve.
Uniswap’s app UI defaults users to v3, but displays a button to switch to v2 if users can get a better price there. So, we should expect to see trading volume naturally flow to the version of Uniswap where users get best price for each trade. If users can get better prices for like-for-like pairs, then we should also see more demand for these trades come from competing products to Uniswap.
If that made no sense, then check out ‘How Uniswap Works’ as a starting point.
So, do we see a difference in the most traded pairs for v3?
Yes! The top 10 most traded pairs by USD Volume for Uniswap v2 (since v3 launch) were all ETH pairs (e.g., ETH/USDT, ETH/WBTC), whereas in v3, 3 stablecoin pairs reached the top 10 by volume (USDC/USDT #4, DAI/USDC #6, DAI/USDT #9).
v3 was supposed to be more favorable for these pairs, and so far it’s succeeding at doing that.
Wrapped Bitcoin Surprise: I was surprised by how much WBTC/stable volume has already gone over to v3 (WBTC = Wrapped Bitcoin). 99.95% of WBTC/USDT USD volume ($7.5M, 83.4% of trades) has already gone to v3 since launch, and 88.5% of WBTC/USDC USD volume ($7.6M, 51.8% of trades). Is this a sign of demand to capture WBTC volume, the relative stability of bitcoin versus other altcoins, or maybe a combination of both?
ETH Pairs Stuck on v2 — Value Opportunity?: The ETH/stable pairs see the most trading volume on Uniswap, but as opposed to WBTC, they’re still mostly traded on v2 (ETH/USDT 27% v3 USD volume, ETH/USDC 36%). These pairs have both the most total USD traded on v3 and the most USD still traded on v2. It’s not a surprise to say that ETH pairs are where the money is, but it may be surprising to know how much volume still remains to be captured by v3 liquidity providers (brb, doing this).
“v3 Dominance”: The total share of a pair’s Uniswap trading volume that has gone through v3
Use the dashboard to hover over the scatter plots and view each pair. WETH is the wrapped version of ETH (essentially the same thing for purposes of analysis).
Where is and isn’t v3 being adopted?
We already explored where v3 adoption is happening, so for this one, we’ll shift to what value remains to be captured.
I’ve only heard of a few of these pairs, so there could be any number of reasons why their trading is still only happening on v2. The top example, BADGER/WBTC is a pair incentivized by BadgerDAO with rewards for providing liquidity. So, until BadgerDAO flips their liquidity provider incentives to v3, we may expect this to stay put.
I’m not aware of any incentives for the other pairs, but since it costs users extra gas to open a brand new Uni v3 vault, the breakeven point to try to arbitrage some of the opportunities is likely pretty pretty high.
Uniswap v3: A success or a flop?
I jumped into this analysis out of curiosity, but once I started to dive in, the questions started to hit me. There’s only ~10–20% of trades going through v3. I know it’s only been a few days, but is this bad? Or, since it’s only been a few days, is this really good? How should I even think about judging this? I then saw the tweet below by Uniswap founder Hayden Adams, with how he was evaluating it:
I liked this approach. So let’s see what this looks like. I anchored total USD traded by # of days since each version of Uniswap launched and — oh:
Uniswap v3 is doing just fine. 🤯
