# Optimism Retroactive Public Goods Funding — My Voting Framework Source: https://nickbtts.medium.com/optimism-retroactive-public-goods-funding-my-voting-framework-9caf0b218055 Press enter or click to view image in full size There is a total of 30m OP tokens being distributed in Retroactive rewards for Optimism Collective impact this season, with Optimism Badgeholders selecting and assessing the worthy projects, DAOs, applications, individuals and groups. This season I am one of the aforementioned Citizen’s House Badgeholders, and with the amount of value (~$50m) being given out this season, plus the respective visibility, fair allocation is a critical task to prove the viability of RPGF as a system. As such, I have spent time developing and refining a framework for allocation, which I’ll outline below in order to hopefully shed some light on what has been a somewhat chaotic and opaque process. Firstly, some stats: 642 projects to assess. 145 badgeholders. 30,000,000 OP to be wholly distributed. Some new infrastructure and processes are in place this season, notably Lists, allowing Badgeholders to curate lists of projects and assign OP for other Badgeholders to use as a direct guide. Unfortunately, Lists only require a text disclaimer around Conflict Of Interests, which means the system has, in some cases, been used for self-dealing, which when coupled with previous evidence of collusion between groups and paid grant-applications-as-a-service, puts the integrity of the Lists system in doubt. As a result, the only real way to be able to allocate the funds fairly is to assess every project individually, so as to be able to assess relative impact; although the overarching mission of impact = profit should allow for a fairly definitive calculation of OP allocation to each project, the fact the entire 30m OP is being distributed means we need to ensure the most impactful projects receive the highest percentage of the distribution. As an aside: as a project co-founder, longer term investor, erstwhile developer and also having spent time working with a high performing web3 marketing team, I believe I am able to assess fairly the majority of the projects for relative impact. There were a small number of omissions due to Conflict Of Interest or lack of depth of understanding. For many I sought information or expertise from my network. To simplify a relative distribution, the process is automated according to the following: Assign a cap to limited impact activities Apply a multiplier to high impact activities Calculate impact based on composite scoring across categories, adjust with the above modifiers, and then redistribute OP across all assessed projects. The scored categories are Technical Impact, non-Technical Impact, User Experience and Dev Experience, each scored out of 10. The following assessments apply the appropriate cap or multiplier: No Impact. A project that has little measurable impact to the Collective, has been funded for the specified impact through previous grant programmes, is spam or has misused previous grants programmes. Cap: 0 OP OP Stack. Technical contributions to the OP Stack generate huge impact for the Collective whilst we are in the process of building an onchain world. Multiplier: 3.00 Superchain. The Superchain is nascent and has very limited tooling and interoperability. Acceleration of Superchain is the single biggest value-add for the ecosystem, and as such should be seen as a Force Multiplier. Multiplier: 4.00 OP Mainnet Impact. Environment-specific or infra based impact returns more to the Collective if deployed solely, or mainly, to OP Mainnet. This is due to 100% of sequencer revenue being returned to Governance, vice a small percentage of revenue for other Superchain rollups. Multiplier: 2.00 VC Funding. VC-funded projects can create impact and generate public good, but often this is a ‘freemium’ type model, where the VC project can accelerate quicker, displace bootstrapped competition, capture the market and then introduce fees. Similarly, this public good provision can be used as marketing, and are considered as such an expense by the company. We should not exclude these projects, but we should ensure bootstrapped, opensource projects have the ability to compete. Multiplier: 0.1 Multiple Organisations. Multiple applications from subsets of the same organisation is disingenuous, attempting create a greater surface area for attention in order to increase rewards, whilst increasing workload for Badgeholders. Multiplier: 0.2 Impressions. For all content, including technical, where impressions are the primary metric. It is hard to assess impact from impressions, and is easily gameable. Multiplier: 0.2 Outsize Impact. Any project that excels in a single category attracts this multiplier (>9.00) . Multiplier: 2.00 Non-technical Content. We should bias toward technical progression, as our bottleneck is not currently users. Content generation through retroactive reward is generally hugely inefficient, un-coordinated and badly targeted. Cap: 1000 OP Events. As above, for both online and IRL non-technical events. Cap: 2500 OP. A number of iterations in Javascript later and the framework is ready to go. I have worked closely with Zoomer on a number of projects and we are closely aligned on impact, so as fellow badgeholders we teamed up to take on the bulk of the assessments between us. Press enter or click to view image in full size How things looked… It took well over 100 hours to assess every project, with at least one sentence describing each (or a reason for a cap). Some projects required some deep diving, particularly around undisclosed VC funding or associated entities. Zoomer did some great work in creating a list showing impact via revenue generation, and some of the lists eventually published provided some good context for scoring technical projects. I will publish a post-mortem once the ballot is submitted, as there are many, many lessons to learn from this process. I’m also looking forward to not receiving any more DMs from projects trying to shill their way to funding!