[SCP-131][Incubation] Added fees for Thorchain swaps

Summary

If adopted, this proposal would enable the DAO to experiment with implementing fees on top of all Thorchain based swaps executed through the Shapeshift application.

Abstract

Starting Jan 15th, all Thorchain based trades would be charged a fee of 10 bps, increasing 5 bps every 30 days until fees reach a maximum of 25 bps.

Motivation

Since inception, the DAO has pursued a singular vision and business model with respect to revenue generation. To date, there has been little to no data based evidence suggesting the sustainability of this approach given measured user growth and revenue per user. Therefore it is imperative that the DAO continue to experiment with alternate revenue models in an effort to find a path towards self sufficiency.

Specification

Over the next 12-18 months the DAO needs to iterate rapidly towards finding a sustainable business model. Assuming a customer value of .25$/month, and 250k/month of expenses, we require roughly 1,000,000 monthly active users to break even. Based on the most recent data from Pendo, we have approximately 17,000 monthly active users and are losing >%15 of users per month since our peak MAU of 21,000 monthly active users (11/13). We have two paths to pursue aggressively; increasing MAUs and increasing revenue per user. All reasonable experiments that align with either of these paths are worth the DAOs consideration.

The proposed experiment is meant to determine if users value our platform enough to pay marginally more for it than a completely free option when swapping assets via thorchain

Thorchain Swap Fee schedule:

Jan 15 - Feb 15 - 10 bps

Feb 15 - March 15 - 15 bps

March 15 - April 15 - 20 bps

April 15 - TBD - 25 bps

An additional proposal and governance vote would kick off around May 15th that would examine data collected during this experiment and determine fees beyond May 15th.

Benefits

Current trade volumes for thorswap are approximately 5M USD / month, this would translate into the following revenues for the DAO

Jan 15 - Feb 15 - $5,000

Feb 15 - March 15 - $7,500

March 15 - April 15 - $10,000

April 15 - TBD - $12,5000

While these monthly amounts may seem small, across all of our business models the Shapeshift DAO is betting on very large (> 15% MOM) user growth in all of its projections. Additionally, even the $5,000 number realized with the lowest fee increment would be our largest source of current monthly revenue.

Implementing fees in a single product allows us to gather important data from our users in a lower risk setting than implementing fees across the entire product line.

Drawbacks

There is the potential that our users are price sensitive and these additional fees cause us to lose existing users or fail to gain additional users. We can not run a true A/B test in this scenario and teasing apart effects of this treatment compared with broad market sentiment, product evolution and other user trends may be difficult, leading to inconclusive data.

As with most open source software, there is the potential for someone to fork the Shapeshift codebase and bypass the fees for personal or public use.

Marketing and PR campaigns would need to be modified to ensure users are not being misguided by the original message of “no added fees”. There would be a small engineering effort needed to augment the existing swapper UI to display the added fees and add the transaction memo when interacting with thorswap. Additionally, the DAO would need a RUNE address that is capable of receiving the fees denominated in RUNE.

References

  1. https://dev.thorchain.org/thorchain-dev/concepts/fees

Discussion Topics:

Should we implement a regressive fee? Larger trades could be charged a lower % fee than. This is technically feasible and could make fees more equitable for larger traders. Many CEXes have similar volume based fee tiers which work on aggregated volumes

For reference, I am posting the Pendo chart of Weekly Visitors. Please note, that the last data point for the current week data is incorrect.

Fee % based on, say holding Fox (foxy) for example? or a hedgey timelock etc.?

As our initial step into fee’s are on Thorswap, that we dont have any ref gen on at all, I like this idea.

I am certainly open to any of these ideas. I believe there may be some non trivial engineering to get this to work due to needing to connect multiple wallets if swapping an asset like BTC but needing to reference someones ETH address for proof of ownership.

Strong support for this.

Sustainability for the DAO will be a combination of increased users and revenue generated per user, and this is a step in the right direction, but not a panacea for all our problems.

I’ll post some quantitative data that demonstrates that demand for swaps is largely inelastic to fees, ie the same interfaces can see large volume across a variety of basis points charged.

Indeed, on calls with Nine Realms it was suggested to us that implementing fees is the simplest and suggested approach to monetizing our existing and growing trade volume.

Thanks for posting this, 0xean.

I look forward to hearing the community’s thoughts on this, but my stance hasn’t changed changed. I remain strongly opposed to adding fees onto any of the protocols or services we integrate, and believe ShapeShift has the greatest potential to both survive and thrive if we continue down our path as a public good.

Not only do I support keeping shapeshift free, but I’m a strong supporter of the longstanding plan to reward users that use shapeshift and generate revenue for the dao through one of the growing number of affiliate revenue streams with FOX proportionally. Unlike fees, this can be a sustainable, winning model that makes shapeshift the best way to use a growing number of protocol services.

With current usage levels, enabling fees doesn’t increase our likelihood of survival. In order for ShapeShift to survive, we need to attract and retain orders of magnitude more users. Our best hope at achieving this with in our runway is to focus on making our product as attractive as possible and as frictionless as possible. I believe we should try new, disruptive strategies like free + ownership rewards that create unstoppable network effects rather than old web 2 strategies like fees that are a race to the bottom and add friction.

Let’s execute on the proposed roadmap over the next 3 months to get the product to a solid place, make FOX rewards the top priority after this, and see whether the world finds a free, open-source, decentralized interface valuable. If we give no fees + FOX rewards a real chance and find out that it doesn’t work, I’ll be more open to considering fees.

I’m afk in the middle of moving, but look forward to continuing the discussion. In the meantime, here’s a talk that covers some of the reasons why I believe no fees are the wei: https://m.youtube.com/watch?v=y-TOlyguMjQ

Also just to clarify, I’m not against fees on the protocol level, and support fees on Yieldies or any other protocols the DAO develops. I just think adding fees on top of open protocols and services is not a winning strategy in a web 3 world, that ShapeShift can and should DAO better, and that no fees + fox rewards is the wei.

g

But in thorswap example, there is no protocol level Revshare is there? affiliate or otherwise?

Correct, there’s no protocol level, so in the absence of that, clearly users find value in the service we provide (an interface to the protocol), as they do with large competitors, and we should add fees to monetize this value, as recommended in calls with Nine Realms.

  1. I largely would like to see this explored more in 3 months after some base-functionality is added and we have a better handle on the app having industry parity features + Arkeo is completed. I would like to better assess app use and other ways we can leverage the costs/ "fees" onto integrated partners instead of the consumer to create a more frictionless experience through either sponsorships, paid promotion/feature promotion, and marketing initiatives. I think there is a large appetite of partners who we have integrated that would love to sponsor/push their product to our consumer base more and would be willing to pay comparable rates to those generated by THORChain fees on users. For example, just this last month as a test of this concept, marketing has tentatively secured $5k+ of funding for some web sponsorship opportunities from different partners such as Hedgey, Banxa, and OnRamper. This is comparable to the revs generated in month 1 of the proposal.

    The big question people keep asking (or rather the question I keep hearing) is “so what if we get a large userbase, how can we monetize them if the app is free/no added fees?” From my perspective, having an engaged community of users is a huge selling point for other protocols to want to then pay us to list them/host them. There is a future where there is another protocol is developed that offers the same security/offering as a feature like THORChain. When that happens, we are then in a position to negotiate with the new protocol as well as THORChain for incentives to push their feature to our users and at that point are generating revenues from the users, but it’s coming out of the pockets of our partners to access then our ecosystem we have nurtured and created. I am imagining a future where we can even use our MAU and platform to renegotiate fees on partners where a protocol is paying us to be listed (or giving us a better rev share agreement) and we have enough other options that offer comparable things so that if they don’t want to pay the fee to be listed or increase the Rev Share, we just turn that feature off and give that business to their competitors (something they naturally will not want if we have an active enough user base). That to me is a very compelling business case for us that gives users a frictionless experience and loads the fees onto the partner so the user can conduct business freely.

    Last thought in regard to this: I am currently working with Product to try to monetize CTA’s and clickable from our partners that would feature their products to our users based on their interests. This provides 2 things:

    Revs from product integrations

  2. An improved user experience by being able to offer users things we see they might need/are interested in and being able to offer the best/most affordable rates in one UI (while not imposing additional friction on the customer experience)

One thing that I thought might be a cool meet-in-the-middle (and an option I think would satisfy both camps here) would be creating a browser extension of app.shapeshift.com (or plugin or whatever the cool kids call them today) and then charging fees on that for everything (for the convenience of using ShapeShift wherever). If someone was really hell-bent on getting no fees, they can go to the browser version that offers free features (where they will still be generating revs for us by viewing/engaging with partner related content that is relevant to them). The other alternative/thought I have in this lane is offering fees on the Arkeo version of ShapeShift (you pay extra for the extra decentralization) and maybe that is the place we push people to when marketing it (but still have the free /not as decentralized version available for people to use)

I would love some time to explore the potential of this generating revs for the above reasons before we start adding fees to features.

I like those ideas hpayne. the current thorchain connection, with no rev bugs me alot is all. but finding another partner that does similar, and getting aff fee from them, would push thorchain to either start aff fee, or lose out. not a bad concept either.

Thanks for all of the responses and am glad to see people thinking about different business models for the DAO.

We have been saying we need more time for parity and features to see user growth for a very long time (https://forum.shapeshift.com/thread/an-open-letter-on-strategy-and-execution-40626). To date, it would be hard to back up with data that the market has seen value in what we have built. I believe if we are to survive we need to start listening to the market more and less to our preconceived ideas of what the market values.

Experimentation is crucial to us finding this out. The time to start these experiments is now when we have a small, but measurable user base to work with. The skeptic in me says we have been running the same experiment for the last year and getting the same results from it. We need to find something that someone is wiling to pay for and continue to iterate down that path.

As a thought experiment, lets say we find out that people are okay with added fees on Thorchain and our revenue per user goes from .25$/month to closer to $1/month, a 4x increase per user. In addition to cutting down our target user base from 1MM MAU down to 250K MAU we now have data to suggest we should aggressively pursue more thorchain trading market share. Perhaps that leads us to consider value add features on Thorchain, like limit orders, stops, or some other wrinkle that others front ends are not offering (this may or may not be technically feasible, but you get the idea).

We have all spent a lot of time at this point looking at different models of user growth and revenues using the affiliate revenue only business strategy, and none of them have shown a sustainable path forward. I am not sure adding thorchain fees is the right experiment, but firmly believe if we don’t start experimenting with different strategies now, we will end up in the same place in 3 months time.

I want to add some quantitative data here to counter the point that adding fees will significantly impact user growth. Data was pulled from flipside crypto SQL queries, and massaged in the following sheet https://docs.google.com/spreadsheets/d/1DMnHTkbRLXnYizpZDsWjPLblj0rhn5o97-lGpqNj0/edit#gid=1047496145

  1. I'm happy to provide access upon request, links to the queries used are in the sheet. The data was pulled mid-December, I can update it if desired, it's still illustrative, albeit stale.

    The leftmost column is the affiliate address, the middle is the basis points charged, and the right is the aggregated trade volume for the given affiliate at the given fee.

    Several things are apparent

    There’s large volume at the 30bps level, across multiple affiliates, evidence users aren’t adverse to paying in the region of ~30bps.

  2. We have done a fraction of the volume of other affiliates charging fees. Our zero fees does not seem to be attracting outsized volume, compared to competitors that charge fees.

Affiliates volume does not seem sensitive to fees charged, i.e. the same affiliate charging higher feels often do higher volumes. Note, I am not suggesting higher fees cause higher volumes. I am suggesting that demand for THORChain swaps is largely inelastic to reasonable fees charged. Econ 101 read on price elasticity for those interested https://en.wikipedia.org/wiki/Priceelasticity

  1. of_demand

While I don’t deny that qualitatively, adding fees will make it harder to acquire users, I think the way arguments have been framed as a dichotomy of “we can have users or fees but not both” (hyperbolic) is disingenuous.

I think these queries/data go a long way to helping quantify how user acquisition will be impacted, and indeed it does not seems to be an issue for numerous competitors.

Adding fees is also helps us cross the penny gap https://redeye.firstround.com/2007/03/thefirstpenny.html - are users willing to pay anything for your product? Note the crude charts in the article depict what the on chain data suggests - users will pay fees, not only that, they’re somewhat indifferent to fees charged, as long as we’re not price gouging (which the suggested fees in the proposal do not).

Monetizing our products not only provides crucial revenue for the DAO, but also helps close a feedback loop of where we can generate a return on our efforts.

Users will pay for value. We’ve created something of value. Users pay for the same value at competitors. We should charge for the value we provide. We are leaving money on the table.

This proposal is not a panacea for the DAOs cash flow troubles. It won’t make us profitable overnight. 30bps fees on our current, and thankfully, increasing volumes, would fix our cash flow problem by ~10% at the current burn.

If anyone has specific questions, I will try to provide quantitative answers with more on-chain data.

I want to add some quantitative data here to counter the point that adding fees will significantly impact user growth. Data was pulled from flipside crypto SQL queries, and massaged in the following sheet https://docs.google.com/spreadsheets/d/1DMnHTkbRLXnYizpZDsWjPLblj0rhn5o97-lGpqNj0/edit#gid=1047496145 I’m happy to provide access upon request, links to the queries used are in the sheet. The data was pulled mid-December, I can update it if desired, it’s still illustrative, albeit stale.

The leftmost column is the affiliate address, the middle is the basis points charged, and the right is the aggregated trade volume for the given affiliate at the given fee.

Several things are apparent

  1. There's large volume at the 30bps level, across multiple affiliates, evidence users aren't adverse to paying in the region of ~30bps.
  2. We have done a fraction of the volume of other affiliates charging fees. Our zero fees does not seem to be attracting outsized volume, compared to competitors that charge fees.

Affiliates volume does not seem sensitive to fees charged, i.e. the same affiliate charging higher feels often do higher volumes. Note, I am not suggesting higher fees cause higher volumes. I am suggesting that demand for THORChain swaps is largely inelastic to reasonable fees charged. Econ 101 read on price elasticity for those interested https://en.wikipedia.org/wiki/Priceelasticity

  1. of_demand

While I don’t deny that qualitatively, adding fees will make it harder to acquire users, I think the way arguments have been framed as a dichotomy of “we can have users or fees but not both” (hyperbolic) is disingenuous.

I think these queries/data go a long way to helping quantify how user acquisition will be impacted, and indeed it does not seems to be an issue for numerous competitors.

Adding fees is also helps us cross the penny gap https://redeye.firstround.com/2007/03/thefirstpenny.html - are users willing to pay anything for your product? Note the crude charts in the article depict what the on chain data suggests - users will pay fees, not only that, they’re somewhat indifferent to fees charged, as long as we’re not price gouging (which the suggested fees in the proposal do not).

Monetizing our products not only provides crucial revenue for the DAO, but also helps close a feedback loop of where we can generate a return on our efforts.

Users will pay for value. We’ve created something of value. Users pay for the same value at competitors. We should charge for the value we provide. We are leaving money on the table.

This proposal is not a panacea for the DAOs cash flow troubles. It won’t make us profitable overnight. 30bps fees on our current, and thankfully, increasing volumes, would fix our cash flow problem by ~10% at the current burn.

If anyone has specific questions, I will try to provide quantitative answers with more on-chain data.

I want to add some quantitative data here to counter the point that adding fees will significantly impact user growth. Data was pulled from flipside crypto SQL queries, and massaged in the following sheet https://docs.google.com/spreadsheets/d/1DMnHTkbRLXnYizpZDsWjPLblj0rhn5o97-lGpqNj0/edit#gid=1047496145 I’m happy to provide access upon request, links to the queries used are in the sheet. The data was pulled mid-December, I can update it if desired, it’s still illustrative, albeit stale.

The leftmost column is the affiliate address, the middle is the basis points charged, and the right is the aggregated trade volume for the given affiliate at the given fee.

Several things are apparent

  1. There's large volume at the 30bps level, across multiple affiliates, evidence users aren't adverse to paying in the region of ~30bps.
  2. We have done a fraction of the volume of other affiliates charging fees. Our zero fees does not seem to be attracting outsized volume, compared to competitors that charge fees.

Affiliates volume does not seem sensitive to fees charged, i.e. the same affiliate charging higher feels often do higher volumes. Note, I am not suggesting higher fees cause higher volumes. I am suggesting that demand for THORChain swaps is largely inelastic to reasonable fees charged. Econ 101 read on price elasticity for those interested https://en.wikipedia.org/wiki/Priceelasticity

  1. of_demand

While I don’t deny that qualitatively, adding fees will make it harder to acquire users, I think the way arguments have been framed as a dichotomy of “we can have users or fees but not both” (hyperbolic) is disingenuous.

I think these queries/data go a long way to helping quantify how user acquisition will be impacted, and indeed it does not seems to be an issue for numerous competitors.

Adding fees is also helps us cross the penny gap https://redeye.firstround.com/2007/03/thefirstpenny.html - are users willing to pay anything for your product? Note the crude charts in the article depict what the on chain data suggests - users will pay fees, not only that, they’re somewhat indifferent to fees charged, as long as we’re not price gouging (which the suggested fees in the proposal do not).

Monetizing our products not only provides crucial revenue for the DAO, but also helps close a feedback loop of where we can generate a return on our efforts.

Users will pay for value. We’ve created something of value. Users pay for the same value at competitors. We should charge for the value we provide. We are leaving money on the table.

This proposal is not a panacea for the DAOs cash flow troubles. It won’t make us profitable overnight. 30bps fees on our current, and thankfully, increasing volumes, would fix our cash flow problem by ~10% at the current burn.

If anyone has specific questions, I will try to provide quantitative answers with more on-chain data.

@0xdef1cafe thank you for posting this! It is exactly the type of data we should be paying attention to. I also 100% agree that this alone doesn’t solve our cash flow problems, but is a step towards changing the revenue equation while we collect critical data about where we should be investing more resources.

I was voting against playing with fees until reading this comment, I support playing with fees if we garantee the possibility of ZERO fees based on FOX/FOXy holding/burning

I really like this industry data that you pulled and am largely on the same page with you on this now–my biggest concern was that the fees added here on our end would be at the high end of what users are expecting to pay but it seems like from the synthesis you did, it seems like that isn’t the case. From a marketing perspective, I would love to know the protocols/players who are charging more than 30bps so we can look at how we might be able to acquire their users/assess what they are doing that makes them feel justified to charge that much (and also the people who are charging 30bps to just to a side by side “value” comparison).

Ideally, I would love to experiment with this in a lighter manner and maybe even offer users a button that allows them to opt into ‘donating’ 5bps, 10bps, 30bps, and 40 bps might be a good MVP to check out–or maybe finding some way to make fees more of a more feels-good/free experience.