I decided to include all the metrics as I find some value in each one. I did not separate Logscale and the standard I felt that the more metrics there are the easier it would be to reward the best projects. All the metrics have their own pros and cons but Gas Fees is our source for RPGF so I allocated 50% overall to Gas Fees and Logscale Gas Fees. It was difficult as my initial allocation had Gas fees with an overall of 75%. but at the end I gave some more weight to other metrics.
As for the rest of my allocation
LS Interactions from Trusted Optimism Users, LS Total Transactions and Trusted Optimism Users’Share of Total Interactions’ each received 8% OpenRank Trusted Users, Power User Addresses & Trusted Recurring Users received 5% each. Interactions from Trusted Optimism Users I allocated 4% with the rest receiving 1% each
As for the OS multiplier I had to go with the max. Open source it the best way forward.
There is a common denominator for me, and it reflects network usage and demand for blockspace, as well as how much users are willing to spend, in sum. So I distributed my ballot in the following way:
Gas Fees (65%): This is the most important metric since it really shows the consumption and demand of the network from the perspective of how blockchains work: computational resources needed. It is neutral since it is the metric that objectively shows which apps are most used. It is a must to allocate a big portion to it.
Total Transactions (15%): Following the importance of network usage, it shows how much demand there is from users to perform operations in apps. Each operation indicates that the user wants to modify their state and then update their activity, whatever it is they are doing. I did not allocate more than this because, in the long run, it is not desirable for the signal to be distorted, to the detriment of the user experience (e.g., one single transaction to perform multiple operations).
Recurring Addresses (7%): Reflects interest over prolonged periods, which are signals of retention, which is good. I did not allocate more than this for reasons of farming cases where usability does not spill value nor showcase utility.
Average DAAs and MAAs (5% each): Reflects the importance of having more addresses interacting with each dApp; we want more of them over time.
Users Onboarded (3%): Where users had their first experience is theoretically a good metric but in a very limited portion. Since it is based on the definition of a “trusted user,” it would have a limited reach over the actual number of onboarded users.
I hope this proportion approaches a fair distribution of what we want to drive forward: more users, more usage, which is where the impact ultimately defines itself.
Additional aspects
About the Open-source multiplier, I decided not to use it this round, in line with many of the comments and issues raised during the round.
About LOGSCALE metrics, I decided not to use any since the logarithmic curve in base 10, while simple, penalizes the top projects too much in my opinion. I have held the view that customizable logarithmic curves (base n, where n is defined by other methods) make more sense here.
About Trusted metrics, I recognize the effort in this area but did not find it acceptable to use for most of them, so I decided to forgo their use where possible.
About the voting experience, I have nothing more to say except that it has been excellent. Congratulations to everyone who made it possible.
I believe that supporting smaller projects has a disproportionate impact. Hiring the first developer can more than double the speed of a one-person project, whereas hiring the tenth developer might increase productivity by less than 10% due to coordination costs. This is why I prefer to use log-scale impact; larger projects should receive the most support, but not proportionally so, making log-scale a decent measure.
Further, recurring activity and power users are good proxies for returning customers. We want folks using the superchain and blockchain as a habit.
Lastly, I give a full multiplier to open source. We want encourage actors who contribute back to the ecosystem via open systems.
I had initially decided to use the log metrics in order to make the distribution flatter among projects, however, after concerns were raised of using log scales and diving into the individual values of each project here, I decided to not use the logscales given the huge disparity it created on projects who have made the biggest contributions to the metrics offered.
Additionally, I did not use any of the trusted user metrics as the data chosen was very limited in nature: only 64 Badgeholders from the total pool actually had accounts linked to Farcaster (enabling seeding through them) and the selection of the web (the nature of what would turn into a trusted user) was also not refined enough to deliver a high signal. It’s a good start though.
Gas fees directly feed into the sustainability of the RetroPGF and therefore the Optimism Collective. Yes there may be apps that don’t have a lot of users but generate a large amount that keeps this an infinite game. I would have like though for there to be a cap on how much a project can get in relation to their Gas fees. Maybe we can think of to develop a multiplier so that even if this metric is “gamed” there is a cap. I think of this metric as one where we can leverage those trying to game it to our own benefit. I was also very shocked to see the very little amounts most projects generated, which reaffirmed my decision to discard the log metrics.
I did not include daily nor monthly metrics since I don’t see there is an intrinsic value in pushing people to interact with products that do not generate any benefit. This reminds me of pushing for people to get hooked on Instagram and other dopamine generating things, not really adding any value to society. I’d rather products focus on doing a good job at whatever it is they are trying to achieve, and if that is being used once every 3 months, then so be it.
I included onboarded users since I believe the concept is valuable, I think the execution and data selected for the metric is not good though, so I did not give it a large weight. I included it though to signal that this is the direction in which I would like things to move.