Introduction
Over the past 10 weeks, Chaos Labs has performed a deep analysis of three protocols’ STIP programs from a risk and efficiency point of view. This post focuses on aggregating the learnings from these case studies into a consolidated review to help delegates and protocols improve their decision-making in the future.
The post starts with a risk analysis of the program on these protocols. It then examines the efficiency of incentive spending, offering actionable insights from case studies to form recommended best practices. This leads to a discussion of incentive elements that should ideally be included in future programs.
The original analyses can be found here:
Risk Footprint of the STIP program
Overall, there is little evidence that the STIP program contributed to an increase in adverse outcomes manifesting due to the risks introduced. The potential risks identified, such as protocol bad debt, cascading liquidations, overleveraged activity causing price wicks, etc., did not occur during the program.
However, the absence of adverse outcomes does not mean that the program did not alter the risk profile of the Arbitrum ecosystem. Risk management must acknowledge that there is always a distribution of potential outcomes, and we need to contain the worst possible outcomes to maintain long-term stability.
We believe that the starting point of the program coming in a bear market with little appetite for leverage or experimentation by users allowed for the dramatic but safe growth on Vertex Protocol, Silo Finance, and Pendle Finance as these protocols went from being under-utilized to a reasonable level of utilization. This does not mean that similar risk outcomes would likely be observed should this growth start from a higher level.
Chaos Labs recommends that the Arbitrum DAO stay vigilant, particularly where exposures are leveraged and composed of multiple protocol dependencies. To help ensure that risk management remains part of decision-making at all times, it is recommended to add a risk section to future incentive program application forms. This section should ask protocols to explicitly declare potential risks, cross-ecosystem dependencies and integrations, and a plan for mitigating these over the program.
Productive Incentive Attributes
All three of the case studies examined grew their activity by multiple orders of magnitude, compared with an average TVL growth of 9.2% by STIP recipients. This presents us with many clear, actionable incentive strategies and tactics from the program for the Arbitrum DAO to learn from.
- Experiment and Iterate: Each of the studied protocols did this successfully in their own way. This was the clearest takeaway from the program and incentive programs should be thought of similarly to paid marketing campaigns in web2. The optimal design is constantly evolving and not always obvious purely to reasoning. There is no substitute for controlled experimentation.
- Keep the Incentive Criteria Clear and Simple: Users need to know what to do to react optimally to incentives. If the criteria are too complex, education becomes an issue, and attention will go to easier-to-earn incentives. All three protocols made the incentivized activity and payoff extremely clear and easy to understand.
- Use Incentives to Amplify New Features and Announcements: This was done by all three protocols studied and significantly impacts bootstrapping new markets, bringing visibility to new features and generally leaning into growth areas.
- Avoid Incentivizing Mercenary Behavior: Wash trading, sybils, and other inorganic activity is solely the result of incentives rewarding such activity and these strategies will be unwound as soon as the incentives end. There is no long-term benefit to incentives spent on these strategies as they are not retained after the program ends. Legitimate strategies earn fewer rewards as a result, and the program impact is diluted. There was limited evidence of this in the case studies as activity was mainly subject to a small positive fee.
Use Incentives to Increase Integrations and Composability: Incentivizing the end use case was a powerful growth driver for growing the ecosystem of a protocol. - Do Not Interfere with the Natural Balance between Supply and Demand: Focusing too much on either side of the market can upset the state of equilibrium. This will likely correct once the program ends rendering this allocation of resources sub-optimal. This does not apply when the market under-compensates certain activities and require long-term subsidies.
- Maintain Momentum with Native Incentives: If feasible, it is advisable to maintain momentum in the most impactful areas through protocol incentives once the incentive program ends. This strategy can better retain activity and position the application more favorably than others in its sector.
Potential Improvements
Despite the positive STIP outcomes of the protocols analyzed, certain potential improvements were identified that could have made their use of incentives yield even greater results.
- Onboarding New Users Has the Greatest Positive Impact: This is easier said than done, given the general microstructure of DeFi. It would be highly beneficial for applications running on Arbitrum and the Arbitrum ecosystem broadly for incentives to experiment with onboarding new users without getting exploited by sybils.
- Aspire to Broaden the Incentive Distribution: Expanding on the previous point highlights that the wider the incentive effect, the more diverse and extensive the onboarded activity will be, leading to a healthier application marketplace. Exactly what determines the optimal distribution is application-specific, with different types of apps naturally growing different activity concentration microstructures.
- Align Incentives with Ecosystem Goals: It does not make sense to grow elements of the Arbitrum ecosystem that are being deprecated over the long term. Bridged USDC.e growth is one element identified that grew as a result of these incentives and is being wound down, making it not an optimal area to target with incentives.
Conclusion
The three case studies performed on Vertex Protocol, Silo Finance, and Pendle Finance provided many vital takeaways for Arbitrum DAO members and protocols to consider in future incentive programs. This analysis covered the impact of the STIP program from a risk and efficiency standpoint.
Although little consequence was observed due to risks manifesting, this can be attributed to market conditions at the program outset, and future iterations will likely play out differently. We recommend close scrutiny of the potential for incentives to drive users too far along the risk curve, to the point where the ecosystem destabilizes. Protocols should be asked to include a plan to contain the impacts of incentive-induced excessive risk-taking in their application forms going forward.
Several consistent behaviors led to the success of these protocols’ incentive programs. The consistency of these behaviors, along with their overlap with traditional digital marketing strategies, gives us confidence in recommending their implementation in future incentive programs.
Lastly, a brief list of incentive attributes that were not included but could have further enhanced the impact of these programs is provided. These elements are not straightforward or as evident as the positive aspects and should be considered as areas for future experimentation and improvement.