The idea of a DAO - a Decentralized Autonomous Organization - rests on the belief that groups of people (an organization) can act (be autonomous) without relying on a single (centralized) person to make decisions. Sadly, the vast majority of DAOs struggle with this vital capability.
Simple DAOs with simple goals (“maximize returns on our fund investments”) have found ways to make decisions efficiently through proposing and voting. But almost any DAO with a complex goal (“build and run a platform” or “fund the best public goods”) has resorted to small groups of “leaders” making operational decisions and strategically crafting proposals. Why? And is this as good as we can do?
RnDAO seeks to deeply understand decentralised collaboration, and to empower it with new tools and systems. Our Decision Making Research unit kicked off with an exploratory study of decisions in DAOs as we know them today. We wanted to hear what DAO participants see as “good” decisions and “bad” decisions, and gather detailed stories of the process that led to them. Our goal was to find patterns and insights about what leads to good outcomes.
All but one of our participants expressed a desire to have DAO decisions be public and open - DAO leaders crave the wisdom of their crowd. And yet each of the “good decision” stories involved a small (1-3 person) team crafting a detailed proposal privately, selectively gathering input from influential DAO leaders, and only making the proposal public after passage was promised by known votes. In sum, stories of “good decisions” were always stories of closed decisions.
If those are the stories of good decisions, stories of bad decisions were where the process of proposing and deliberating was open and public. In our synthesis of the interviews [link to something cleaned up], we found three common pains that drive DAOs away from open debate.
When a decision is made in the open, there is a lot of interaction, which can also be a lot of noise, and thus this creates a lot of work. The work is needed to find the important signal in the noise of discussion and critique. The most concentrated burden falls on DAO leaders and/or proposers, who have to read every post, then consider, and then write in reply to at least some of them. Voters (DAO members with tokens) bear a work burden that isn’t quite as bad (they can be selective about topics and depth for reading and don’t need to respond), but is enough to push them to not vote on most topics, a pattern that looks like “voter apathy”.
A second motivation to avoid public debate on proposals is that it is invariably a slow process. With vague comments spread across the proposal, specific concerns aren’t resolved over time. Instead they build up in length and complexity
DAO leaders told us the final decisions about wording and direction had to happen in calls. In real time, you can make sure that the people you need are there, you can answer questions to everyone’s satisfaction, and the personal context leads to more polite behavior. All of that adds up to resolutions happening faster in synchronous calls.
Discussion in public forums causes anger within the community through three pathways. The first starts with two people arguing for different views, and then snowballing into factions as others get involved. While this is typical for the internet, DAO leaders see it as corrosive to productive work and collaboration.
The second pathway is that individual concerns are never visibly resolved. Even when an author changes a proposal to satisfy a commenter, those comments are left visible and people who share the concern assume it has either been ignored or is blocking progress.
The third pathway is that when a problem is important enough to motivate a significant proposal, there is a passionate base of DAO members who are frustrated by the status quo. Most of those members don’t have the bandwidth to join, or even read, the full debate. Without visible progress, it looks like the DAO is ignoring them and their pain.
In the end, these pains lead to one vital danger: the wise and talented people within your DAO won’t deal with wasted work and high conflict for long. Either they don’t engage with governance, or they leave. DAO’s depend on a voluntary community to drive their mission, and talent bleed is an existential threat. When given a choice between open debate and keeping talent, successful DAOs keep the talent.
Should DAOs adopt closed decision processes then? It could be that DAOs have simply rediscovered the same management model as corporations and would be better off using it.
It could have been, but the data told us otherwise.
Our participants (DAO leaders) defaulted to crafting their proposals or decisions in private, but sensed that the proposals were weaker because of it. One person had a specific story of getting feedback and buy-in from the important players and pushing a proposal through with a quick and clear vote. And then they were blindsided when a technical flaw wasn’t raised until the proposal passed. In trying to avoid conflict and solve problems quickly, we know that we lose out on the diverse knowledge that organizations intend to bring together.
At the same time that our participants explained why they kept decisions quiet until perfect, they also talked about times they felt locked out of important discussions. This was especially frustrating in decisions where the flaws felt obvious, but raising concerns publicly would lower one’s social standing. Why, they said, should they feel invested in a community where everyone was just going to do what the “founders” want anyhow?
Without open decisions, are we really acting as a “DAO”? At least one of our participants is keeping their tooling project a corporation, exactly because these issues aren’t solved.
DAOs and DAO leaders today are caught on the horns of a dilemma - do they have high-conflict open decisions or untrusted closed decision making? Either choice puts the viability of their DAO community at risk. Our hope at RnDAO is that this is a false trade-off, and that by finding the root causes we can solve this dilemma elegantly.
So, what causes the noise, the anger, and the delay in open debates?
The goal of open debate is to pull in the diversity of expertise and goals across the DAO. But most DAO members don’t have the technical expertise to read, much less critique, highly technical solutions. Engagement and participation is easier when people can communicate what they know, and all DAO members know most about the problems and frustrations they personally face. Giving input about your problems is easy.
But most proposals are about a specific solution to a vague problem, and most of the debate quickly dives into the technical issues with that approach. This is easy for those few people who have that technical expertise, but engaging is hard and frustrating work for the rest of DAO members.
People with a problem are a natural tribe, as are people passionate about a given solution. Those two tribes are natural allies when the context is framed in that way, but in today’s DAOs they end up in needless confusion and conflict.
The high work of conversational noise shows up having to read (or at least skim) ALL the conversations. There is very little signal about what is important - to write about, to read, to comment on- because only I know what my interests and focuses are. Individuals need to hunt for the conversations where their expertise is important or they care about the problem being solved.
DAO leaders and major token holders have to be ruthless about which proposals to even read. DAOs either create a formal filter (e.g. a centralized guild to review and approve proposals) or just ignore anything that doesn’t already have a minimum set of votes or comments. Both of these actions are opaque though. And there’s no mechanism for DAO leaders to commit to addressing a problem until there is a solution already at vote.
Debate today is scattered across systems (Discord, Forum, Telegram) and across channels within those systems. The nature of open debate is that topics come up serendipitously as you are discussing something else. Sometimes you have a long and detailed thought for a group, sometimes it’s a quick comment to a colleague. Some of the people you want to involve are in one place, some are in another. In the end, we discuss the same topic everywhere.
This adds work and time by making people hunt across channels. It also adds noise because people will raise a claim in one place not realizing it’s been responded to elsewhere. With no good way to see topics across systems, tracking the debate takes time while it is constantly accruing repetition and noise.
Proposers fear there is a hidden sea of opinion below the visible signs of a few comments. It's impossible to know if one vocal person represents a large class or if they are just enjoying the attention. In consequence, those invested in the proposals don’t want to close off specific threads for fear of offending a silent majority. This drives them to leave sub-debates open and unresolved until a vote is finally called and they can see the real sentiment.
In most tools (e.g. Discord, Discourse Forum), comments are applied to the entire proposal. Unlike Google Docs or Notion comments, you can’t select a specific area and create a thread about your concern. And that means that tracking questions is a burden on both the proposer and the voter.
The open structure also encourages vague and emotional objections, rather than concrete concerns that can be resolved in good faith. This may be one driver pushing people to draft proposals in Google Docs (a more closed system) and only afterwards moving them to forums.
In this project, we have found evidence that the “common knowledge” about DAO decision making doesn’t capture the true problems. In many ways, what people see as the causes of the problem are really the effects.
Low participation in voting, and low engagement with discussions, may cause governance problems, but members don’t participate because debate and voting is so broken.
Closed deliberation isn’t caused by power imbalances, rather the pains of open deliberation lead those with power to use it successfully.
Noise and conflict and delay show up in open debates even without conflicting interests.
If the common knowledge is weak, then these new theories should be tested rigorously. We can extend the qualitative research to uncover the best stories of successful open deliberation. We can test some hypotheses quantitatively across DAOs, and test others experimentally within individual DAOs. The value of good decision making is too high, and too critical to the DAO ecosystem, to leave these theories unexplored. Collaborate with RnDAO to extend this research!
We're actively working on solving this. If you're part of a DAO that could be interested, contact us via our DAO Deliberations telegram group to discuss and keep up with future research.