The ARC’s vision on historical incentive programs and its future
Treasure’s Arbitrum Representative Council (ARC) would like to share our vision on the current state of incentive programs, more specifically all STIP variants and the LTIPP, and our vision of what an ideal future incentive program would look like. The reason we are sharing this now, is because we want to reflect and state our opinion on the state of incentive programs and to encourage having an open discussion about this topic before any new incentive proposals go live. We will first share some historical context of pain points that we experienced, to then be able to give recommendations towards a system of incentives that is more data-driven, and has a higher degree of accountability and flexibility.
Proposal Voting
When looking at the STIP we believe that the approval structure of STIP proposals is fundamentally flawed. The expectation placed on delegates to conduct detailed analyses of user incentive programs and to vote on each proposal is unrealistic and counterproductive. The STIP-Bridge tried to solve this issue by introducing the ‘optimistic voting mechanism’, but our concerns
that it could lead to a higher degree of passive acceptance by delegates (while still placing the review burden on them), proved to be true. In fact, there were only two delegates (Seed Latam and L2BEAT) that challenged the proposals of projects and exclusively did so on the basis of the ARDC recommendations. This not only means that the delegate basis that challenged projects was very small, but also that the ones that did choose to challenge proposals did so on the basis of an external body of review. Effectively, the review of proposals during the STIP.B (the challenge) has largely been outsourced to a single entity (the ARDC), and arguably so has part of the possible blame that could come at the cost of challenging a proposal; as can also be seen by the argumentation for the challenge given by these delegates, such as: ‘our posting of the challenge is not a statement as to whether or not this application should be funded, but merely an administrative act to solicit delegate input’ and ‘Following the ARDC recommendation’. The conclusion we take from this, is that a more streamlined voting process should not come at the cost of proper review.
Besides the time intensive and possible political pressure reviewing proposals puts on delegates, there is also the matter of expertise. If we want to grow Arbitrum, we need incentives to grow with it. Expecting all delegates to have subject matter expertise on all verticals and sectors within Arbitrum is an unrealistic ask. For this reason, the council and advisor structure were introduced in the LTIPP. Although this is definitely an improvement to all STIP voting mechanisms, there is still not enough flexibility and room to iterate on incentive structures for specific sectors; but also, more generally.
Data collection
The data collected and provided through STIP rounds proved to be insufficient to use for incentive structures that have come after it, and there is not enough research done on STIP data to measure the effect of the STIP on network growth. This is especially worrisome considering that the top 10 wallet addresses that received incentives through the STIP have gotten a combined 18.43% (6.22 million $ARB) of all the incentives, and the top 100 wallets 44.83% (15.43 million $ARB). Some projects took it upon themselves to provide data on their STIP incentivization, which is also problematic considering it will lead to data being used differently by projects and therefore making them less comparable to each other, and other non-incentivized projects.
We have data that we do not effectively use, while also continuing to fund incentive programs and iterating on them merely on an experiential and theoretical basis. We need data and analyses, to be able to compare and iterate on incentive structure mechanisms. Analyses on STIP and LTIPP metrics should be done, these analyses should be debated by the DAO and referred to for the creation of future incentive programs.
Election structure
For the sake of decentralization and democracy within the DAO, we think that all major positions should be voted for by the DAO. This becomes ever more important if we want to grow the DAO, and delegate more responsibility away from current DAO delegates towards representatives with subject matter expertise. In relation to the previous incentive structures, this means that roles such as the PM and the Data Provider would become electorally decided.
Feedback structure
The creation of the council and advisors for the LTIPP has been a significant improvement from having a ‘direct DAO democracy’. Advisors have been helpful in analyzing and iterating on proposals together with relevant parties. Yet, a communication divide seems to have impacted the advisors’ ability to ascertain clarity on the (approval) requirements of the council. Partly, because the evaluation criteria in the rubric were qualitative and therefore subjective in nature. For instance, there was a lack of clarity regarding what types of incentive mechanisms could or could not be used. The reason that seems to be at the cause of this is that this was a reaction to the STIP rules being too stringent, and therefore inhibiting projects’ abilities to create innovative and project/vertical specific incentive designs.
Conclusion
To address these issues, we propose the establishment of a professional, well-funded structure dedicated to overseeing, refining, and promoting best practices for user incentive programs on an ongoing basis. This entity would function similarly to for instance OpCo, AVI and the GCP. The Easy Incentive Program (EIP) is a solid piece of infrastructure that could in part set us on our way to achieve this vision. Namely, the proposed dApp has the ability to create ‘program categories with independent budgets, councils, and applications.’ Working towards the creation of independent incentive programs means that:
- Incentive programs can be created independently and approved through a DAO vote, but can also be wound-down without it impacting the overarching incentive structure altogether. This gives the DAO more control and flexibility on what to fund and when, thus being able to continuously keep incentivizing based on market needs (i.e. data-driven).
- Increased competition will improve the effectiveness and ROI of incentive programs. By having multiple concurrent incentive programs, the DAO can learn and iterate way faster on the effectiveness of various incentive structures and programs. In turn, this way the DAO is also able to reward the most effective members within an incentive program, as well as reward and pursue effective incentive programs as a whole.
- There is a more viable argument for the DAO to delegate power and agency towards independent councils. Projects applying for incentives will benefit from this, because the team reviewing their proposals will have more subject matter expertise and will thus be able to help them perfect their incentive proposal better than a ‘general’ council. Delegates can be relieved of having to review individual proposals and merely have to review these independent incentive programs for the same reason that more confidence can be placed in these subject matter experts, but also because there are actual implications to how an incentive program is being ran (i.e. the incentive program as a whole can be reviewed by the DAO, held accountable and have their budget be adjusted accordingly). This puts more responsibility in the hands of the council in having specific incentive programs succeed.
- There is more freedom for incentive programs to create their own rules. The STIP showed to be too stringent in its rules to be able to nurture innovation, while the freedom in the LTIPP caused the advisory and council bodies to not be able to sufficiently create and communicate objective measures for feedback and review. Smaller and more independent councils create the ability to create tailor made forms of review and approval, such as a more objective and clearcut rubric.
- It removes direct conflicts of interest for delegates and builders within Arbitrum. A lot of the delegates within the DAO are also building their own projects on Arbitrum, and have a responsibility to both their own project and Arbitrum. In terms of a general incentive proposal, the projects with a larger delegate base will always have an enlarged influence and can therefore impede on innovation by attracting more incentives towards them and/or less towards possible competition or other (especially newer, and developing) verticals. By creating independent incentive programs, contrary to the general incentive programs, there is a collective interest to create a successful incentive program to be able to continue to receive funding.
Independent incentive programs can cause chaos if not done right, and have delegates lose oversight of what is going on. For this reason, an overarching oversight committee and data provider(s)/analyst(s) should be instated. The oversight committee should be responsible for general oversight, an advisory body for the setup of new incentive programs, communication between the DAO and the incentive programs and other administrative tasks. The Data Provider is there to collect and provide data, processing this data and standardizing this data across projects and councils, to be able to accurately inform the oversight committee and the wider DAO on projects’ and councils’ effectiveness on a data-driven basis.
edit: tagging @GFXlabs @404DAO @WintermuteGovernance @coinflip @JoJo @seedlatam @CastleCapital @Matt_StableLab @paulsengh for visibility.