AMA Highlights: CowSwap
By Daniel Dal Bello, Director.
December 9, 2021–9 min read.
On Thursday 9 December, we welcomed Anna George of CowSwap into the Hillrise Group Telegram chat for an AMA about increasing efficiency in decentralized exchanges and fighting ‘Maximum Extracted Value’.
CowSwap is an Ethereum-based DEX aggregator that allows trading with MEV protection while using ETH-less orders and ‘Coincidence Of Wants’.
The name is based on a novel system that matches buyers and sellers before routing through an AMM, leading to higher efficiency in trades.
CowSwap is the first trading interface built on top of the Gnosis Protocol v2.
We were curious to learn more about how CowSwap brings innovation to the DEX space.
In this post, we have compiled key questions and answers from the event.
Daniel Dal Bello
Hi Anna and welcome! Great to have you here. We’re excited about this conversation as we see CowSwap bringing some tangible innovation to the DEX space.
Before we get into the details, could you introduce yourself and the CowSwap project?
Anna George
Of course, happy to! And thanks a lot for organizing this opportunity to introduce CowSwap.
Well first off, I can talk a little bit about myself. I did join Gnosis a bit more than 4 years ago and at the beginning was wearing many hats — as I guess we all do when the team is still small.
I focused mainly on partnerships across the different products Gnosis has been working on: the Gnosis Safe, Prediction Markets, Dutch Exchange, Gnosis Protocol v1, etc.
Since the beginning of this year I shifted towards a product-focused role at CowSwap, and now — that the team will spin out of Gnosis — I will be heading the team to this new adventure together with Felix (our teach lead).
Regarding the product, I’m not sure how much people know already, but our DEX CowSwap, which is in a way a ‘Meta-DEX Aggregator’ that protects users against value exploits (such as MEV), has launched in May this year and has already achieved quite significant traction — we currently have a daily trading volume of US$30–50 million.
The unique feature of CowSwap is that it adds a batch layer, where we are able to match trades peer to peer.
Daniel Dal Bello
As the project focuses on what is quite a technical concept in Miner Extractable Value (MEV), could you introduce this concept at a high level?
Why is it important to us as regular end-users and what is your approach to solving this problem?
Anna George
Miner Extractable Value — or how it has now been renamed to Maximal Extractable Value (as the value can not only be taken by Miners but also by other parties), refers to the maximum value that can be extracted from block production in excess of the standard block reward and gas fees by including, excluding, and changing the order of transactions in a block.
In more simple language, whenever you place a trade, there’s likely an opportunity by others to extract value from this trade.
What is the impact? There are trades we have seen where a user lost 7 Ethereum in one single transaction (or sometimes even more).
Essentially the slippage tolerance a user sets to ensure their trade goes through can be exploited. Exploited means that someone else takes this slippage from you.
Maybe not the best explanation but if someone wants to dig deeper, feel free to ask.
Essentially the slippage tolerance a user sets to ensure their trade goes through can be exploited. Exploited means that someone else takes this slippage from you.
This is a cool dashboard by the Flashbots Team showing the value that has been extracted from traders over time.
Rowan Zwiers
In an attempt to counter this MEV manipulation you propose delegating trade executions to ‘solvers’, who execute the trades on behalf of traders.
We came across an authentication process for these solvers in your documentation, and have seen that you incentivize these solvers with a small fee determined by the users.
Can you tell us more about these solvers and how they operate?
Anna George
What happens in CowSwap is that external parties — called the solvers — look at the CowSwap order book and then compete against each other to find the best possible way of executing the existing orders. Then, the solver with the best solution executes the trade. Obviously, they have to spend gas for this, hence the fee that is charged to the user includes the coverage for the execution cost + a small bonus.
What happens in CowSwap is that external parties — called the solvers — look at the CowSwap order book and then compete against each other to find the best possible way of executing the existing orders.
This fee is not determined by the user by the way, but by the UI.
Rowan Zwiers
How is it decided if a solver proposes the best possible trade?
Anna George
Solutions are ranked based on an objective criterion, which is looking at the prices and surplus users receive, minus the execution cost of the trade.
The exact criterion is ‘total_surplus + total_fees — total_costs’.
For example. Sometimes there could be a tiny price improvement by splitting a trade not only via two pools but potentially via three pools. However, the increased gas cost of this transaction has to be accounted for , which might not be worth the cost-benefit of splitting the trade 3 ways.
Rowan Zwiers
How do uniform clearing prices of batched auctions guarantee average lower gas fees for executed trades?
Anna George
Imagine this scenario. You are placing a trade, selling USDC, and buying BAL. You would usually incur gas prices, slippage, and protocol fees (and worst-case some MEV exploitation).
Now, if there’s someone else selling BAL and wanting USDC for this, why would you pay all these fees? You could simply exchange your tokens with this other user. And this is exactly what CowSwap facilitates.
By matching these orders directly peer to peer, the user doesn’t have to pay the protocol fee (e.g. on Uniswap it’s 0.3%), also the gas cost is lower than matching the trades directly, and the user might save on some slippage.
Now, if there’s someone else selling BAL and wanting USDC for this, why would you pay all these fees? You could simply exchange your tokens with this other user. And this is exactly what CowSwap facilitates.
Raymond Reijnders
Coincidence Of Wants enabling executions outside of liquidity pools is basically a form of ‘OTC settlement’, which obviously increases efficiency and lowers routing costs/fees.
However, for these coincidences to occur frequently we can imagine a high volume of off-chain ‘wants’ need to be at the disposal of the settlers (solvers) to match.
Is this essentially a conversation around the network effect? How important do you feel the influence of higher user numbers will be in capturing such coincidences and increasing trade efficiency?
Anna George
Definitely, the benefits for all users increase the more trades we have in the system.
We have done some research into this, basically projecting what price benefits we can realistically offer users, and we have seen that if we take randomly 20% of DEX trades taking place on Ethereum today, and put them into CowSwap, then we can see price improvements of 0.5%.
Of course, this also depends on which tokens are traded, with more frequently traded tokens having a higher chance of finding COWs.
We are also playing around with different concepts to increase COWs. For example, we could offer users to pay lower fees, if they are willing to wait until the next COW settlement. With Cow-Settlements set fix to every quarter of the hour. So users could decide to check on ‘wait until next COW Batch’ (which would be either 15, 30, 45, or full hour timelines — whichever one is closest).
This would significantly increase the number of COWs already today. But even with the current trading volume, we frequently have batches with at least 2 trades.
Rowan Zwiers
Do you also foresee manners to cooperate/aggregate these wants with other aggregators/solutions providers?
Anna George
100%. So essentially, we collect all trades in one single order book, regardless of whether they were placed via the CowSwap interface or via other integrations — for example, the Balancer integration.
The goal is to have as many DeFi platforms, aggregators, but even some native AMMs and DEXs (as Balancer) integrate the COW Protocol!
Raymond Reijnders
I’m curious to hear, to what extent will the implementation of ETH 2.0 in the coming year have an impact on CowSwap?
Anna George
Not so much to be honest. Our focus will remain on Ethereum for the foreseeable future, but we are also already looking at sidechains and layer 2s.
Rowan Zwiers
Will the PoS settlement on Ethereum 2.0 not change the technicalities exploited in MEV activity?
Anna George
Yes, the Flashbots team is looking into auction mechanisms. However, the extractable value of course still exists on Ethereum 2.0, and the way CowSwap works as a batch layer on top remains unchanged.
The idea of CowSwap is to already minimize extractable value by exchanging as much volume as possible peer to peer.
So the impact is only on the remaining volume that still has to be settled via on-chain liquidity.
Solvers are currently integrated with the Flashbots RPC to handle this part for additional protection, and once Eth2 is live, they will update to the new Flashbots RPC built for Eth2 (and/or alternatives by other teams).
Daniel Dal Bello
The DEX space is getting fairly saturated and standing out in the crowd of DEX aggregators is getting more and more challenging — even for those that provide the best solutions.
What is your approach to marketing and attracting users to your platform? How do you emphasize and communicate your USPs to the user?
Anna George
We definitely still have some room for improvement here, especially around communicating our value proposition more clearly, but we also have a lot of cool ideas in our backlog.
One thing that recently launched is our affiliate program allows users to share a personal link to CowSwap with their network and we keep track of how much volume is generated via this link. Currently, no profit expectation is communicated yet and we can already see quite some traction generated there, but it might increase more once the reward structure is announced.
There are many more ideas piling up but I don’t want to spoil the surprise!
Rowan Zwiers
For a protocol that brings meaningful technical innovation, you have chosen a very playful style in visualization!
Naturally, we can assume ‘Coincidence Of Wants’ leading to the acronym of ‘COW’ has served as an inspiration for this direction.
To us, it is fair to say that on the surface, a new user might think nothing about the name is significant — like Sushi or Pancake.
How important is the branding piece from your perspective?
Anna George
I believe branding is super relevant.
We were approaching it from this end. The Gnosis brand has huge value in the sense that it is being trusted and the development team is recognized. What we were lacking, however, was building a community that gets excited about our product.
We compared ourselves to competitors in the space and noticed that we never tried out the ‘meme-approach’, we were giving it a chance to see if it would make a difference — and it did! Since we launched CowSwap, people have a brand to identify with, they like to hang out in our Discord and they are having fun with the brand.
The Gnosis brand has huge value in the sense that it is being trusted and the development team is recognized. What we were lacking, however, was building a community that gets excited about our product. We compared ourselves to competitors in the space and noticed that we never tried out the ‘meme-approach’, we were giving it a chance to see if it would make a difference — and it did!
Of course, there will always be different users with different needs, there are some who would prefer a more ‘serious’ interface, but even those got used to the brand and like it now. You can never cater to everyone, but we felt the need to change something to build a community around us, and this really worked. In the future, there will likely be different UIs though to cater to different needs.
Part of this spin-off is to also establish a more serious brand. CowSwap, the meme interface, is just the interface.
The Protocol, that our team is focused on building, is the Coincidence of Wants protocol. This one will focus on what it stands for and have serious branding linked to it.
Rowan Zwiers
We’re getting to the end of the year now, and it has been a wild one at that. What are your main focus areas leading into 2022?
Anna George
Well, the biggest next thing is obviously the GIP-13, which will decide whether CowSwap will have its own token.
Other than that, things on the roadmap are limit orders, deploying on sidechains/layer-2s, onboarding external solvers, and much more!
Hillrise Group supports ambitious Web3 startups with early-stage venture capital and fundamental research.
CowSwap is a DEX aggregator that enables trading with MEV protection.
Connect with Hillrise Group
https://hillrise.group
https://hillrisegroup.medium.com
https://twitter.com/hillrisegroup
