Entropy will be voting AGAINST this proposal. Before diving into our rationale, we would like to state that Entropy has a conflict of interest in that we have designed and authored a separate incentive program, which hasn’t been posted to the forum yet. Additionally, we’d like to acknowledge the amount of time and effort that the Patterns team has put into this proposal. Their interviews identified key insights regarding LTIPP projects’ level of knowledge about onchain user acquisition funnels and ability to calculate metrics like CAC and LTV. The proposed analytics tools and structures (enabling participants to work out core marketing metrics), as well as the ongoing reporting structure, additionally seem robust.
However, despite the recent changes, we are still not confident that this program is a significant improvement to LTIPP and STIP, and we fear that the DAO would be repeating several of those programs’ mistakes. For a program that is based on predefined measurable goals, the listed KPIs for the DeFi category in particular raise cause for concern, as metrics such as volume and number of trades can both be easily gamed. We fear that the program fails to incorporate the learnings of previous ARDC research work. For example, this is what Blockworks Research stated about optimizing for TVL in the lending category:
Furthermore, as we see it, analyzing absolute metrics instead of relative figures (e.g., total market share change) is flawed, since the absolute metrics are largely driven by current market conditions. When it comes to the program’s high-level structure, we foresee that the application and selection process will again put a lot of pressure on the projects, similar to previous programs. We’re especially concerned about projects having to design their own incentive and marketing structures, KPIs, as well as having to distribute incentives themselves(?). This concern is shared by other delegates:
As we have seen in the past, placing a majority of the responsibility for the campaign design with the projects, especially those that are smaller/newer teams, is likely to lead to subpar results. Taking into consideration that LTIPP saw over 170 applications, we expect the proposed program to see 100+ applications at minimum, which is more than the current committee structure can realistically help revise before selections. Thus, in the current design, the DAO would once again be selecting a small number of projects to run “experiments”.
Our team echoes several of the concerns @CastleCapital raised. It is our belief that generalized programs that include so many different types of projects make it impossible to evaluate and iterate upon. Having a large number of separate, project-specific campaigns creates fragmentation, lowering the overall impact and efficiency. Based on previous incentives research done within the Arbitrum DAO, as well as looking at the retention performance of concluded incentive programs run within other L2 ecosystems, e.g., ZKsync, it’s clear that generalized programs cause an increase in high-level market share metrics for incentivized actions, but this market share gain quickly reverts to the old, organic level once the programs end. Although the program proposed here differs from the conventional structure by focusing on marketing more heavily, it isn’t targeted enough on the demand side in our opinion, and we foresee that all increases in relative activity will revert within 90 days of the program.
Our overall view on incentive programs is that they should be extremely targeted, focusing on bootstrapping new products/markets that emerge from within the ecosystem, thus amplifying the effect deriving from natural product-market fit, as well as encouraging prospective verticals to develop by driving native as well as non-native projects to offer products within those verticals. In short, our belief is that once a vertical within an ecosystem reaches an organic market share and stabilizes, a structural or exogenous shock (such as efficiency gains from a new DEX model built on Arbitrum or an Arbitrum competitor’s sequencer experiencing prolonged downtime) is required for organic market share growth to materialize. When it comes to capital allocation, one possible avenue through which the DAO could create a structural shock in an already well-established vertical is by funneling sequencer profits back as incentives to that vertical—an idea we think should be seriously considered in the future as the DAO’s overall profitability continues increasing.