A quick analysis of KIP voting history

Thanks to @Wes² for comments on an earlier version of this post.

The recent passage of KIP-23 seemed an opportune time to take a step back and look at the evolution of governance participation since the launch of governance in October 2021. Snapshot makes this relatively easy by providing information on wallets and number of tokens voted for each KIP, so I used that for the analysis that follows.

General Voting Patterns

As a first pass, here are the levels of participation in voting for each KIP (broken out on the horizontal axis). The bars in the top chart capture the total number of tokens voted for each KIP, and the bars in the lower chart break the votes out instead by the number of wallets voted.

Some preliminary patterns in the data:

  • Voting by tokens and by wallets have generally tracked each other.
  • The period from 10/21 - 2/22, corresponding to KIPs 10 - 15, was an exception to that pattern, with the number of wallets voting declining while the number of tokens voted was relatively high
  • Although they operate on very different scales, both tokens and wallets voted have generally declined since that time, with the decline being steeper in number of tokens voted.

These general patterns appear more sharply when combining wallets and tokens into a single representation. The box plots below chart the distribution of vote sizes (where each wallet’s tokens voted are the unit of analysis within each KIP) for each KIP on Snapshot. The solid boxes represent the interquartile range (IQR) or the 25th to 75th percentiles of the distribution of votes for each KIP. With the exception of KIPs 10 and 11, the majority of wallets voting in each election have voted less than 1,000 tokens each. The dots higher in the graph represent individual votes of thousands of ROOK each, and (in statistical terms, again taking the individual wallet as the unit of analysis) represent outliers in each vote.

Some observations:

  • The concentration of vote distributions toward the bottom of the graph indicates that the bulk of voters are not whales, which is reassuring from a community participation perspective. That said, the low number of wallets evidenced in the chart above implies that we have significant room for improvement in terms of expanding participation.
  • The bear market appears to be having an effect on whale involvement, with participation by large wallets falling off markedly from KIP 20 onward.

Contentious KIPs

Layering voting outcomes onto this analysis shows that KIPs have - once they make it into voting - have been relatively uncontroversial. Several KIPs have passed with some objection, but overall contention has been concentrated in relatively few KIPs, notably KIP 5 (CVX), KIP 10 (Visor Finance) and KIP 15 (the Q2 Rook Labs budget).

A closer look at these three KIPs shows a significant divergence between KIP 5 and the other two. Where the community (loosely defined as those with relatively small voting wallets) dominated support for KIP 5, the objecting votes were spread across significantly larger wallets. It’s unclear whether these wallets participated in the governance process that led to the vote but even if some did, the contention that followed that vote appears in retrospect to have been inevitable given the stark differences between those voting “Object” and those voting “No Objection”.

KIP 10 was almost a mirror image for KIP 5 in this regard, with a larger range of wallet sizes voting “No Objection” while objections were more concentrated among smaller wallets. The votes were more evenly distributed across voting positions for KIP 15.

Closing Thoughts

While each vote is somewhat unique given the nature of our KIPs, we can take away some preliminary conclusions here. The first is that this first stage of our governance has - with the exception of some very difficult times - gone smoothly once KIPs have gone to vote. That doesn’t change the fact that participation has been declining over time, across all sizes of voting wallets.

We keep analysing what is going wrong and we still do nothing about it for months.
We see something is not doing well → we analyze it → we discuss a bit more → we forget about it → we get back to it few months later → we analyze it once again → we forget about it 2 community calls later etc. This is the reason I dont participate in voting anymore nor in community calls. We discuss ideas 10 times and we do nothing afterwards. We keep procrastinating with so many good ideas.

The answer to my feedback is gonna be: “we agree, we acknowledge that there is room of improvement but we have first do 10 community calls to analyse more and then do nothing anyway. But in the meantime feel free to keep providing feedback. Thanks for your feedback.”

Thanks for your feedback (sorry, had to be done!). Yes we have moved slowly on some of these governance issues but are preparing some more substantive proposed changes to discuss sooner rather than later. Certainly sooner than 10 more community calls. But point taken.

1 Like

Would you like to aggregate a list with me of these good ideas that are stuck in this process? I’m not sure I agree with you, but would love to work with you on a solution to these problems you mentioned.

Great Work Jason!

One thing I didn’t consider is how voters for this DAO compare to voters of other DAOs. Are the trends here similar or different?

A larger sample size from older DAOs would also be interesting to look at: how are their voting trends during extended bull markets, extended bear?

The MakerDAO post you shared with me is great. If this DAO can gather data and learn from their implementation, it could act as one driver to improve voter turn out for the next 100 KIPs and beyond.

Thanks Wes! You’re right, it would be great to have some way of comparing this to patterns in other DAOs. The data is all there in snapshot so it’s a question of API wrangling and then figuring out how to compare. Definitely worth pursuing.

1 Like