A viable alternative to percentage-based transfer fees?

[Thom]

Having the same fee for transferring both 1 USD and 10,000 USD is absurd - it is obvious that for the user these two transfers have extremely different levels of usefulness.

Saying it's absurd is a bit strong, don't you think?

It costs the network no more to transfer 1 BTS than it does 1,000,000 and that holds true for all assets.

However, the price of an item is not only related to what it costs to produce it. It raises the issue of what profit margin is reasonable, or better stated, what profit margin can be obtained from the market. It's a matter of demand, and right now demand is low on all fronts of this ecosystem.

In general I think there is merit in a 2 tiered approach as you describe here, and basing it on the asset itself rather than in terms of the equivalent BTS value is an excellent way to simplify the determination of which fee rate the transaction falls within.

This same approach might also work for other assets transfers, but b/c of how UIAs are created the approach isn't necessarily as clean. Even if it was allowed for UIAs, even to the extent of allowing the threshhold to be set by the UIA creator, the effort required to implement it may not be worth the benefit to the ecosystem. For bitAssets no GUI effort is required. The threshhold value could be just another set of parameters set by the committee.

I thinks it's best to try this out on bitAssets first before implementing it elsewhere.

[abit]

Imo 10 BTS is still too high for a common Chinese user.

10 BTS transfer fee for payments below 10 bitCNY is still a significant reduction over the current 30 BTS.
It cannot be too low because it will encourage people to game the system by splitting their payments into several smaller parts.

It's a bit high for individuals, however maybe merchants don't think it's high?

[Empirical1.2]

Edit: There may be some value in an intermediate tier.

I think 3 tiers is the maximum. I'd start with 2 and see how it goes.
Having more tiers will make it too complex for the user and it will also weaken the referral program too much.
(or maybe we should introduce 3 tiers but make the transfer fee significantly higher than 30 BTS for the last tier - this would compensate the referral program businesses)

But I think we definitely need it.
Having the same fee for transferring both 1 USD and 10,000 USD is absurd - it is obvious that for the user these two transfers have extremely different levels of usefulness.
As we have it now, we either overprice people for small payments or forgo the revenue we could potentially have for bigger payments - it's just stupid business-wise.

   I agree. I also agree 3 may be the maximum for a tiered approach & that the last tier could be higher that 30 BTS.

[jakub]

Edit: There may be some value in an intermediate tier.

I think 3 tiers is the maximum. I'd start with 2 and see how it goes.
Having more tiers will make it too complex for the user and it will also weaken the referral program too much.
(or maybe we should introduce 3 tiers but make the transfer fee significantly higher than 30 BTS for the last tier - this would compensate the referral program businesses)

But I think we definitely need it.
Having the same fee for transferring both 1 USD and 10,000 USD is absurd - it is obvious that for the user these two transfers have extremely different levels of usefulness.
As we have it now, we either overprice people for small payments or forgo the revenue we could potentially have for bigger payments - it's just stupid business-wise.

[Empirical1.2]

Imo 10 BTS is still too high for a common Chinese user.

10 BTS transfer fee for payments below 10 bitCNY is still a significant reduction over the current 30 BTS.
It cannot be too low because it will encourage people to game the system by splitting their payments into several smaller parts.

That might not be a bad thing. Those who don't care about $0.1, which according to BM is most people, wouldn't bother splitting their payments, whereas those for who a few cents mattered could split their slightly larger payments into smaller sizes.

Edit: There may be some value in an intermediate tier.

It seems a TX fee of $0.05 would be more appropriate in the sub $30 range too. (This would still be 0.2%)

Having said that $0.2 USD transfer fees make some project unviable.

Im pushing Bitshares for a joint effort of two NGOs and the local crypto community which one the objectives is to provide entrepreneurs from marginal economies with a way to use digital cash. (Its a bigger project ex-Worldbank, ex-UN people working on it)
 
Basically they will be moving an IOU, but $3.2 ARS is too much for the average size of that transactions. ($30-$50)

are we competing with "every single other coin" or the legacy systems?

I think one of our main competitors is uphold (BitReserve), who offer centralized BitAssets, they are supposedly the fastest growing money platform in the world & they are 'free' for any amount. https://uphold.com/

I'm not suggesting we should/can be free, but we do have stiff competition among customers that don't value decentralization.

[jakub]

Imo 10 BTS is still too high for a common Chinese user.

10 BTS transfer fee for payments below 10 bitCNY is still a significant reduction over the current 30 BTS.
It cannot be too low because it will encourage people to game the system by splitting their payments into several smaller parts.

[abit]

I like that your solution attempts to balance the needs of the referral programme with the need to be able facilitate, if not micro, at least smaller size transactions, like for example tipping. It is also simpler to implement that percentage based fees. Definitely worth considering.

I'm sure we can find some common ground and arrive at a simple solution that:
- respects both tipping/micro-payments and the referral program
- is simple for the users to remember and the devs to implement
- prevents spamming

The solution I proposed also enables some pseudo-regional customization, i.e. we could set the fee thresholds like this:
- for bitCNY: 10 bitCNY (~1.5 USD)
- for bitEUR: 1 bitEUR (~1.1 USD)
- for bitUSD: 1 bitUSD
And this way the Chinese users will get a bit of preferential treatment but without any real temptation to game the system.

Imo 10 BTS is still too high for a common Chinese user.

[jakub]

I like that your solution attempts to balance the needs of the referral programme with the need to be able facilitate, if not micro, at least smaller size transactions, like for example tipping. It is also simpler to implement that percentage based fees. Definitely worth considering.

I'm sure we can find some common ground and arrive at a simple solution that:
- respects both tipping/micro-payments and the referral program
- is simple for the users to remember and the devs to implement
- prevents spamming

The solution I proposed also enables some pseudo-regional customization, i.e. we could set the fee thresholds like this:
- for bitCNY: 10 bitCNY (~1.5 USD)
- for bitEUR: 1 bitEUR (~1.1 USD)
- for bitUSD: 1 bitUSD
And this way the Chinese users will get a bit of preferential treatment but without any real temptation to game the system.

[Empirical1.2]

I like that your solution attempts to balance the needs of the referral programme with the need to be able facilitate, if not micro, at least smaller size transactions, like for example tipping. It is also simpler to implement that percentage based fees. Definitely worth considering.

[jakub]

Percentage based transfer fees are completely untenable. Every single other coin in the world has a low, fixed fee for any amount sent, bitshares would lose out to the competition and it would be a PR nightmare. Percentage based trade fees on the other hand, are acceptable.

But are we competing with "every single other coin" or the legacy systems?

Anyway, if you read the OP I'm not advocating percentage-based transfer fees (because it's too hard to implement).
What I propose is a two-tier transfer fee schedule to address the problem of micro-payments.
If we don't do micro-payments somebody else will.

[abit]

Percentage based transfer fees are completely untenable. Every single other coin in the world has a low, fixed fee for any amount sent, bitshares would lose out to the competition and it would be a PR nightmare. Percentage based trade fees on the other hand, are acceptable.

If the fee is charged by IOU/Smart coin issuers then would be OK (in addition to basic network fee). Best if the issuers are able to define several thresholds and % of the fee. Under this condition the committee can control fee of transferring BTS to be low or higher (as demand). Maybe the ones who benefit by the reference program won't agree though.

[monsterer]

Percentage based transfer fees are completely untenable. Every single other coin in the world has a low, fixed fee for any amount sent, bitshares would lose out to the competition and it would be a PR nightmare. Percentage based trade fees on the other hand, are acceptable.

[jakub]

In the last Mumble @bytemaster mentioned that percentage-based transfer fees are a complex subject, mainly because of these two reasons:
- it's hard to reliably determine the exact BTS value of an asset at any given point of time
- for stealth transfers the amount is not available

The current transfer fee policy (30 BTS or $0.10 per transfer no matter what the amount is) effectively kills businesses based on small tips.

So if percentage-based fees are such a headache, why don't we apply a simple tier-based solution:
For each SmartCoin let's introduce a threshold expressed not in BTS value but in terms of the asset itself, so that e.g.:
- if you transfer less than 1 bitUSD you pay the equivalent of 10 BTS
- if you transfer more than 1 bitUSD you pay the equivalent of 30 BTS
And this would only be applied to SmartCoins and normal transfers (i.e. stealth mode excluded).

It will not be perfect but relatively straight-forward and easy (?) to implement.
And it will address most of the problem without hurting the referral program in any significant way.
But most importantly, it will open up huge opportunities for businesses relying on micro-payments.

Is there any flaw in this approach?

EDIT: I think I've come up with a better idea described this new thread:
https://bitsharestalk.org/index.php/topic,20789.0.html