[Discussion] Safe DAO Resource Allocation Model (OBRA)

Title: Safe DAO Resource Allocation Model (OBRA)
Status: Draft SEP, open for feedback and discussion
Author: Peter ‘pet3rpan’ (1kx), Justin ‘AccelXR’ (1kx)
Date: 22nd Thursday, September 2022

// This is a draft proposal that we are seeing feedback and looking to iterate upon

About:

Decentralized governance / DAOs established a vision for the possibility of decentralized, neutral public goods owned by the public. However, after 3+ years of live data emerging around decentralized networks, decentralized coordination is still a work in progress.

During this period, 1kx has supported, invested, and participated in over 40+ DAOs and incubated/launched many ourselves. Through our experience, the case for decentralized resource allocation has a long way to go with finding a long-term viable model, with most DAOs experiencing major coordination failures such as the following:

  • Unclear consensus and clarity on core DAO activities and product strategy, leading to excessive spending and misallocation of community resources to misaligned initiatives.

  • Lack of oversight, feedback loops, and accountability of community initiatives leading to overhiring and working groups with unreasonably high burn rates and low output efficiency.

  • Inadequate contributor quality curation and accountability leading to a high amount of incompetent, low context individuals gaining senior leadership positions, leading to a further degradation of the community’s operating standard and contributor value extraction at the cost of the entire community.

  • Lack of data driven of decision making processes that lead to the governance capture of the community to contributors who brute force governance via beuracractic lobbying.

  • Bloating on the DAO’s decisional surface area to a point where collective governance contributors no longer have the adequate context to govern effectively, leading to over delegation and concentration of social capital to a few select voices who are time rich but not nessearily those creating real impact for the DAO.

Rather than trying to address these problems incrementally, 1kx looks to propose a surface area reduced model for the governance of Safe DAO’s resource allocation known as ‘outcome-based resource allocation’ (OBRA).

Outcomes-based resource allocation (‘OBRA’)

The OBRA model minimizes the focus on the governance of resource allocation to focus only on the following areas:

1. Goals & metrics:

  • What are the vision, goals, and north star of Safe DAO?
    • eg. ‘To become the asset management standard of web3’
  • What metrics does the DAO use to measure progress towards its north star?
    eg. ‘# of multi-sigs deployed, TVL of multi-sigs, volume transacted via multi-sigs, revenue/token buy back rate, multi-sig market share"

2. Strategy:

  • What are the different approaches to achieving the DAO’s agreed-upon goals?

    • Eg. ‘Focus on building more third-party safe modules to drive safe transaction volume, which would lead to more utility and thus greater TVL’
    • Eg. ‘Focus on institutional crypto adoption to drive TVL’
  • We expect there to be three main categories of strategies:

    • Adoption (focused on gaining greater adoption and market share)
    • Value accrual (focused on accruing value to Safe DAO)
    • Risk (focused on addressing and managing risk for Safe DAO)
  • Example of a strategy proposal template/framework:

    • Goals and metrics the strategy looks to drive progress in
    • Outline the execution strategy or thesis
    • Detail around existing data or evidence to support this thesis
    • Detail around any risks associated with this strategy
    • If this strategy succeeds, what is the happy case?
    • If this strategy fails, what would be some reasons?
    • What are some example initiatives that would fall under this strategy
    • Assessment of the strategy’s maturity and if there is additional data that needs to be collected for the validation of this strategy
  • We expect there to be many valid strategies that the DAO can adopt to achieve its goals. We should both carefully assess the validity of strategies while also focusing on experimenting and validating new strategies.

  • The maturity of each strategy and progress being made on each front should determine the level of resources being allocated by Safe DAO towards initiatives under that particular strategy

3. Initiatives

  • Initiatives are work proposals that are funded by Safe DAO
  • Initatives are more traditionally thought of as more traditional ‘DAO proposals’
  • Example of an initiative proposal template/framework
    • Which pre-approved strategy is this initiative driving forward?
    • Which metrics and KPIs will the initiative be measured against?
    • Who is the accountable initiative lead? (individual or organization)
    • What is the initiative about?
    • What resources are being requested from Safe DAO?
    • What risks does the initiative entail?
    • Timeline/roadmap of milestones for the initiative
    • Are there any upfront funding this initiative needs at the beginning
    • The requested $SAFE success reward if the team achieves all work milestones (vested token bonus for completion of work)
  • Operating procedures around initiatives:
    • All initiatives must be driving forward an existing DAO strategy
    • All initiatives are funded by streamed payments via a tool such as superfluid or sablier unless the sum of funds is < $10,000
    • The proposer may request an initial lump sum to kickstart the initiative, but the lump sum may not exceed more than 20% of the total funding requested for the project unless the project timeline requires less than 3 month to complete
    • The $SAFE success reward for the completion of initiatives will be locked up for 1 year, followed by a 1 year linear unlock via vesting contract

4. Review:

  • Safe DAO will operate on seasons lasting 4 months each consisting of a work period and a review period:
    • Initial 14 weeks of operational focus
    • Followed by a 2 week review period in which the funded initiatives of Safe DAO will be reviewed by token holders and delegates
  • During the review period, all initiatives with streams of funding are subject to review and be required to share progress updates - they however do not need to re-apply for funding and drive a proposal through governance if their pre-approved funding extends beyond the prior work period
  • The DAO will also review existing approve strategies and may also decide to remove supported strategies too
  • Any new strategies and initiatives for the DAO must be proposed before this period to be considered for funding or approval
  • If token holders and delegates wish to terminate an initiative, the rest of their approved funding are returned to the DAO (all unspent funds and unvested tokens)

We expect the first review cycles for Safe DAO to take a bit longer in practical reality but, over time, get to a better place of operational predictability. For the first review cycle that emerges of Safe DAO, we want to take extraordinary care to walk through the process carefully to educate all community members involved about the process and invest time with involving delegates.

Voting thresholds for governing strategy, initiatives, and review

We suggest there be minimum thresholds in governance power for being able to vote on the DAO’s core strategies, initiatives, and review cycles as a means of encouraging coordinated delegation parties to emerge in the governance process while also discouraging low effort and low signal participation. We suggest the following parameters:

  • Governance on strategy: minimum 1% of circulating supply
  • Governance on initiatives: minimum 1% of circulating supply
  • Governance on review: minimum 1% of circulating supply

Incentives for reviewing initiatives

Once the first review cycle emerges, there will be an open question on how to properly incentivize the review of this work, as this becomes an operational burden on token holders and delegates. We can expect professional delegates that take a delegate management fee on tokens to emerge as a result of such incentives. MakerDAO currently has a model for rewarding delegates (MIP-61), and we can expect a similar design to emerge here eventually.

Why no working groups?

Typically working groups within DAOs are funded on an ongoing basis, with very little oversight where many critical decisions around strategy left for contributors to decide. By first defining strategies as a DAO via token holder governance, it enables clarity of operating strategy and direction for contributors to focus on contributing as opposed to overly involving themselves in governance and other areas of their core competency. Initiatives are then funded based on pre-approved strategies and curated based on their ability to efficiently produce metris driven results for the DAO.

As such, with Safe DAO, there are no ‘default’ working groups. While working groups may form and request resources to drive initiatives forward for Safe DAO, there are no ‘official’ working groups to which the DAO will default responsibilities towards. Any arbitrary individual or group will be able to apply to contribute towards the DAO via initiatives.

We instead look to foster many groups contributing to Safe DAO, some cross-functional, others specialized in certain functions. By assessing contributors based on metrics/KPIs and fostering a free market of working service provider options, the DAO is able to assess and reward its contributors in a manner that is politically minimized and free of governance capture from any single contributor group.

The goal here is to open up a free market for teams to offer bids for work in relation to the DAO as opposed to many of the other DAOs that ‘officiate’ working groups and their working group leaders into permanent positions purely based on being ‘early’ or having social capital in the community.

While this model might introduce sub-optimal resource allocation towards initiatives, it is a design that minimizes the political friction of Safe DAO. We can think of the current status quo of working groups as highly inefficient government-subsidized departments and the initiative-based model as a free market-based model that encourages the formation of service providers for the state that operates under the pressure of competition and a free market.

While a working group model for earlier stage DAOs might fit better due to the need to be more iterative with feedback loops and strategy, Safe DAO has product market fit and a natural monopoly. At a smaller scale, you can operate with greater uncertainty without the overhead of politics, but as you scale, the political environment is a variable you want to minimize.

In every case, we’ve seen an over-reliance on working group-based models suffer similar fates to centrally governed economies which we’ve historically seen to fail.

In this outcomes-based governance model, we look to place a heavier emphasis on debating strategies first before deploying resources and having metric-driven feedback review cycles on the DAO’s allocation of resources. reviewing the DAO’s work.

Slow is smooth, smooth is fast.

How do we retain crucial functions for the DAO without working groups?

There may be areas of critical function for the DAO that may need ongoing functions, such as smart contract security or full-stack security audits to which the DAO should look to create a risk strategy to which various service providers can bid for the ongoing responsibility required to be filled by the DAO. The difference here is that there will continue to be regular review cycles for service providers retained by the DAO and that there might even be several service providers that are engaged to support the same very initiative. We expect risk-based initiatives to be much longer than and steadier supported by the DAO as opposed to other initiatives surrounding adoption or value accrual, as these areas are far more dynamic than risk management.

Design principles

  • Limit the scope of the DAO’s scope of governance and push complexity to the edges
  • Establish a resource allocation process that limits politics and rewards merit/data
  • Allow for bottoms up open source resource allocation and opportunities to contribute, but only if there is pre-existing clarity and due diligence around strategy
  • Keep it simple stupid (K.I.S.S)

The evolution of Safe’s operating model

This operating model looks to soften the landing of Safe from a centrally operated organization into an open-source DAO. We expect this operating model to evolve as the core organization currently driving forward Safe gradually decentralizes over time. For the time being, this operating model simply looks to accommodate community-driven resource allocation toward grassroots initiatives.

Non-resource allocation-related proposals

There will be many other proposals that are non-resource allocation related, such as smart contract parameters or token list upgrades, and we expect those to have their own set of processes governed as needed by the rest of the DAO. This operating model mostly has looked to zero in on resource allocation, which has been mostly been the number one point of failure in other DAOs.

Governance references that inform our point of view:

Our goal here is not to refute the tremendous contributions these DAOs have made to the space for their experiments are necessary to understand the viable design space of coordination better but instead to acknowledge the risks involved and seek an alternative structure that accounts for these common pitfalls.

Next steps

  • Feedback from community members on this operating model
  • Ratify a revised version of this operating model as a DAO via a community vote
  • If the community ratifies the operating model:
    • Propose a set of metrics and KPIs for Safe DAO as a community proposal
    • Propose a governance framework for developing strategies (a way to propose new strategies and a process to validate them)
    • Propose a governance framework for proposing initiatives (that drive forward approved strategies)

====

This initial draft here aims to open up a conversation and gather feedback.

We’ll be iterating on the design and execution of this model per community discussion.

ps. Please fork these ideas and apply elsewhere as needed.

8 Likes

Attaching diagram:

// Was only allowed to post a single embedded media in the forum post

1 Like

This will likely be a canonical post referred to by metagovernance geeks like myself for at least months if not years to come. Thank you for writing it.

Why didn’t you make a higher threshold I wonder for Governance on strategy? Wouldn’t we want something “constitutional” to have a higher threshold? We wouldn’t want the Goals and strategies changing as frequently or easily as individual initiatives / proposals would we?

I know this post is focused on OBRA as opposed the particular goals & metrics, strategy and initiatives. That said - I’d argue that the most important projects in web3 should have some component of their strategy being to grow web3 itself. My observation is that with only ~5k active developers in web3, we are still early days and need to all be intentional about 100xing that number in the years to come. What do you think?

Last quick point - once metrics are agreed upon - would it make sense to have a function that gathers these? I’m new to the SAFE DAO and maybe there are already guardians or something doing this role in which case my question is of course moot.

5 Likes

I fully support an OBRA approach. Although reviewing and monitoring outcomes is critical, I believe the real power of this approach is the process it instills and what that process produces - a plan.

From a messaging standpoint, I view this as having DAO governance move from a proposal-based approach to a plan-based approach. By plan, I mean having discrete objectives (what you call strategies) and the initiatives to achieve those objectives.

This is an approach I have spent considerable time researching, as a founder in the DAO coordination space and as a genesis core operator with Wilfdfire DAO (a meta-governance services DAO founded by the Fire Eyes team).

Referring back to my renaming of strategies, I recommend a slightly different naming structure. I have thought long and hard about terminology (you need to because it is far more important than people realize) and this is the position I came to. DAO “seasons” are coordinated through a strategic plan. A strategic plan is comprised of objectives and initiatives.

A strategic plan, if you think about it as an object contains the narrative for why you have this particular plan and it defines the time period you are working in. In the case of Safe DAO that would be four months. The objectives are what the DAO wants to achieve (or avoid). You can have multiple objectives.

Each objective should have an unambiguous way to define what success means if the objective is achieved. These are the “goals and metrics” in 1k(x)'s model. You can see how it makes more sense to embed these in an objective rather than have the metrics across all objectives in a single bucket. For example, an objective could be, “To become a leading project in the web3 space” with a metric, “TVL is in the top 10 of all web3 projects”.

From a process standpoint, the first step is to create a strategic plan (a name, i.e., “Safe DAO Season 3”, timeframe, and possibly a draft narrative). Then the objectives are proposed, debated, and decided on. Once finished, the draft strategic plan is opened up to the community to propose initiatives that support achieving the various objectives. The community debates and then approves or rejects the proposed initiatives based on the community’s belief in the outcomes they will produce.

At this point the final narrative can be written, the strategic plan “published” or made active, and the DAO can start to execute. The plan now becomes the communication vehicle for what this DAO is working to accomplish. A great onboarding and project management tool. I have lots of ideas on this as well but I’ll stop here.

Anyway, these are my initial comments. I would enjoy participating in a working group to make this proposal a reality and to make Safe DAO a testbed for OBRA (aka, plan-based governance).

2 Likes

I like this structure and flexibility of the OBRA model. It may be a good way to avoid ‘working group complacency’ where progress can grind to a halt when people get busy or don’t have clear KPIs.

I’d consider adding something in addition to the $SAFE success reward for completion of initiatives that don’t go to plan. If a team fails to deliver wrt the metrics and KPIs outlined in the initial prop, they could recover some of the reward by presenting how and why they didn’t meet the expected goal so the rest of the community can learn & adapt.

I’m not sure if this needs to be an explicit guideline or could emerge as a cultural norm, but I think it could minimize KPI/ metric misalignment.

I also like having risk as a core strategy because it will encourage more meta initiatives that look at the long term resiliency of the DAO/ ecosystem. Although metrics for the risk category will be harder to define.

1 Like

@isaac you bring up a critical point. Although, you may not realize just how big an issue it is. Right now, web3 comp is broken - badly. Most people don’t get paid for their efforts or are at the mercy of the crowd to retroactively comp them when they have no idea of the effort involved or the results achieved. They basically guess.

More to your point, paying for results makes complete sense, except when it doesn’t. Where it doesn’t is when a community is doing complex problem-solving. For these situations, these communities need to support experimentation to discover the innovation that solves the problem.

In an all-or-nothing payout mindset, you will get little to no risk-taking, resulting in just incremental advancement. We need to reward the experimenters and pay for their efforts even if the desired outcome was not achieved. In reality, a great deal was achieved. The community discovered a path that does not work (for this org, at this point in time, with the given resources).

If you set up an objective/initiative, plan-based model you’ll realize the initiatives competing for resources to achieve an objective are actually a set of possible paths to success. If the chosen pathway project (aka initiative) doesn’t pan out. The community can learn from it and now make a more informed decision of what the next-best initiative should be resourced to achieve the breakthrough result.

When a community gets good at this process they will learn to fail fast, saving the remaining resources to be allocated to the next-best experiment. This is where I will add my next modification to the above model. I would not wait to do an end-of-season retrospective/review on the outcomes. I would make it an ongoing process. As soon as an initiative should be killed, kill it. Then move on.

The originally proposed process will likely lead to a major project management problem - the sunk-cost fallacy. This is to say, getting the project to the finish line, no matter what the evidence is saying, in the hopes that maybe the team can pull off success in the end. This wastes resources and is how many bureaucratic environments operate. The very orgs we’re trying to disrupt.

The end-of-season review should still happen. There still needs to be an accounting of the overall season results and for any non-experimental initiatives that ran for the season. Also, it’s good to stop and reflect and reset. This is when you can start the next season’s planning cycle by creating the next strategic plan.

1 Like