Increasing Voter Participation Rates in Genesis

papa_raw · May 27, 2019, 10:50am

unnamed

Currently, Genesis has an issue with proposals in the regular queue, that is, via absolute majority. As @parrachia has correctly noted:

Right now the only way for a proposal to pass is boosted, other wise it’s 30 day + 51% of all rep in favor (our record is 41% total)

At the request of @Kate I’ve been asked to offer my general commentary regarding it, and posit some solutions:

First, the “Why is this happening?”

Unlike other DAOs, Genesis, as an alpha experiment, was instantiated with a fairly arbitrary Reputation distribution and has added members very loosely. With this “loose policy,” many members simply do not participate after they submit their first proposal as they don’t have any incentive to do so: extrinsically, there isn’t enough capital to generate activity; i.e. Genesis cannot administer grants large enough for full or part-time attention; intrinsically, there’s just not enough interest to sustain prolonged engagement; DAOs are still a novelty and the Genesis Alpha as a virtual tester community does not have so much emotional punch as to sustain activity with the majority of new joiners.

Holographic consensus anticipates this issue and offers, through the GEN prediction mechanism, the ability for people to “force” a vote to occur on any given proposal. This is the mechanism the Genesis Alpha has been using for the most part, and it’s been working: within the Genesis .2 protocol there’s an extrinsic incentive to boost proposals as one can profit from the DAOstake. With this, it could be argued that there’s no need to be able to pass proposals within the regular queue, but in my opinion I disagree with this logic: not being able to reach a global majority when needed is a bug, not a feature.

I am sure the above reasoning is by no means exhaustive, but it offers a high level overview of the issue.

Second, the solutions:

There are arguably five solutions that will increase the likelihood of proposals passing in the regular queue:

Remove inactive voters: This could be accomplished through social consensus; i.e. a policy can be formalized wherein inactive voters are slashed of their Rep after x amount of time. This type of altruistic punishment has been demonstrated to work in order to maintain the quality/activity of a process – in our case, voting. This would be ultimately a manual implementation of a process that can be made automatic through some sort of Reputation decay function; however, Arc needs to be upgraded to facilitate off-chain voting in order to automate decay, this is not expected until Q1 2020.
Massively inflate the Rep pool towards active voters: Similar to (1), this simply adjusts the balance of Reputation towards those voters who are actually voting, thus increasing the likelihood of a global majority forming. It is, in a sense, a tax on inactive voters.
Fractalize to a lower n-stakeholder count with a bloc voting model: There’s an ongoing discussion around the merits of fractalization here, and I present a bloc voting model here that could be utilized to increase Reputation voter % in the regular queue. The jist of this is that Reputation would be allocated by voting blocs, and not individual stakeholders, and those blocs would be orgnanized into their own DAOs; eg. DAOstack Technologies could be a single voting entity with 20% of the Rep, instead of 20 separate stakeholders. @adam has confirmed that DAOs will have the ability to vote in each other within Alchemy as soon as the Generic Action scheme is released; this is expected in the next month.
Implement single or multiple layer delegation (aka Liquid Democracy): Both single and multiple layer delegation would be expected to increase Reputation thorough-put, assuming that Reputation is delegated to active voters. This could be a great project/module for an open-source developer to tackle.
Lower the threshold for absolute majority: One of the key distuinguishing features of DAOstack is the ability for DAOs to upgrade themselves, this is accomplished through the Scheme Registrar, a scheme that will be made available within the next two-ish weeks. With this, we will have the ability to update/change the Genesis parameters; one parameter is the Majority Parameter, that is, the required % to pass a proposal through the regular queue. By lowering this from 50% to some arbitrary number, say, 30%, it immediately becomes easier to pass proposals within the regular queue – the tradeoff being that the results may not, in fact, reflect the global majority.

It goes without saying there are numerous marketing technical solutions as well that can provide indirect solutions for this issue by channeling increased voter attention through owned channels: e-mail and push notifications, the ability to subscribe to DAO updates, Telegram and Twitter bots, etc… these should not be ignored, but they do not directly impact the issue itself (you can receive 100 notifications for a DAO and yet not vote in it at all).

Kate · May 27, 2019, 12:51pm

Thanks a lot Pat - this is great and so necessary for our conversation on how to unblock proposal making and proposal passing in the DAO - in my view a key challenge in right now

I’m evaluating these steps with a bias to action in terms of solving the problem in Genesis right now:

As you’ve mentioned before (1) seems to be in opposition to one of Genesis’s goals - which is to increase membership - thoughts? Also does the Arc upgrade mean this is not an option?

(2) Adjusting rep in light of Genesis Beta being launched in the next few months seems unadvisable as deeply considering and potentially adjusting the rep pool is part of this work.

(3) IMO It seems too early in the experiment to try this - this seems like it would necessary in a bigger DAO but putting the horse before the cart to try this as a solution to our problem

(4) What needs to happen for an open source dev to be able to tackle this?

(5) This seems to be an easy way to make an adjustment in the lead up to Genesis Beta once the line once the scheme registrar is ready to go - is there a way we can do a calculation to get to a less arbitrary number?

Further: the original 50 % parameter is fine if all members are relatively active but in collaborative decision making systems I’ve been involved with like Loomio at Enspiral (this is a totally different context I am aware) full participation or even 50 percent turnout is rare (same for most elections) and so we don’t even try get 50 percent quorum for ‘standard decisions’ (this is saved for significant decisions). Obviously Holographic Consensus is designed to address this but we’re not quite in the scale for this to be fully functional right now https://handbook.enspiral.com/agreements/decisions.html

Would be keen to explore the implications of lowering the turnout required and wondering if can we come up with a less arbitrary number (calling mathematicians)

AdrienDLT · May 30, 2019, 10:42am

Thank you Pat for bringing this issue to our attention. Disclaimer: I realize I’m still getting to use to the Holographic Consensus so their may be aspects that I am missing.

I will tackle the points you brought up but first of all I’d like to add a couple of comments. It’s slightly off topic but my guess is that the issue we’re tackling here isn’t just “voter participation rates” but how to also get regular proposals to pass (not “just” boosted proposals) & get pollinators excited about the “regular proposal” process too. My comments are regarding the “regular proposal” lengthy process:
a/ 30 days is a very long time, especially if it’s for a proposal that has request a low budget (i.e. less than 2% of funds held in the DAO)
b/ 30 days is a very long time, one has done a preliminarily effort on a draft proposal (before submitting thesaid proposal) as this can take days/weeks. As a pollinator, I don’t like the idea of submitting a proposal that might not pass so that preliminarily effort is significant.

Addressing your points:
1/ Remove inactive voters: tldr: I am against this. Let me explain: I call the initial proposal where one gets rewarded REP (reputation) the “Registration Proposal” when helping onboarding a new contributor (pollinator in Genesis DAO lingo). Following that logic, I compare this to signing up when joining Reddit/Stackoverflow/Imgur or any other successful platform. Would any of these platform unregister me if I’m inactive for x months/years? No. Simply because they know you may come back at any time & having already done the sign up process removes a big pain point (which is much bigger in the Genesis DAO as it takes days/weeks to get that first proposal passed).
2/ Massively inflate the Rep pool towards active voters: tldr: it depends. There are so many ways to implement that idea & it’s a hard problem. We’d want to make sure that we don’t lose potentially very strong contributors just because they missed the early days of the DAO & have a voting power that’s extremely low compared to other participants. That’s a typical problem that have all the online communities I’m part of & I haven’t seen any community solving this so far. Thinking out loud: a solution for this may be to give extra reputation to contributors that are extremely active within a given timeframe (i.e. the “contributors of the month” get his REP earned in the last 30 days multiplied by 2).
3/ Fractalize to a lower n-stakeholder count with a bloc voting model: tldr: nice, but prob not enough. I like the idea of D2D, the bloc voting model is very interesting. The main issue I see is that making a DAO rely on another DAO’s votation, adding strong dependency on another DAO will add a lot of complexity. It may be a very good thing in some cases (i.e. when you have a ParentDAO --> ChildDAO relationship) but should it be the answer to the voter participation rate? I don’t think so.
4/ Implement single or multiple layer delegation (aka Liquid Democracy): tldr: maybe yes. For all I have read about this idea, I’m worried that one manages to accumulate too many votes. I’d suggest to experiment first with a non-multilayer & limited delegation system (i.e. one can only be delegated up to 3 contributors’ voting power).
5/ Lower the threshold for absolute majority: tldr: maybe for a temporary solution. The Genesis DAO (in Alpha right now) is the first DAO using the Holographic consensus so we may get used to the idea that we’ll need to adjust that setting until we find the correct one. We might observe recurring patterns with other DAOs being launched with the Holographic consensus & get very good at adjusting that setting according to the phase any given DAO is into (i.e. brand new vs 3 years old with 1000+ contributors).

Following Kate’s comment:

solving the problem in Genesis right now

That’s a very good point & only see point 5 as a possible solution.

nylon · May 30, 2019, 3:35pm

Thanks, Pat, I think this is a critical issue that should be addressed soon.

In my opinion, reputation decay is the only long term viable solution.
@AdrienDLT touched on a social media platform unregistering you, I don’t think that is the proposition here, you are not removing someone from participating, you are, over time of inactivity, taking away some of their Rep and influence in the platform.

If your account hasn’t been active for a long time your reputation would be decreased but not completely removed.

From Pat’s link of the colony WP:

All reputation decays over time. Every 600000 blocks, a user’s reputation in every domain or skill decays by a factor of 2. This decay occurs every hour, rather than being a step change every three months to ensure there are minimal incentives to earn reputation at any particular time.

I think this a great approach but we might want to change the parameters to fit Alchemy better. Maybe a few days or a full week? also, be a lower decay of 1-2% every said period of time.

Reputation should be dynamic in both ways. You can get reputation with participation and contribution but, If you’ve stopped contributing/participating then your rep should decay slowly. A DAO is a dynamic and changing organization and so should the (top) rep holders change according to their recent (as defined by the DAO) contribution.

Also in #2 you wrote:

We’d want to make sure that we don’t lose potentially very strong contributors just because they missed the early days of the DAO & have a voting power that’s extremely low compared to other participants.

What about the other way, person X contributed a lot to the DAO 2 years ago, but he’s been inactive for the past year, yet still holds a substantial amount of reputation. Now he is not familiar with the "Current State of the DAO™️ and maybe even the goals have changed. In a sense, the reputation that he has received is now irrelevant to the DAO. And so his voting power should be diminished.
I think that to start as a newcomer and know that reputation doesn’t change is discouraging.

A well defined and slow reputation decay will also incentivize people to stay and participate, incentivize newcomers to contribute, and solve the issue of reputation participation. While allowing a collective consciousness whcih is relevant.

That being said I understand that enacting a reputation decay is impossible right now, and I agree that the only available right now solution is probably #5, but it’s a temporary solution which doesn’t really solve the problem.

Another suggestion is to make a manual proposal manual for reputation decay. For example a 2% rep decay proposal every month. Not sure if this is even possible?

jer979 · June 3, 2019, 2:08pm

Thanks for this, Pat.

Trying to get my head around everything, but it sounds like #4 above is kind of like a Delegated Proof-of-Stake model, but for reputation. I’m sure there are unintended consequences I am not considering, but at a high level, there’s something cool to me about this. It would give people an opportunity to increase their reputation by convincing/demonstrating to others that they vote in the best interests of the DAO.

Then, they could share GEN rewards with their “REP backers”. Perhaps the REP decay/slash could still happen, but as long as someone “renewed their vote/backing” of a delegate w/in a specific timeframe (tbd by DAO, I suppose), they would keep it (or increase?)

philh · October 5, 2019, 4:17pm

Thanks for this! The fact that no non-boosted GenDAO proposal has ever passed was one of the things that bothers me a lot with Alchemy, and I’m super happy to realize that it has already been spotted and discussed by you and several members of the DAO. Thanks also for the creative suggestions to fix the issue.

Do you happen to know whether other Alchemy-powered DAOs have the same issue, and if it has been discussed elsewhere?

I pretty much agree with @nylon’s comment regarding the fact that reputation should be more dynamic and reflect the actual activity. I understand that it might be covered in a future version of Alchemy, is there any news about that?

The solution #5 seemed to be an easy fix. I’m wondering why it hasn’t been tested. Most people seem to agree with your view that the systematic failure of non-boosted proposals is a bug rather than a feature. I wonder what it would take to actually change the reputation threshold to, say, 25%, and see what happens. I can’t see how it could be a problem, but I may miss something. But how could it be tested? Is it an option to actually make a proposal about it?

@AdrienDLT also raised the issue of the 30 days period for non-boosted proposals. Do you know the reason for it, and whether it has ever be discussed since the start of GenDAO?

Thanks!

wkarshat · October 6, 2019, 10:14am

A noob to DAOstack and Genesis, I still have to ask why growing the total number of “registrations” matters, if a large majority end-up not participating.

Back in the day, the ‘growth’ fellow at some eCommerce company would never delete accounts that could be variously labeled as fake, duplicate, abandoned, etc. The number of new and total accounts were reported monthly to the VC as proof that the site was still inching toward the near vertical part of the “hockey stick”. Measures that one can brag about, yet have no direct correlation with the business became known as ‘vanity metrics’ and are frowned upon in the Lean startup community.

The game of vanity continues in the industry, where Apple will no longer report the total number of phones sold per quarter (as it is now declining) but instead brags about “over a billion ‘i’ devices activated worldwide”. Facebook reports 1.44B accounts, supposedly only including the ones that logged in within 30 days. I suspect that they count all the mobile device and laptop app instances that communicated with their servers at least once, even if the user hasn’t opened the app or interacted with the service. Facebook may be incentivized to mislead their advertisers and the business media and their readership, but does that logic apply here?

I think it is devastating in case of DAOs, which are specifically “organizations” aspiring to socially reach and affect “communities”. Maybe the mindset of tuning ‘funnels’ and tweak ‘cart abandonment’ works for eCommerce, but anytime participants get a sense that they are just numbers, all the allure and promise of DAOs is done and over with.

Specifically with the Alchemy REP arithmetic, if over half of the signups have been variously abandoned, it directly affects those joining afterwards. Having just joined I received 0.33% and the average for 229 accounts is 0.44%, so it appears equitable. However, those abandoned and otherwise inactive accounts that jointed earlier have locked-up perhaps 40% of the REP issued at that point. As a result my % is 6.4 times lower than the maximum, where it could have been perhaps 4 times. Large fraction of REP having been abandoned prior to one joining multiplies the advantage of earlier adopters, beyond what was intended. No inflation proportional to the REP already held nor deflating inactive accounts gradually, while both fine ideas, will change this situation, as it is set based on the order of initial allocation.

I’ve observed that new “applications” are routinely approved by as few as 2 votes. Even high visibility proposals, like #rekt and ‘real names’ have received respectively 16 and 7 votes so far. I guess that perhaps 10% of 229 are active enough to even notice pending proposals.

Current outlook appears to be that the participants have two duties - to field proposals and to vote on them. My experience is that recruiting, filtering, approving and supporting new participants is by far the most important task in the long term. Now the “application” by a prospective participant is structured as just another proposal, and one with no fungible value and perceived as having low impact. Reviewing those have been relegated to a couple of dedicated founding members, signaling to the active participants that voting is largely optional and that dealing with membership matters even less.

I am not prepared to offer what should be done, particularly in light of current state of participation, but would suggest tracking logins and voting as indicators of availability and activity respectively. I’d like to see a third option (Abstain) to the Yes/No vote, to capture the Recuse, Undecided, Postpone opinions, so a proposal could be allowed to fail without signaling ardent opposition.

papa_raw · October 6, 2019, 6:48pm

This isn’t the long-term plan by any means. I’m sure different folks have different reasons for how we initially ended up with this; for me, at least, I think the registration process is/was appropriate for an alpha product. At the time Genesis was launched (June 2018) joining a DAO was an extremely novel process. So in a sense, it was a “user acquisition” process to expose folks to the alpha, and DAOs in general. But I believe you’re correct: now it’s not super helpful, and as we transition to GenDAO1.0 it should be phased out; perhaps an “OnboardingDAO” has some potential for new Alchemy users.

I’m sort of skeptical this could work with the current Gen.2 voting machine. It would only be relevant to unboosted proposals, and I doubt users would want to pay gas to abstain. But as a step towards designing alternative voting machines, this is another parameter to consider.

papa_raw · October 6, 2019, 6:59pm

I think the issue with #5 is that if we made an actual effort to activate the entire DAO’s attention (e.g. there was a highly contentious, high budget proposal, with significant stake on both sides, and everyone is canvassing), there could be a potential global consensus achieved. If we view a global majority as the yardstick for fairness, it is unjust to apply #5, and to me, seems like a last resort.

Since the time I wrote my post I’ve been giving greater consideration to the idea that this isn’t necessarily a bug. I think there’s an argument to be made in light of Moloch-style DAOs that possess nothing but relative majority (and struggle with the same absolute majority requirements) that staking is a valuable and necessary resilience mechanism, and an improvement overall. I think there’s an argument to be explored for this non-global modus operandi we’ve implicitly adopted.

wkarshat · October 28, 2019, 11:25am

As an update, the #rekt proposal garnered a total of 16 votes and 14% REP, the ‘real name’ seems stable at 19 votes and 15.6% REP and Daniel’s onboarding back in April saw 21 votes and 21.3% REP. And optimistic number of ‘active’ accounts is under 15% of the total, now at 238 [minus 3+ without REP], though they could command closer to 25% of total REP.
This is important for the “absolute” majority to be even a theoretical possibility in a very visible and important vote… The proposal to lower the threshold from 50% to 30% will not accomplish this.