[Discussion] What's Balancer's current Primary Objective?

[bakamoto20]

Some of you who are a part of Discord will have noticed the discussions over the last few weeks (within the Liquidity Mining committee in particular), about the types of pools we create and incentivise. Within this there have emerged two quite different ideas about Balancer’s current primary objective, and I want to get a better sense of what the community thinks, in a less free flowing context where parts of the debate often get lost.

These two differing views on what our current primary objective should be can be summarised as follows:

1. The current primary objective of Balancer is to provide useful liquidity to traders and a diverse range of options for Liquidity Providers: both in terms of assets, and market positions. Primary KPIs are TVL & Trade Volume. Optimising for fee generation is primarily a future concern, that requires us to first establish a critical mass of liquidity and attract good flow. LPs aren’t only interested in fees, they also care about the market position they’re taking, and recognise that often market position is much more important than fees. Currently fees are subsidised with BAL rewards & pools should be trying to slightly undercut the market to attract volume.
2. The current primary objective of Balancer is to generate fees. Each pool should be designed in order to generate the maximum fees we can from arbitrage & people doing less-than-all asset deposits/withdrawals to pools. Pools should be included/excluded from BAL rewards based on whether they are currently generating high fees. Our focus is attracting LPs into pools that do high fees, regardless of whether these pools are useful to traders. Balancer will never get “good flow”, so don’t bother trying. What’s most important is that Balancer can show high fee APYs against pools, start taking significant protocol fees, and build a treasury.

I fall very firmly into camp #1. I believe we should be creating a wide array of pools with competitive fees for traders. Index-type pools can fit into this strategy, but we shouldn’t be trying to force USD into them all (not all LPs want USD exposure) to generate more arbs, and we shouldn’t currently be incentivising indexes with fee levels which make them useless to traders (above 0.3%), as that’s effectively wasting our liquidity’s potential to add value.

I believe it’s important we can get the community to align on what we are pursuing here, and will defer to the community’s judgement. At the moment, the same debate is happening each week, where a fee maximiser (#2) is widely publicly campaigning to increase fees on pools & add pools with fee maximising structures, plus quickly drop pools that don’t generate as many fees; and a TVL/Volume maximiser (#1) is at odds with them.

My view is that we need to align the community on this, not split and end up with a hashed compromise that means there’s constant friction every week debating fee structures/horse trading for pool fees, rather than what pools we should be creating & incentivising to achieve Balancer’s primary objective.

At any stage, Balancer can revert to high fees & extracting fees from arb volume, if that’s what we want to be. The arb volume will never go away. The window to become a competitive DEX is one which will continue to narrow and get harder/more competitive, however. If we want to achieve that objective we need to throw 100% at providing competitive liquidity, and make it crystal clear to the wider community that we fully intend to be a competitive DEX: not solely monetise arbs & less-than-all asset deposits/withdraws.

The more I’ve thought about it, the more entrenched I’ve become in my view: we’re only going to be a competitive DEX if we’re aligned in trying to design & incentivise that, and throw everything at it. The constant debate and campaigns to focus on high fee % high volatility pools, and public negativity around the idea that we can even be a competitive exchange is bad for our marketing/growth (if that’s what we want to be). Other DEX’s we’re competing against are not arguing each week whether they can even be viable or not.

I believe that with the vault, asset managers, internal balances, stable pools, smart pools & wide array of functionality being built over the coming months that Balancer is a competitive DEX and can go on to be DeFi’s primary source of liquidity. I believe the future of DeFi trading is aggregators: integrated into wallets, portfolio manager apps, etc. The “go direct to Uniswap to trade” is going to slowly cease to be a thing over the years as market participants become more educated. We will continue to be competitive with Uniswap V3, Sushi, Curve, etc as the product evolves. But I don’t think we will compete as a DEX unless we’re all aligned around actually doing that.

Perhaps we should consider having a governance vote on some Primary KPIs for Balancer, that we can align on aiming towards during 2021, and all work towards achieving?

If we at least have community agreement on whether we look to serve traders with our liquidity or not, it will make Liquidity Mining decisions easier: fees cease to be the primary issue, and pool design to maximise broad, useful liquidity within the BAL allocation we have available can be the focus.

I welcome the community’s discussion around this, and recognise my view is just one of many (hence the desire to find some solid ground on this we can build from). I’m sure others will have a lot to add, fleshing out both sides of the debate.

[DavisRamsey]

It all sounds nice, but I didn’t see any piece of actual data to support what you’re saying. Meanwhile, I can point to the GTC pool and the conclusion is quite clear that undercutting on fees has zero impact. Uni-v2 has 10% of our TVL but does more volume than us despite our fee being half theirs. Now you’ll argue that things will change in the future, that ppl won’t go direct to uni but to aggregators and THEN we’ll win with lower fees.

I can’t argue against hope and theories. All data I’ve seen + my experience tells me lower fees will not attract traders. I also believe LP’s are farming our 2 token pools not because they want the exposure but because they want free BAL rewards. Would any LP really argue they would stay in 95% of our pools if we remove BAL rewards? “But that will change… etc etc.” There will be a process to educate & get feedback from LP’s on what multi-asset pool design works but to say LP’s will not participate in multi-asset pools designed for fees is extremely short sighted imo.

It is difficult because I know Balancer could become one of the highest revenue protocols very quickly if we stop trying to become a budget Uniswap. We have many advantages but we refuse to use them because we’re trying to become something we’ll never be in my opinion.

I understand you’d prefer that I stop offering my opinion and become a silent yes man, since the only hope of your strategy working is if we commit to it 100% (your words) and I’m distracting from that. I sympathize… but it is not possible for me to stand by and not say anything when I don’t agree with what’s being done. I’m not here for the 100 BAL/month, I’m here because I deeply believe in Balancer’s potential and it saddens me that in my opinion we’re actively choosing not to embrace it.

It is true, I cannot be 100% certain I am correct. There is a chance I’m wrong, you’re right, and me being here is doing WAY more harm than good. I feel like as long as I am here, I have to speak my mind. Perhaps in the future when we get some data showing your strategy is working I could more easily get behind it.

If we aren’t interested in a liquidity mining program driven by data but by hopes & dreams, it is probably best that I’m shown the door as soon as possible so you all can get on with it. I think the Ballers can vote to remove me so perhaps that would be a good option to pursue.

[bakamoto20]

To be clear, I’m not suggesting we should only have Uniswap-style 2-asset pools. It’s perfectly possible to have multi-asset pools which also provide liquidity that is useful to traders. It’s not a binary choice WRT pool design. I am all for shifting more rewards to multi-asset pools, high fees is my top concern.

The liquidity mining committee was given a remit to assign rewards to pools, but as this has gone on it’s become clear to me and others on the committee that we’re now deciding the future of Balancer in a major way that perhaps wasn’t intended with the proposal, of which the conversation focused more on asset selection than Balancer’s fee model. The goals of LM as it stands are too diffuse, such that we’re both trying to optimise for entirely different future visions.

I don’t want to kick anyone out or anything like that, you’ve contributed loads to Balancer, much more than I have. I just want to get input from the community and ensure we’re rowing in the same direction, if the community agrees with you, then I will happily shift towards that model.

What I don’t think we can do is middle in this half way house where we’re hashing the debate out every week and trying to drag things in opposite directions. I believe the middle zone is pointless, as we’re not giving it our all to be a central source of defi liquidity (which we have to if that’s what we want) or building to compete strictly as an index project (ignoring traders altogether), and confusing people w/ two very different ideas of what Balancer is trying to be.

[Andrea81]

Hi @bakamoto20

In my opinion, Balancer should find its own market niche and expand its presence from there. I also agree that at the moment the protocol is neither fish nor fowl and this creates image and brand recognition problems.

I believe that the time has come to experiment with something new with more or less aggressive pools to attract those who intend to manage their capital depending on the market’s conditions (option 1?) I also agree we shouldn’t be so fixated with USD in every single pool. In this therefore you have my support. If the “experiment” fails, we will always be able to change strategy with data in hand.

@davisramsey please do not become a yes man. I always read your comments and ideas and I’m often surprised by the observations you make. You get a view of things from angles that never cease to amaze me. Your contribution and experience is so valuable. That said, sometimes it’s good to let go and hope things work out. What do you think?

[DavisRamsey]

Appreciate that, thank you

And you’re absolutely right. As of last week, I told bakamoto I will auto vote on yes on whatever he proposes. He can choose to listen to my opinion, or ignore me, totally up to him. He is in complete control of the LM program as far as I’m concerned.

[danneskjold]

Complete noob here. Perhaps I can help by being outside this ongoing discussion until now.

When I first discovered Balancer, I realized that the self-balancing index was a superior idea and that protocols like Uniswap passed up a major opportunity by sticking with 2-token equal-weight pools. This is Balancer’s #1 advantage, full stop. We should focus everything we have on showing potential users the benefits of providing liquidity in a balanced portfolio rather than spread across multiple different LPs or just hodling.

Currently, most of the liquidity on the ETH chain is in 2-token pools. Some have weights other than 50-50, but this isn’t taking advantage of our uniqueness enough. I like that the Polygon site is focusing more heavily on 3+ token pools which have already attracted lots of liquidity. Users need to see and use these first and foremost, to break the mold in their mind of 2-token pools only. I would say the #1 KPI by far is TVL, which indicates users are buying into the multipool idea.

Onto fees. The average mature LP (once it has attracted TVL proportional to its volume, and no liquidity mining) achieves rarely more than 40% APY, looking at Uni, Sushi, and Balancer’s main pools. But this is crypto - volatility is the name of the game. If a liquidity provider takes a position in, e.g., ETH-LINK-AAVE-BAL, they are exposing themselves to way more than 40% upside AND downside risk. In any normal crypto market, even 40% APY in fees is meaningless if the underlying token has pumped or dumped. Let alone 10% which is what most Balancer pools on mainnet are getting. We NEED to have a value prop for LPs other than fees, because fees are not actually that valuable in markets this viable.

That being said, I’m not sure that we care too much about traders other than arbs, because our primary product is self-balancing portfolios. I’m not even sure we should market ourselves as a DEX, and we don’t need to worry about undercutting UNI with fees - we’ll never attract liquidity that way. I agree with @DavisRamsey that we need a proper revenue stream, ESPECIALLY for BAL holders, something that most protocols are missing. I also agree that we have not done the proper testing to determine which is a superior fee model. Also, we need to seriously calculate whether 1. high fees mostly relying on arbs or 2. low(er) fees attracting retail traders will generate more revenue.

So I’ve found myself somewhat in the middle. Here’s some principles you could consider.

1. Attract liquidity first and foremost.
   a. Build 3+ token pools with complimentary tokens that users would actually invest in
   b. Reward pools with high TVL that are 3+ tokens or unique weights
2. LPs and BAL holders are our customers, NOT traders.
   a. Establish a fee model that rewards BAL holders and LPs over traders – we will always have arbs

Some potential action items:

1. Do the math on whether fees change volume, and more importantly, which fee models maximize TVL and revenue (I recognize these will conflict but worth investigating)
2. (HUGE) Ask users which 3+ token pools they’d like to see on Balancer, build them, and incentivize them. This is our #1 reason to exist, we need to use it.

@DavisRamsey I’m interested in your thoughts on how Balancer could become one of the highest revenue protocols very quickly, while retaining the main purpose of having multi-token pools.

[bakamoto20]

This is awesome input, thanks. With regards to the math action item, I think the challenge here is that as it stands we need much greater TVL to get aggregator trades routed through us in large enough volumes to move the needle and prove it out. You need both deep liquidity + competitive fees to take the volume. The big unknown (and I don’t really know how to get started on the math) is how much the vault is going to impact how much volume we win from aggregators as TVL increases. Last I heard they’re still not really using batchSwap: if aggregators are doing single swaps through the vault, we’re truly blind to what that difference will mean at the moment on ETH.

Edit: it’s hard to look at current data to predict the future, too. Really a lot of it depends on whether aggregators end up taking the volume from the likes of Uniswap, Sushi, etc direct.

I agree that from an LP perspective, fees are less relevant than the gains in underlying assets. It’s the ability to go long on a basket of assets and forget about it that’s really interesting to a lot of people on that side, which is why I oppose the inclusion of USDC in every index pool (it becomes a token to increase fees, at the expense of creating greater impermanent loss - with just arbs I’d be surprised if this works better than pools without a stable, unless you enter & leave the pool when most assets are similarly priced in USD – which is a specific market position to take).

As an LP, I always generally sought out pools that I thought were long term correlated, but short term would have volatility. I think it’s wrong to just focus on maximising “uncorrelation” for fees: what you end up doing is disguising the impermanent loss with the high fee APY.

[DavisRamsey]

Thank you for your contribution here, you’ve raised many good points.

As for how we develop into one of the strongest earning protocols, it is a process of iterating on what is working versus sitting tight and hoping we attract traders with our low fees.

It is not going down the list and adding USDC to every pool, but given ETH/USDC is our highest earning pair per BAL spent + generates higher TVL per BAL rewards compared to ETH/DAI, one simple way to start getting LP’s used to the idea of multi-asset pools is to trial some 3 asset pools token/weth/usdc. This should also result in much higher sustained fee generation (even if we do not edit the swap fee).

I believe there is market demand for pools including USDC, and polygon is evidence of this imo. When I was pretty much all in on 8 token pools last summer, the reasoning was by pooling everything together and generating passive fee income, I de-risk the entire position. I have no idea if YFI, SUSHI, or AAVE etc will go up the most tmrw and I have no edge in trying to short term trade. If your pool is generating single digit fees without attracting a high TVL, it is evidence that there is no demand outside of free BAL rewards. Without BAL rewards, nobody would LP in it imo.

As you move down the risk spectrum, it’s more tenable to pool without stables as the expected volatility is much higher. I really don’t think we’re offering a differentiated product by throwing up a WETH/AAVE/LINK/BAL pool - these are all established projects with a high market correlation. If someone wanted broad defi exposure, there is ZERO reason to go with this pool over a DPI for example. If someone was bullish on one or all of these projects, why pool them for low single digit APY instead of just holding in your wallet or single asset staking when applicable for similar APY? The answer is, because we’re giving away free BAL rewards.

How we become a revenue generating monster is by looking at the numbers, what pools can generate fees comparable to the necessary TVL to outperform the BAL we’re spending on it. If a pool can only generate single digit APY but cannot attract necessary TVL to bring BAL rewards APY to single digits, I really do not see how this benefits Balancer protocol.

[bakamoto20]

WRT USDC, there is demand yes, especially when we have mega high APYs and are in a sideways/bearish market at the moment. That demand would quickly shift if we switched bullish though. I think that “no one would pool without BAL rewards” probably could be used to argue about most of the pools at the moment, something I see as mainly a side effect of us not generating enough good flow.

The reason we have that pool is because LINK-ETH, AAVE-ETH do good volumes on Polygon & we’re trying to kick-off BAL liquidity there. From an LPs perspective, there’s plenty of reasons to pick that pool over DPI. For one, you know that there’s volume on Polygon in those pairs that can be captured. Secondly, you may want exposure to those specific assets equally but not all of DeFi in a weighted index.

When it comes to correlation, I disagree that WETH-AAVE-LINK-BAL are as correlated as you say. They’re long term correlated but carry short term volatility which generates trading activity/fees while mitigating impermanent loss over the long haul.

I believe the only evidence we’ve generated from Polygon so far is that people will pool most assets for a very high APY. I don’t know what else we can take away from it after a couple of days. The reason the USDC pool has more liquidity in it is that the 1% swap fee was collecting more fees from people doing single-asset adds for a while, driving the APY up on the right (people picking higher number), and people are generally more market neutral than usual at the moment as we’ve been ranging: they want more in USDC as it’s seen as safer in current market conditions, to collect the high APY.

Our stable pool will probably attract by far the most liquidity because of the market conditions and people hunting for stable yield. I think it’s wrong to draw long term conclusions about people’s interests based solely on the current market. There are many people in Crypto who don’t want to touch dollars.

[Eugin]

Thanks for bringing forth this discussions, @bakamoto20. Shout out to @solarcurve ( DavisR) because u make us stronger.

~
I feel that BAL’s primary objective should be inline with V2’s vision of being a “portfolio manager and trading platform”.

In my opinion that means trying our best to attract portfolios (“LPs”- high TVL), and being an effective DEX ( high volume/ capital efficiency). With stronger trading volume, dynamic fees and vault optimization in place, it will eventually translate to more benefits to LP and BAL holders.

And that is more aligned with option #1.

~
My personal experience is that DEX volume originate from 3 sources:

Source 1 (S1): Arbitrageurs - attracted by price mismatch, profit-maximising bots
Source 2 (S2) : Direct website - attracted by network effect, special features, network effects, decent price, variety liquidity
Source 3 (S3) : DEX aggregators - attracted by best price and liquidity

For most DEXs,
Arbitrageurs (S1) account for a “base/floor” level of liquidity, which increases with market volatility.
Direct flow (S2): eg. Uniswap, Sushiswap
Aggregators (S3): eg. routing Curve liquidity ( curve has deep stables liq.)

Option #1 allows us to target all 3 major demand sources.

Balancer has strong competitive edge in the market such as LBP which provides vertical integration - it’s such a wasted opportunity that legit projects use LBP for token launch and eventually leave for Uniswap pools - this can certainly be upsold with LM. UI/UX is also top notch and improving with V2 launch and a major drawing factor. Aggregators are also fully integrated.

Option #2 suggest that we optimise/increase our fee structure to “arb” fees from arb bots to improve protocol revenue.

Breaking down trading flows in this scenario means we will receive flows from S1 (arbs), but low/no flows from S2 (website) and S3 (aggregators). My fear is that bots will eventually smarten up and be optimised to transact ONLY if they can make money through arbitraging, just like how it is currently optimised for gas.

This means that we will not get as much fee income as now, and DEX prices may be inaccurate and we lose competitiveness. Aggregators may make the logical business decisions to cut us off and we lose another powerful source of volume.

~~

On LP provision and being portfolio managers:

I think we need to improve on educating the community and that can be administered through LM too.

In general, i feel that LPs are predominantly concerned with impermanent losses.
In the market and to compensate liquidity, LPs are commonly “made par” with LM rewards. This is often pitched as an intermediary bootstrapping method before fees are high enough to compensate LP. ( eg This is doable - Uniswap is a successful example, UNI does not provide any LM rewards)

We need to educate that apart from LM rewards, BAL has 3 levers such as market dynamic fees, multi asset and adjustable asset weights which is uncommon in DEXs. All these levers can be “used” to configured pools to alleviate impermanent losses and provide good gains. While education takes time, we need users to familiarise themselves with BAL and thats where the LM rewards come in.

Lastly, we need to accept that some liquidity providers prefer 2 assets portfolios and cater to them. On this, I feel that we need to have the flexibility to support more 2 asset pools. I am all for incentivizing different asset configs but find it limiting if we exclusively incentivize multiasset pools. ( such as our Matic LM program)

To achieve all this, we need to be more flexible with our LM programs and have the ability to attract more liquidity, projects and community into Balancer - which in turn integrate Balancer deeper into the wider DeFi community.

[neo-renaissance]

Additionaly to the 3 sources you mentioned, volume can also come from other projects building on top of Balancer (and possibly providing their own UI).

[ValeryWeb3]

Hi Guys .very like what you’re doing and thank you!
I totally agree with You about pools with 3 tokens instead of 2, they’re more competitive than pool’s with 2 token. (But also I have bal/eth pool and will not go from it ) also agree with Davis to cut gtc/eth pool rewards.

I ve been with balancer since june 2020 and strong Bal holder. I see a lot of discussion on forum and discord about Apy /Apr and lividity miming and how to attract traders.
I believe and see that firstly we need educate people how to become an Asset Managers !
As for me , I don’t need Lividity Mining at all, because I make more money as asset manadger and rebalance my portfolio once a month or once a week ( depend on market condition) . As for nearly months I see big potential to increase my Eth stake with Eth2x FLI and DPI and therefore increase ETH in USD. And it will be much more money than I can get from lividity mining(will be just a small bonus ). So I suggest :

1. Focus on 3/ 4 pools (with 3/4 tokens each )
   2 . Educate people how to grow their stake with these pools. ( And they will Not come Only for Purpose Mining Rewards ) . And this is our Goal!

Some people want increase USD, others want increase Eth, also we have OG with BTC. We will have all categories of users in our pools and therefore hudge TVL. Hudge TVL → more fees , more traders , more arbitrager , more protocol revenue , more happy people, better tokenomics. And Balancer will become Number One Protocol in the World !

I think you already knew all of the above , but I hope my thoughts will be useful !

[bakamoto20]

Hi Valery, thanks for your input here. In terms of pools with 3 tokens being more competitive than 2, can you explain what you mean, and who for? I definitely agree on the education side of things, although we do probably need to be careful about bridging into a “becoming a financial advisor” position. It probably needs to be quite community led.

When you say focus on 3/4 pools also, do you mean 3/4 pools in total? Or do you just pools with 3/4 assets in them?

[ValeryWeb3]

1. to focus on 3 or 4 basic pools that everyone will use. for example Eth/btc and Eth/usdc . Eth is the most traded asset on the market. many use aggregates for trading. with big TVL and our low commissions people won’t even know that trades go through Balancer.
2. in some cases pools with 3 tokens being more competitive than 2 tokens for those who have concern with impermanent loss. eth/btc/usdc for those asset manadgers who count their money in USD . dpi/btc/eth or link/eth/bal/aave for those who count in tokens.

[AHurwitz]

@bakamoto20’s objectives outlined in 1 and 2 encompass two strategies followed by high growth startups and lean or “indie hacker” projects.

1. High Growth Strategy

Optimize for user growth by taking advantage of the resources (BAL) available. Done successfully, the team and stakeholders are aligned on the long-term vision for the users AND the business model.

For example, Amazon communicated to shareholders for a decade that they are going to take an accounting loss in order to capture market share and build the business model via recurring fees with Prime and infrastructure services like AWS.

In addition to finding alignment throughout the Balancer community through proposals and voting the community must have a long-term business model in-place that they are working towards. Otherwise, the time and capital spent to obtain market share will not be constructive.

With Balancer’s 100m tokens, and a max of a 65m distribution to LPs at ~$7.5m/yr, that provides a runway of max ~8 yrs for distribution to build out a sustainable business model. However, it is likely in a much shorter time frame there will be clear winners such as in the next year.

2. Lean Startup Strategy

In the indie hacker approach the focus is allocating resources to what is working or will work in a much shorter time horizon. This approach is safer when building with a small team and limited resources as it allows for failing fast, and to scale what works best.

The risk for the AMM/DEX/Liquidity Pool space is that the rest of the competition seems to be taking the high growth approach and is fine taking an accounting loss to distribute rewards and gain market share.

3. Hybrid Approach?

Perhaps a hybrid approach can be agreed upon in order for the community to move forward more effectively without the constant pull in opposite directions as described above.

For instance, what if 80% of the resources (rewards, engineering) went to scaling the user base, while 20% building out some “killer pools” with high APR to build a smaller passionate segment of the user community.

Liquidity Providers (LPs)

I also believe LP’s are farming our 2 token pools not because they want the exposure but because they want free BAL rewards. Would any LP really argue they would stay in 95% of our pools if we remove BAL rewards?.. Without BAL rewards, nobody would LP in it imo. - @DavisRamsey

There is “smart” and “dumb” money.

My hypothesis is that smart money will stay if the APR from the swap fees makes depositing in the pools profitable.

Dumb money will follow the pools with the best rewards across AMMs and protocols in the short-term. Smart money will look for sustainable business models in the technologies that they have long-term conviction in. Smart money typically has stronger conviction in specific protocols vs. index fund like tokens.

…fees are less relevant than the gains in underlying assets. - @bakamoto20

Continuing the theory above, smart money will look at this holistically. The underlying gains is not a significant factor since the assets are likely already held by smart money whether it’s in cold storage or deposited into a pool. In order for the underlying assets to come out of cold storage into a pool there needs to be a long-term incentive worth the inherent smart contract risk.