Insights and learnings from recurring DAO failures
Blockchains enable trust and transparency without a central authority, forming the backbone of a new digital onchain economy. Onchain organizations, most notably Decentralized Autonomous Organizations (DAOs), utilize this infrastructure to manage resources and achieve their goals in a peer-to-peer setting.
Over the past five years, DAOs have grown significantly in size, capacity, and complexity. Over 3,3 million addresses actively vote and participate in onchain organizations, with nearly USD 35 billion in assets controlled by DAO treasuries (source DeepDAO).
While DAOs have significant potential, this new form of coordination faces many challenges related to governance and operations. To reduce their complexity, DAOs often start as simple, centralized organizations and grow to become more resilient and decentralized over time, a process referred to as progressive decentralization.
This report aims to contribute to developing and adopting functional governance frameworks that increase DAO's operating capacity and allow them to grow and decentralize over time.
It draws lessons from nearly 40 DAO failures from their inception in 2016 until 2024 to identify patterns that negatively influence survivability and effectiveness. In the second half, the insights are used to provide actionable insights to improve the Governance of the Optimism Collective.
We are particularly grateful to Optimism Grants for approving our Mission Request grant, which made this research possible. We also thank everyone who provided input for this report through in-person conversations and written contributions.
Our team analyzed 37 DAO failures between 2016 and 2024 to identify key failure patterns. The data was collected through stakeholder conversations and desk research. The DAO Failure Dashboard is openly accessible and provides an overview of our data.
We differentiate between DAOs that are currently still operating (Active), have minimized their operations (Dormant), or have ceased to operate entirely (Ended). The DAO's current status is used as a proxy for the severity and impact of a failure.
We mapped each DAO's primary category, such as DeFi protocol or L1 blockchain, Network, voting tool, and structure used, and the primary and secondary failures identified.
The graph below showcases the primary failure identified in the current dataset. Failing to find a market fit for the product or services offered by a DAO is recognized as the primary failure, with over 40% of all failure cases falling in this category. In nearly 30% of cases, an exploit caused by a bug or hack caused a DAO failure. Governance and operations-related failures occur less, with Voter Apathy (8%) and Stakeholder Dissallignment (5%) occurring most in this subcategory. Even when combining all Governance and Operations failures (19%), their likelihood and impact remain relatively small compared to product-related failures.
The below graph visualizes the percentage breakdown of primary failures that resulted in a DAO ending (blue column) compared to the average percentage of a primary failure method across the entire data set (red line). The difference from the top of the blue column to the top of the red line showcases the increased severity of a specific failure type.
A sizable exploit resulting from a hack or bug is another common reason a DAO ends, with nearly 44% of catastrophic DAO failures resulting from an exploit. This data further suggests that product fit and security are the most critical responsibilities of a DAO. The existence of liquid governance tokens could be one of the reasons for the high failure rate as a result of exploits, as the hacks often include the protocol native token, which creates significant sales pressure on the project. The reduction in token price often severely impacts the operating capital of these DAOs, as most DAO treasuries are centralized around their native token.
Failing to find a product market fit is noticeably second, with 37,5% of critical failures associated with a DAOs inability to find market fit. In some cases, DAOs may be created to serve a specific purpose at a particular point, such as aiming to acquire specific assets or testing a specific thesis. When such DAOs fail to accomplish their primary objective within a reasonable timeframe, there often is little to no reason to continue operations. This dynamic represents the market forces of the onchain economy, with capital formation and talent moving around at the speed of the internet. These forces make it relatively easy to attract initial commitments but increasingly complex to retain attention and talent.
While governance and onchain voting are often at the forefront of DAOs and their operations, the data suggest that lawsuits and leadership scandals are more likely to result in a DAOs collapse. This suggests that “good governance” is most important at the leadership level of DAOs and indicates that having the right people involved and committed to the DAOs' objectives is one of the key areas on which DAOs' governance should focus. This dynamic is much more akin to the governance process of a country - which often centers around election and representation as opposed to direct democracy on the legislative level - than usually imagined when thinking of DAOs as a means to have all stakeholders participate in Governance.
When including the secondary failures, the frequency of stakeholder dissallignment and concentration of power significantly increases from 3% to 13% each. When analyzing the individual cases further, the data suggest that Voter Apathy often leads to concentration of power, which cripples the overall governance of an organization. Also, the Lack of PMV and Stakeholder dissallignment frequently go hand in hand. Often, DAOs lack the mechanisms and rigor to properly align on and commit to a specific strategy that is worth executing. The lack of a long-term horizon and structures that prefer many small independent initiatives over large strategic plans often results in an environment where objectives and information flows vary widely across the DAO, significantly limiting their capacity to progress. However, the opposite can also be true - as DAOs often operate online transparently, creating the ability to document and make all organization processes and data available. The introduction of strong data pipelines leveraging information management solutions and AI agents to organize further and operationalize the insights has the potential to accelerate DAOs capacity to make the right decisions.
The current dataset and research scope provide limited insights into the influence of different voting structures and governance tools on the capacity of DAOs to make the right decisions, as nearly all DAOs follow the same native governance token structure and comparable tools (Snapshot or Tally). Future research can focus more profoundly on the patterns that increase the likelihood of a specific failure state occurring and provide insights into how to resolve them.
#1 The value generated by DAOs offerings should be the main focus.
A DAO's survival is deeply coupled with its ability to provide value through its products and services. Decision-making in DAOs should be organized around product and business KPIs.
#2 Setting the right top-level strategy is the most critical decision for a DAO to make.
Aligning the best strategies to realize product development is essential. It ensures that resources are spent cohesively and success is measured similarly.
#3 It is essential to select the right operators and keep the ability to withdraw their power.
Attracting suitable talent is key to realizing a DAOs vision. DAOs should aim to incentivize contributors and adequately create processes for removing some or all of a contributor's power when they no longer play a meaningful role in the DAOs course.
#4 Proper data pipelines and information management solutions are key to increasing decision-making capacity over time.
They provide DAOs with a collective intelligence that can be leveraged to make the right decisions, especially when paired with AI agents that can synthesize large data sets.
One of the largest and most influential DAOs is the Optimism Collective, a collective of communities, companies, and individuals that aim to scale Ethereum and build a better internet. It governs the Superchain, a network of interdependent chains built on shared standards and values using the OP stack, and funds developments with the OP token.
This analysis aims to strengthen the Optimism Collective by cross-referencing Insights and learnings from recurring DAO failure with the Optimism Working Models for Decentralization to identify vulnerability and actionable recommendations for future improvements.
The key domains identified and visualized in the Decision Diagram Working Model help focus this analysis, and the reasoning of the Decision Design Framework increases the insights' adaptability and implementation.
Before evaluating the key domains, we review the overall Optimism Working Models for Decentralization based on the key takeaways of the DAO failures study.
Optimism has a clear and detailed roadmap for product development and decentralization, which increases security over time through standardization. This is key in addressing the large thread created by exploits and contributes to its product-market fit.
One limitation is the lack of baked-in product adoption guidelines and clarity on how the product produces and captures value. While DAOs are often compared to nation-states, the reality is that the governance of technology ecosystems is highly dependent on their adoption rate. As such, key organizational north stars should be tied directly into each of the layers of decentralization to set clear expectations of the results each layer optimizes for. This allows for better identification and review of the layer’s parameters.
On the protocol layer, adoption numbers, such as the Superchain % of Major Crypto Txns from the Superchain Health Dashboard, could be utilized. On the economic layer, the Monthly Est. Collective Revenue could be compared to monthly spending to visualize P&L over time and simulate potential trajectories toward self-sustainability.
The Working Constitution, Superchain Product Vision, and Optimism Season intents provide guidelines on the overall vision and priorities for the Optimism Collective. However, they don’t clarify the best strategies to accomplish these objectives. The current Mission request-driven grant allocation framework allows for the submission of stand-alone missions, which in the past led to unwanted overlap and unallocated missions. The complexity of deciding the proper mission is multivariate, as their effectiveness is interdependent and influenced by many factors, including available talent and urgency. Determining and assigning Missions requires high expertise and increased degrees of ownership. We propose to operate Missions in a manner that closely resembles how Government ministries function, with key civil servants with leadership and professional analytic experience owning a cluster of Missions that optimize for a specific domain. These Mission clusters could contain a degree of pre-assigned contributors to guarantee the ability to execute certain core responsibilities within a specific budget and timeline while opening up missions as tenders. The Cluster leads are responsible for the execution and tracking of their missions and communication toward the overall Optimism Collective. An extreme example of an execution cluster is the OpCo in ArbitrumDAO, a new entity formed to execute ArbitrumDAO’s strategy, which is projected to be funded with 22M ARB for 30 months.
Another limitation is the short duration of Season and Grant Cycles, as most initiatives take weeks or months to be assigned. By the time the projects are completed and results become measurable, the goalpost has already shifted, and the focus has moved to the next Season. It is key to have some consistency in objectives and strategies across seasons. A way to resolve this could be to introduce objectives or intents across multiple Seasons, which is parity done by the Working Models for Decentralization. Alternatively, the duration of a Season can be extended, or various Seasons can be merged into a larger frame, such as an “Era”, which could consist of 3 to 4 seasons.
The bicameral governance model designed by the Optimism Collective separates purely token-based voting from a form of elected representation. The Citizen structure provides a path toward withdrawing one’s power; however, this function seems underdeveloped.
The Citizen House and the Token House have little re-election taking place, resulting in stakeholders with no or limited active commitment remaining influential at the expense of attracting new committed contributors - often against their own will. A time-bound redelegation function, where tokens should be redelegated every x season, should be introduced to increase the effectiveness of the Token House. Citizens should be elected to provide perspective on domains they have context on - instead of randomly assigned to clusters and elected through private processes. Separating between “regular” citizens who offer input from a board perspective and “expert” citizens with high expertise in a specific subdomain can be a solution. The election of regular citizens should be straightforward and take place through public mechanisms, for expert review should be based on publicly verifiable experience and contributions. Onchain activity should play a role in assigning citizen roles, as one of the key innovations of Blockchain is the fact that users can be owners, which goes beyond token ownership with the concept of Citizens.
It’s currently time-consuming to understand and remain up to date with the Optimism Collective. Many resources are spread across different locations, with many key pieces living solely on the Optimism Forum. New tools like Governance Hubs that organize key resources should exist. In addition, reports should be generated at the end of each Season to revise developments and showcase progress. By creating solid data pipelines on all layers, Optimism Collective members can navigate decisions with the help of AI agents that can process large amounts of data efficiently.
Overall, the Working Model for Decentralization is well in line with the key suggestions from the DAO Failures study and, with some modifications, can become the best model for progressive decentralization.
We prioritize our efforts based on the potential impact, likelihood of happening, and relevance of the topic to the current state of OP Governance.
The current working models for further decentralization provide an initial roadmap; however, the public discourse and communication around it have been minimal. A clear example of this is the publication of the Superchain Product Vision, which paints a two-year product vision but fails to reply to any of the follow-up and clarification questions asked by members of the Token and Citizen House.
The DAO Failures studies showed that governance models with multiple substantial stakeholders are more rigorous and active than a governance system where a small group (often the core team) directs most decisions.
As such, Governance must be included in elements of the North Star and priority setting to improve the content of the working model and increase the buy-in of the larger Optimism Collective. The Optimism Foundation must engage more with stakeholders using public channels such as the Optimism Forum and respond to feedback, especially from Optimism Governance members.
The suggestion is for the Optimism Foundation and key ecosystem stakeholders to compile measurable targets that meaningfully progress Optimism's mission and make it more concrete. Governance should be involved in approving any large north-star setting efforts.
In addition, the priorities identified in the working models for decentralization should go into a formal input and review phase, after which governance should vote on ratifying the collective priorities and roadmaps. A process should be implemented where priorities are reviewed at a fixed cadence, and Governance can propose modifications.
Citizens, given their expertise, should be deeply involved in scoping targets and priorities, while the Token house should be the primary approval to ensure Token holders.
Once in place, all Optimism Collective-funded initiatives must be linked to the agreed targets and priorities to receive funding.
Many DAOs fail due to poor financial management, which combines ineffective budgeting and treasury control. Creating clear targets and priorities, as described above, can help increase the efficiency of overall spending; however, additional measures should be taken.
The Token House should approve the proposed budget sizes for each Season, as these allocations directly influence the number of OP tokens hitting the market. The budgets should clearly articulate their relationship to past initiatives and alignment with the targets and priorities.
A report should be published at a fixed cadence, initially at the end of each Season, that allows Optimism Governance to evaluate the budget's efficiency. This report will help the Collective craft more effective budgets over time and increase accountability.
Having a large majority of a DAOs treasury in its native token has proven risky, as it intensifies the impact of adverse market movements on the organization's capacity to coordinate. The suggestion is to diversify a share of the Optimism Collective. Many options exist for how to diversify, such as the conversion of OP to stables through a sale using decentralized mechanisms or for one of the Optimism Collective affiliated entities to organize a direct purchase of OP tokens to decrease the changes of the purchased OP tokens to create sales pressure on the token.
Alternatively, the Optimism Collective can develop mechanisms to increase ownership of Superchain-native tokens, such as OP Chain tokens. Token acquisition can occur through revenue-based contributions or direct token swaps with key ecosystem players.
The question of who should become a collective delegate or citizen is critical as voter selection processes result in low participation and a concentration of power among a small group of stakeholders. Optimism Collective should continue to create inclusive and well-structured voter selection systems that encourage active participation and align governance decisions with the collective goals of their communities.
Besides the potential benefits of time-bound redelegation for the Token House, the separation of regular and expert citizens, and the inclusion of onchain metrics in representative selection described previously in this report, transparency should be highlighted.
Opaque systems, where it’s unclear who makes the decision and based on what they make, negatively influence the system's overall functioning. The proposed method is to work back from desired outcomes in line with the overall optimism strategy to estimate the right type and function of governance participants. From there, precise evaluation and election processes can be published and operated. This is especially relevant for the (re) election of Citizens, as they play a key role in the Optimism Collective.
With most OP tokens still unallocated, with just over 1,35B out of 4,3B OP Token in circulation (31%), a key method for increasing the Token House quality and resilience is increasing the share of new OP tokens entering circulation that are delegated and used to govern over time. Various methods exist to accomplish this, but initial pilots could focus on adding conditional bonuses in OP to airdrop and governance fund reward participants who continue to engage meaningfully with the Optimism Collective over given periods, comparable to equity incentives in startups.
With Season 7 entirely dedicated to Interoperability, ensuring all layers of the Optimism Collective make meaningful progress on this front is crucial. The protocol layer developments are well-resourced and progressing quickly; however, the social and economic layers lag behind.
Without social-economic coercion, the vision of a unified Superchain can not be realized.
It’s essential to identify the current and desired states of the social and economic layers and take significant steps forward. However, we believe the resources dedicated to and earmarked for social-economic coercion are insufficient. As such, some of the progress on the social and economic layers will depend on the collective ability and willingness of all stakeholders to unify.
As Superchain Eco, we’re dedicated to accelerating the Superchain Ecosystem by functioning as a resource and coordination hub that promotes sustainable and collaborative practices across various stakeholder groups. We believe in the power of optimism and will continue proactively initiating efforts to increase social and economic coercion across the Collective.
