Hey, it’s me again.
I've spent quite a bit of time analyzing Arbitrum's evolving approach to ecosystem incentives and its implications for DAO governance. This piece explores that journey, highlighting shifts from programs like STIP and LTIPP to the recent 'Incentives Detox' and the broader lessons learned.
I hope you like it.
"Show me the incentives, and I'll show you the outcomes."
- Charlie Munger
You don't have to look far to find examples of protocols that have poured tokens into programs with inconsistent results. We've seen varied outcomes challenging simple narratives throughout crypto's numerous market cycles. While some follow a pattern of TGE → treasury funds get allocated → network metrics spike → results fade as programs conclude.
The Arbitrum ecosystem offers a more complex story. The ARB token price saw strength when incentives began, with significant downward pressure coming later when token unlocks occurred. While this correlation is noteworthy, we should recognize that broader market factors also influenced these price movements.
This creates a unique dynamic in crypto; unlike traditional grant programs distributing fiat currencies (whose value isn't directly impacted by the act of distribution itself), our tokenized communities navigate complex market forces: while poorly structured incentives can indeed create sell pressure, well-designed programs—particularly those targeting established protocols expanding to new ecosystems—can drive sustainable growth and positive price action.
The relationship between incentives and ecosystem health is more nuanced—context, program design, and strategic targeting all matter tremendously in determining outcomes.
There is (some) hope, though. We're seeing opportunities to learn from past mistakes and more efficiently align incentives for growth (and maybe even sustainability). This piece dives into the Arbitrum DAO and its progression through multiple programs and the recently concluded "Incentives Detox" - a deliberate pause to analyze results and design better approaches to funding ecosystem growth.
Arbitrum: Watching Incentive Evolution in Real Time
Arbitrum's journey through STIP, LTIPP, and the "Incentives Detox" shows the challenges and opportunities in developing effective funding mechanisms. Their experience highlights a fundamental mismatch: Traditional grant disbursement methods—with upfront payments, committee-based selection, and milestone tracking—don't always align with the goals of targeted incentives.
While grants excel at funding potential and promise, Arbitrum ultimately desired something more precise: measurable outcomes tied directly to network growth metrics. This disconnect between the funding mechanism and desired result became increasingly apparent as the programs evolved, forcing a reevaluation of how treasury resources should flow to maximize impact.
Their willingness to recognize this misalignment and adapt offers valuable lessons for the entire ecosystem. Let's examine how this played out by looking at Arbitrum's incentive evolution, starting with its first major attempt at ecosystem funding: the Short-Term Incentive Program (STIP).
The Short-Term Incentive Program (STIP)
Arbitrum's STIP started as a one-time, community-driven initiative in mid-2023. Initially allocating 50 million ARB tokens (about $40M at the time), the program was designed with specific goals: accelerating network growth, experimenting with incentive models, gathering data on effectiveness, and discovering strategies to drive user activity.
What made STIP unique was its community-driven approach. Unlike many incentive programs run by foundations or small committees, STIP was initiated and governed by the Arbitrum DAO. Projects submitted proposals through the Arbitrum forum underwent community review, and then a Snapshot vote was held, where delegates decided which applications to approve.
This process meant the community – not a centralized team – determined who received funding. Once a project passed the Snapshot, there were a variety of accountability mechanisms, including:
Funds weren't distributed upfront but streamed biweekly through Hedgey
A 5-of-9 multisig with community signers had the power to halt streams
All recipients completed KYC and signed formal grant agreements
Projects had to provide bi-weekly forum updates and create public dashboards
Recipients were forbidden from converting ARB or using funds for their farming
STIP's popularity exceeded expectations – in the first round, 97 proposals went to vote, with around 78M ARB requests. When the demand outstripped the initial 50M budget, the DAO took an unusual step: they passed a "Backfund" proposal in December 2023, adding 21.1 million ARB to fund the remaining approved projects. This brought the total distribution to 71.4 million ARB across 56 projects – significantly more than initially planned.
According to StableLab's analysis, STIP drove impressive short-term numbers—a 24% boost in total transactions, 29% more unique users, and a 47% increase in total value transacted. But beneath these positive topline metrics lies a more complex reality.
While STIP had established clear high-level objectives, L2BEAT's governance lead, Krzysztof Urbański, later pointed to a measurement challenge:
"The issue with the incentive programs was that the exact goals were never set, so it was hard to measure whether they worked or not."
This apparent contradiction highlights a nuanced problem: STIP had broad ecosystem goals but lacked specific, measurable targets for individual funded projects.
The program tracked metrics like TVL and user growth without predetermined thresholds for what constituted "success" for each protocol. This made objective evaluation difficult – was a 20% TVL increase amazing or mediocre? Without benchmarks established beforehand, evaluation became more subjective.
The sustainability challenge became increasingly apparent as incentives concluded. Some recipients, like MUX protocol, saw their metrics decline significantly after incentives ended, with TVL dropping from $63.5 million to just $21.4 million. This dramatic volatility suggested mercenary capital chasing temporary rewards rather than building lasting ecosystem value.
What impressed me most about STIP was the genuine community spirit and coordination. While the intention to decentralize decision-making was admirable, it created unexpected challenges in practice. The community-driven selection process, while inclusive, significantly slowed decision-making and created a lack of specificity in targeting and measurement.
The gap between STIP's high-level objectives and defining granular metrics for individual projects also created a challenge. Without predetermined thresholds for success at the project level, the program defaulted to vague measurements that made objective evaluation nearly impossible. Was a 20% TVL increase incredible or merely adequate? Nobody had defined this beforehand, making post-hoc analysis subjective.
STIP provided valuable lessons but highlighted the need for a more refined approach. Rather than abandoning incentives entirely, Arbitrum took the feedback and created a more structured follow-up program.
The Long-Term Incentives Pilot Program (LTIPP)
LTIPP was conceived as a follow-up to STIP, addressing gaps left after the first round exhausted its budget. This wasn't just "STIP Round 2"—it was designed as a pilot to test improvements before committing to a full-fledged long-term program. The program was intended to run for roughly three months, giving the DAO data before deciding on a more permanent framework.
Several structural changes set LTIPP apart from its predecessor:
Instead of open voting on every proposal, LTIPP introduced a five-member Council elected by the DAO to evaluate applications. This Council included GFX Labs, 404 DAO, Wintermute, GMX, and Karel Vuong from Treasure DAO. In addition, they added three Application Advisors (JoJo, SEEDLatam Gov, and Castle Capital) to provide feedback and guidance to applicants—something STIP lacked entirely.
Like STIP, LTIPP maintained protections against misuse, including:
Funds weren't distributed upfront but streamed biweekly over 12 weeks through Hedgey
All recipients completed KYC with the Arbitrum Foundation and signed agreements
The Council could halt any project's stream with a 4-of-5 vote if they detected violations
The DAO retained the power to override Council decisions via Snapshot
The Arbitrum DAO authorized up to 45 million ARB for the pilot, worth about $65 million at launch. In practice, the Council exercised restraint—approximately 30.65 million ARB was allocated to protocols, with roughly 14 million eventually returning to the treasury unused. About 70-75 protocols received funding, spanning DEXs, lending, perps, bridges, games, and more.
Each project needed to provide biweekly updates and metrics before, during, and after their campaign, and most created public Dune dashboards to track performance. They also had to disclose all wallet addresses used for distribution, allowing on-chain activity to be monitored.
In some places, the results showed improvements over STIP (at least in the short term). Gauntlet's analysis of Uniswap's liquidity incentive program under LTIPP found that $9.11 of TVL was added per $1 spent during the program, and $5.99 of TVL was maintained after the incentives ended. Additionally, after the incentives concluded, Gauntlet saw a 24.8% sustained increase in optimal-routing market share.
OpenBlock Labs' data revealed that stablecoin and staking token inflows remained in the ecosystem post-LTIPP (whereas some mercenary ETH capital did leave), showing that targeting the right liquidity could pay dividends.
Digging beyond surface-level TVL numbers reveals a more nuanced story about yield-seeking behavior. Mercenary capital might have left because of staking yield differentials and bridge yield opportunities on other networks. This highlights an important lesson: Effective incentive design requires understanding which assets to target, the competitive yield landscape across the ecosystem, and how it influences capital flows.
Not everything worked perfectly. Onchain perp exchanges underperformed, and Blockworks found that Arbitrum's perp sector was losing about 5% of its L2 market share in TVL during the program. Some projects tried unconventional incentive mechanisms that created too much user friction, particularly "complex incentives that involve purchasing vesting ARB," which hurt results compared to more straightforward liquidity mining.
As LTIPP concluded, the Arbitrum community faced a decision point. While the program showed promising improvements over STIP in some areas, the mixed results and persistent challenges with mercenary capital highlighted that even with the better design considerations and lessons learned from STIP, the programs couldn't fully solve the incentive alignment problem.
Rather than immediately launching a new program, L2BEAT's governance team proposed a radical idea: What if the DAO deliberately took a step back to learn from its experiences?
This strategic pause became known as the "Incentives Detox" - three months dedicated to analyzing results, identifying patterns, and rethinking how incentives should work in a maturing ecosystem.
The Incentives Detox: Taking Time to Think
The most intriguing part of Arbitrum's journey might be the three-month pause that just wrapped up their "Incentives Detox" period. It was a deliberate effort to analyze results and design better approaches.
One community member (0xDonPepe) summed it up nicely in the forum:
(By the way, I love quoting someone named 0xDonPepe in a research piece- Crypto is undefeated.)
The detox process, led by L2BEAT's governance team, kicked off with biweekly community calls to dissect data from previous programs. Instead of rushing into another funding round, they analyzed what drives sustainable adoption across programs.
Through the working groups and biweekly community calls, several key insights emerged. During the 3rd call, Tnorm (Gauntlet) presented an analysis of past programs, while Vending Machine did the same during the 9th call. (I should know; I was lurking/listening). Their research, which was supported by data from onchain, showed some problems that kept coming up: committee-based selection favored established players, upfront token distributions favored short-term capital over long-term engagement, and raw TVL turned out to be an unreliable measure of success because it showed the same growth patterns across chains regardless of the incentives.
This bias toward established protocols wasn't necessarily intentional but emerged naturally from several factors: established teams often had greater resources for creating polished applications, an existing reputation within the ecosystem, stronger connections to committee members, and were perceived as lower-risk investments. Meanwhile, newer teams and innovative but unproven concepts faced steeper barriers to accessing funding, potentially limiting ecosystem diversity and innovation.
As these insights solidified during the detox period, several frameworks emerged with different approaches to incentive design:
The "ARB Incentives: User Acquisition for dApps & Protocols" (by kamilgorski): This proposal emphasizes funding specific user acquisition strategies for protocols, moving beyond TVL to potentially include off-chain methods, and focusing on measurable user growth based on LTIPP learnings.
Entropy Advisors' DeFi Renaissance Incentive Program (DRIP) provides a mixed approach. It offers smaller initial payments for setup expenses while holding back bigger rewards that are given out only when certain performance goals are met.
IOSG Ventures' Results-First Model: This model suggests moving away from giving upfront grants and rewarding protocols only after they meet certain clear Key Performance Indicators (KPIs).
Merkl/Jumper Proposal ("Make Arbitrum Great Again"): Suggests utilizing Merkl's existing automated infrastructure to distribute incentives efficiently based on customizable, on-chain KPIs, aiming for objective and targeted liquidity mining rewards.
This diversity of approaches showcases the ecosystem's maturation but also highlights a central challenge: the governance mechanisms designed to democratize decision-making often impede the precise, targeted incentive designs that data suggests would be most effective. Arguments frequently erupt over specifics, leading proposal authors to deliberately craft broad frameworks that can gather consensus at the expense of clarity.
The result? Months of thoughtful analysis might not translate into action as quickly or decisively as the ecosystem needs. This tension between inclusive governance and effective execution represents perhaps the most significant challenge facing Arbitrum and DAOs broadly.
Ultimately, the Incentives Detox revealed a tradeoff in DAO governance: decentralized decision-making versus speed and precision. While the Detox had admirable intentions to be inclusive and gather diverse perspectives, we've seen the same pattern repeat across crypto governance – more voices lead to slower decisions and less specificity in outcomes.
In many ways, the Incentives Detox itself fell into the same pattern as STIP and LTIPP – good intentions around inclusive decision-making that ultimately slowed the ability to implement clear, opinionated strategies. There's a strong case that having some degree of centralized, specialized decision-making is beneficial when defining specific outcomes and aligning metrics precisely at the outset.
Learning From Arbitrum's Journey
What makes Arbitrum's evolution particularly valuable is their willingness to learn publicly. Rather than stubbornly doubling down on approaches that showed mixed results, they've continuously refined their methodology:
They started with a community-driven selection process (STIP)
They evolved to a more streamlined council model with advisor support (LTIPP)
They took a strategic pause to analyze results and identify patterns (Incentives Detox)
They're now working toward more automated, results-based mechanisms
These results-based systems are different from regular grants because they set specific goals (KPIs) that automatically activate rewards when met. Examples include rewarding protocols when they reach specific transaction volume thresholds, maintain a predetermined TVL level for an extended period post-incentive, or achieve verifiable user retention metrics. The key innovation is that funding becomes conditional on measurable outcomes rather than promises or potential.
Strategic upfront incentives can provide value for established protocols with proven product-market fit on other chains. These “battle-tested” protocols represent lower-risk investments since they've already demonstrated execution capability elsewhere. Giving protocol incentives to bootstrap activity on Arbitrum would likely generate strong returns compared to funding unproven projects.
The key insight: It's not about abandoning upfront incentives entirely but targeting them at proven winners expanding to new ecosystems. This approach acknowledges that projects with demonstrated staying power on one chain deserve treatment different from those of speculative ventures. It's a more nuanced strategy that leverages market signals rather than committee opinions.
Each iteration has gotten them closer to what might actually work. By sharing their data, hosting open community calls, and transparently discussing challenges, they've turned their journey into a learning opportunity for the entire ecosystem.
This evolution reflects a growing sophistication in treasuries' approach to resource allocation rather than a simple shift from grants to incentives. The Arbitrum community is developing a more nuanced understanding of when each approach works best:
Results-based mechanisms: excel for ecosystem-native projects where metrics can be clearly defined and measured
Strategic upfront incentives: remain valuable for established protocols with proven track records expanding to new chains
Public goods funding: continues to require traditional grant approaches for infrastructure that benefits everyone but doesn't generate measurable returns
The Arbitrum DAO's incentive journey reveals as much about the challenges of decentralized governance as it does about effective treasury management. The evolution from STIP to LTIPP to post-Detox proposals shows a constant tension between inclusive decision-making and decisive execution.
What's becoming increasingly clear is that effective treasury allocation requires specialized knowledge, swift decision-making, and targeting – qualities that often conflict with fully decentralized governance. The most successful approaches will likely involve:
Delegating specific decisions to smaller, specialized groups with clearly defined mandates
Ensuring effective accountability mechanisms exist to verify delegates execute their responsibilities
Establishing success metrics before programs launch rather than evaluating subjectively afterward
Differentiating between funding mechanisms based on risk profiles (proven protocols vs. speculative ventures)
Recognizing that sometimes more centralized, opinionated approaches can deliver better results than trying to please everyone
While complete decentralization might be the ideal, Arbitrum's experience suggests that hybrid models – combining community input on broad strategy with delegated execution by specialists – may ultimately deliver superior outcomes.
The most valuable lesson from this entire journey might be that governance requires as much experimentation and refinement as the incentive mechanisms it oversees. The central challenge remains creating systems that can be decisive, targeted, and accountable while maintaining the core ethos of community ownership.
In my next article, I'll explore how this shift from traditional grants to programmable incentives is playing out across the entire ecosystem and what it means for the future of building sustainable protocols and communities.
Peace,
Sov
