As we step into 2025, we continue to summon Ether's Phoenix through the evolution of Retroactive Public Goods Funding. This post outlines our plan to improve the Retro Funding mechanism, as well as the proposed scope of contributions Retro Funding will reward in 2025, with the end goal of more effectively and accurately rewarding those creating the impact on the Superchain.

TLDR

• Learnings from 2024: While Retroactive Public Goods Funding rewarded hundreds of builders in 2024, the program still has room to improve in accuracy and reliability, ensuring builders are rewarded fairly and continuously for their impact.

• Shifting to Continuous Rewards: The program will transition from annual rounds to ongoing impact evaluation and regular rewards throughout the year, offering builders greater consistency and predictability.

• Metrics-Driven, Human-Guided Evaluations: Combining onchain metrics with human expertise will create a more accurate evaluation system, while still allowing voters with relevant expertise to express qualitative feedback.

• Scope in 2025: This year, Retro Funding will focus on Dev Tooling (developer tools like libraries and debuggers), Onchain Builders (projects driving cross-chain activity), and OP Stack Contributions (with a focus on Ethereum Core Development). Details on these programs will be published in January.

• Vision for Ethers’ Phoenix: The end goal of Retro Funding remains to create an economy that fairly rewards impact across the entire value chain, addressing common market failures, and making the production of public goods as profitable as building closed-source software.

Building the Impact Oracle

Retroactive Public Goods Funding is the economic engine of the Optimism Collective. It rewards the impact contributors have on the Collective, in order to realize Optimism’s guiding principle of “impact = profit”: the idea that positive impact to the Collective should result in a reward to the individual. With 850M OP dedicated to Retro Funding, it’s one of the most ambitious initiatives to date aimed at re-envisioning how to better align incentives to impact.

Since measuring and defining impact is complex and hard, Retro Funding runs as an iterative experiment in which insights and learnings from each round inform the design of future experiments. In 2024, we made significant improvements to our approach to experimentation, which generated insightful learnings across the 3 Retro Rounds conducted.

Where we are today

In 2024, Retro Funding rewarded over 400 builders with 20m OP across 3 retro rounds. While we have made promising progress on measuring impact, the program still fails to reliably support important contributions to Optimism. Furthermore, we lack sufficient evidence that rewards have caused significant ecosystem growth.

Key builders have repeatedly highlighted that Retro Funding is not reliable enough to drive their contributions to Optimism, as they have no way of knowing how impact is measured, and when upcoming rounds will take place.

1. Taem from Sunnyside Labs (formerly Test in Prod), the first external OP Stack core development team, shared that: “I’m not convinced the current allocation mechanism can accurately measure the impact of technical contributions. I believe it's going to be improved because the core contributors are very serious about making it better.”

2. tmm from Wevm, which creates developer tools, such as Wagmi and Viem, that are beloved by onchain builders, mentioned: “Right now we can’t solely rely on Retro Funding to reward our contributions to Optimism. It happens too infrequently and there isn't enough predictability with reward amounts.”

3. Trent from Protocol Guild, a collective of Ethereum Core Developers, and the top recipient of Retro Funding across all rounds to date, stated: “While we've been a top recipient in past rounds, the unique large-scale nature of our mechanism still means it is not being rewarded proportional to the impact provided. The eligibility of rounds and allocation processes can still use a lot of improvement.”

4. WagmiAlexander from Velodrome, one of the most successful DeFi Protocols on the Superchain, highlighted: “High quality projects drive outsized impact, but the allocations to date don’t reflect that. As of right now, for application builders across the Superchain, Retro Funding is not seen as a significant factor.”

In its current form, Retro Funding rounds have room to improve in accelerating the Superchain’s growth. The program’s opportunities can be narrowed down to three buckets:

1. Alignment: Retro Funding lacks a structured framework to ensure that rewards reflect the value of impactful contributions, limiting its ability to truly operationalize the principle of "impact = profit."

2. Accuracy: Retro rounds can improve in their ability to reward impact accurately and proportionally. This has been a point of reflection on Retro Funding 5 & 6 with allocations showcasing very low variance of reward amounts.

3. Consistency: Rounds for a given scope have taken place once per year, with big changes in scope and eligibility. The rules of the game have constantly evolved, making it hard for builders to rely on the program.

Based on what Optimism has learned about Retro Funding in 2024, the program will evolve in 2025 to focus on measurement-driven impact analysis, consistent rewards, and a comprehensive framework to evaluate which public goods the mechanism can support.

The Impact Chain - A framework for realizing impact = profit

Optimism is the Superchain, a network of L2 chains, known as OP Chains, which share security, a communication layer, and an open-source technology stack. At the heart of the Superchain is the blockspace it provides to builders and consumers.

Optimism is built on impact=profit — the principle that positive impact to the Collective should be rewarded with profit to the individual — but thus far, Retro Funding has failed to translate this principle into tactical mechanisms within the program. The Impact Chain aims at addressing this problem, providing a framework for how the Collective will achieve impact = profit over time.

The Impact Chain is a dependency graph which maps out all contributions across the Superchain: from the protocol’s foundation  (such as Ethereum software, the OP Stack and more) to contributions that drive Superchain blockspace consumption (such as onchain contracts, developer tooling, and more).

In the table below you can see the key components of the Impact Chain and how they each relate to Superchain blockspace:

Although these key components frame the Impact Chain as a linear graph, in reality, it is a complex interdependent graph in which multiple components have the same dependencies across the stack.

Market Failures

The Impact Chain may suffer from market failures, where the market fails to effectively allocate goods and services (e.g. the allocation is not Pareto efficient), leading to a loss of economic value. Market failures include monopolization, negative externalities, information asymmetry, and especially, the production and maintenance of public goods.

In well-functioning markets, impactful builders are naturally rewarded (realizing impact = profit). This is not the case when there is a market failure. Retro Funding's job is to correct possible market failures by providing reliable rewards to impactful builders in otherwise unprofitable markets.

Bootstrapping markets

In addition to correcting failures within existing markets, Retro Funding aims to catalyze net new markets as well. Retro Funding plans to achieve this by identifying economic goals, and rewarding the impact of progress towards these goals. This capability will be particularly useful to bootstrap new markets, especially in kickstarting new features corresponding to technical innovation, such as interop.

The endgame

Our endgame is to get Retroactive Public Goods Funding to a state in which it is capable of  monitoring the whole impact chain and intervening in order to correct market failures and bootstrap new markets, while accurately rewarding impact to realize impact = profit; this state is what we refer to as Ethers’ Phoenix.

Accuracy

To realize the vision of impact = profit, Retro Funding needs to accurately measure impact.

In 2024, we have radically increased our advancement on how to measure impact by shipping experiments on metrics-based impact evaluation and expert-based impact evaluation.

Moving forward we hope to combine the best parts of what humans and data are good at, capturing both qualitative and quantitative aspects of impact.

Metrics-driven, human in the loop

This approach leverages data’s systematic reach with human intuition’s nuanced adjustments. Metrics establish a quantitative foundation, ensuring that projects are assessed objectively and fairly, while human input adds layers of qualitative nuance, including quality, innovation, and momentum. An iterative feedback loop lets us adjust metrics based on the insights of humans with domain-specific knowledge. This process is similar to RLHF (Reinforcement Learning with Human Feedback) in machine learning, but with an emphasis on retaining clear, interpretable inputs for human adjustments.

Process Overview:

1. Retro round goals: Each retro round establishes one or multiple measurable goals within its specific domain.

2. Evaluation algorithm: Metrics-based evaluations generate initial token allocation proposals within each domain. Humans with domain-specific knowledge review allocations, fine-tune the metrics, and choose the best allocations.

3. Iterative Refinement: Continuous backtesting helps align metrics with desired outcomes, improving the model over time. Iteration cycles are significantly reduced, allowing us to learn at an accelerated pace.

Retro Rounds Goals

Each retro round establishes one or multiple measurable goals within its specific domain. These goals should drive the system towards impact = profit. In a subsequent step, an evaluation algorithm is selected to achieve this goal.

Evaluation Algorithm

The evaluation algorithm defines the measurement of impact for a given retro round.

Today, Retro Funding leverages an evaluation algorithm which solely relies on the input of a number of badgeholders to evaluate impact.

In the future, a number of metrics will be used to evaluate impact.

Humans with domain-specific knowledge provide qualitative nuance to measurements by rating allocations, providing qualitative input and feedback. Based on this input, metrics are continuously refined to achieve accuracy.

Principles for metrics:

1. Evolutionary: As a measure becomes a target (Goodhart’s law), metrics need to continuously evolve to stay good measurements of impact.

2. Robust: As metrics become a target, they are subject to attacks such as manipulation. The metrics design needs to focus on robustness to minimize the effectiveness of attacks.

3. Reproducible: To enable performance measurement of an evaluation algorithm, impact metrics need to be reproducible over time. Data points which are not continuously generated over time will limit the ability to backtest evaluation algorithms.

While we use objective metrics as a basis for evaluating impact, it is critical that we reinforce the notion that these evaluation algorithms are inherently subjective. Over time we can encode this subjectivity into the system by introducing multiple competing impact evaluation algorithms, each based on different value systems. These value systems may fundamentally conflict, but embracing diversity is key to cultivating a decentralized & creative ecosystem.

Iterative Refinement:

After the conclusion of a relevant timeframe, we measure how the evaluation algorithm performed at achieving its goal and can validate or falsify hypotheses.

Pioneering the metrics-driven model with humans in the loop pushes us into a new territory. While Retro Funding 4: Onchain Builders provided valuable insights in this area, there’s a range of topics to explore, from the governance mechanisms of the evaluation algorithm to the datasets which enable measurement.

In the spirit of experimentation, we want to continuously validate that this model delivers desired outcomes. Here are the results we aim to achieve in order to confirm that metrics-driven, humans in the loop results in higher accuracy impact measurement:

1. Satisfaction with allocations: Badgeholders, and other stakeholders, will rate their satisfaction with allocations consistently higher compared to 2024.

2. Acceleration: The evaluation algorithm will be consistently improved upon at a much faster rate than the previously-used purely human evaluation paradigm.

3. Iterative Refinement: At the end of the Governance Season, we will observe significant progress in the quality of evaluation algorithms of specific domains.

Consistency

In 2024, rounds took place once per year per category, with the detailed scope and eligibility announced shortly before the round. The initially announced round titles and vague description preemptively set expectations for many builders which the program did not fulfill. This led to frustration for many builders, who were expecting to be rewarded, and found out shortly before the round took place that they were not eligible to participate.

To address these shortcomings**,** Retro Funding will no longer reward impact within annual  domain-specific rounds. Instead, Retro Funding will run as a continuous reward mechanism, with a regular cadence of impact measurement and reward distribution.

Ongoing Retro Funding

Metrics-driven impact evaluation enables the ability to continuously reward impact because it reduces the operational lift of measuring impact within a given time period.

This enables some exciting new upgrades to the builder experience:

1. Consistent scope & eligibility: While badgeholders will have the ability to update the details of impact evaluation, the scope of a domain specific program will not change within a Governance Season. Our aim should be to maintain a consistent scope over an extended time period, while keeping the flexibility to support changes in the Collective’s roadmap and goals (Intents) in the short-term. More on this within the “scope” section below.

2. Predictable rewards: By using a set of metrics, we can provide builders with insights on how their impact is measured and rewarded. This allows builders to form a much more interactive relationship with the system, understanding the causal relationship between their actions, the resulting outcomes and their rewards.

3. Rolling application windows: Builders can join the program at regular intervals, reducing the “all or nothing” nature of applying to a round.

4. Regular rewards: Builders receive rewards on regular shorter timeframes (quarterly, monthly, weekly).

Retro Funding Missions

Thus far, builders were faced with different token allocation programs across the Optimism Token House, Citizens’ House and Foundation, making it difficult to discover opportunities. To improve consistency across Collective token allocation, Retro Funding will become part of the larger Missions framework. While Retro Funding continues to pursue the impact = profit vision, this alignment will ensure cohesive operations across all token allocation programs.

Scope of rewards

The following are considerations to decide on the program domains and scope for 2025:

1. Impact measurement capability: What impact can we measure today?

2. Supporting the Intent: Which programs can advance the Collective’s 2025 Intent?

3. Value Chain applied: where can we observe market failures, or the opportunity to bootstrap activity?

Impact measurement capability

Since we’ve learned that metrics-based evaluation is more accurate and effective, let’s start with understanding what impact we can measure today. Our capabilities are largely dependent on the data coverage within relevant domains.

Over the past year the Optimism Foundation has partnered with Open Source Observer (OSO), an open source data pipeline to power the insights behind Retro Funding. Open Source Observer runs open data infrastructure as a public good. To date, OSO processes over 100TB from 12+ data providers to produce 1000+ data models maintained by 80+ contributors in a growing data collective community.

Today’s data coverage includes GitHub data, onchain activity, project artifacts, reputation from sources like Farcaster, ENS, and Gitcoin, and package dependencies (More details here).

There are several areas where we are missing sufficient data coverage today. These include network service dependencies (RPC node providers, indexers, etc), offchain user analytics, individual and team impact within a project, and market data for various services.

Supporting the Intent

In 2025, Optimism is focused on our Collective Intent to achieve “A set of stage 1 chains doing $250m per month in cross-chain asset transfers”. You can find out more about this intent in our [blog post] or [forum post]. This is a step in the journey to realize the [Superchain product vision] published last month. This focus for the Collective in 2025 can also add focus to the scope of Retro Funding in 2025.

Retro Funding in H1 2025

The following programs are planned for H1 2025:

• Retro Funding: Dev Tooling will reward toolchain software, such as compilers, libraries and debuggers, that support builders in developing onchain applications and adopting interoperability on the Superchain.

  • The Dev Tooling round originally planned as "Retro Funding 7: Dev Tooling" in 2024, will be folded into the ongoing rewards for Dev Tooling throughout 2025.

• Retro Funding: Onchain Builders will reward projects that grow the Superchain economy and drive cross-chain asset transfers enabled by interop across eligible OP Chains.

These programs will allow the Collective to double down on metrics-based impact evaluation, help support the Collective’s strategic priorities, and continue to address market failures across the Optimism and broader crypto ecosystems.

Retro Funding in H2 2025

In the second half of the year, the Foundation is likely to propose the continuous execution and iteration of the Dev Tooling and Onchain Builder programs. In addition, the Foundation plans to propose an OP Stack program, starting with a limited scope around supporting Ethereum Core Development. Optimism will work with builders in the community to develop a structure that relies on consistent measurement-based evaluation to reward core developers’ impact. Ethereum Core Development has already made significant progress within this area by implementing self-governance of reward allocation among core developers with Protocol Guild.

What about other contributions to Optimism?

Retroactive Public Goods Funding is increasing its focus in order to improve the core reward mechanism, make the program more reliable and trustworthy for builders, and address the market failures most important to Optimism and Ethereum’s success.

Ultimately, Retro Funding aims to expand to all layers of the impact chain as measurement capabilities improve, and as the world moves onchain.

Retro Funding is far from the only token allocation program within the Optimism Collective. To learn about other opportunities, see the grant overview.

Ethers’ Phoenix or bust

The goal for Retroactive Public Goods Funding remains to achieve the axiom of impact = profit, a world in which every contributor to Optimism is fairly rewarded for their impact. An economy where building public goods can be as rewarding as building closed-source software. A fork of capitalism, that embraces collaboration and market forces.

We have never felt more confident in this vision.

Stay Optimistic!