I’ve been building governance analytics tools and Lido caught my attention
for an interesting reason: your voter quality is exceptional.
Out of 46 DAOs we track, Lido scores a perfect 10/10 on Human Participation
Rate — every single voter appears to be a genuine human. No bots, no sybil
activity. That’s rare.
The flip side: only about 170 out of 32,500+ followers actually vote,
and voting power is fairly concentrated (top 5 wallets hold ~50%).
So Lido has a quality-over-quantity dynamic — the people who DO vote are
real and engaged, there just aren’t many of them.
Curious how the community sees this. Is low turnout a concern, or is
delegation working as intended? Would love to hear from active voters
and delegates.
Following up on my post about Lido’s governance health — we just launched
the full Decentralized Governance Index.
Lido ranks #10 out of 44 DAOs with a GVS of 7.3. Your Human Participation
Rate is a perfect 10/10 — only a handful of DAOs achieve that.
The areas pulling the score down: Decentralization Index (5.6) and
Power Distribution (4.5). Consistent with the concentration patterns
I mentioned earlier.
Full comparison across all 44 DAOs: chainsights.one/dgi
Thanks BCV — appreciated. The concentration patterns in Lido are consistent across multiple metrics, which is why we track them daily now.
Curious: is there a specific aspect of Lido’s governance concentration that concerns you most? Voting power distribution, delegate concentration, or something else?
Agreed. The data supports this — when a handful of large delegates hold most voting power, governance decisions reflect few perspectives regardless of token distribution.
The DGI tracks exactly this: Lido’s Power Dynamics Index is 4.0/10, and the top 5 wallets control 65% of voting power. That confirms what you’re saying.
Question is: how do you incentivize more small delegates to participate actively?
Thanks for the context, the discussion, and the questions. Let me add my two cents.
From the start, the Lido Delegate Program was designed around quality over quantity. Decision-making in Lido DAO is high-context and complex: the DAO governs all aspects of protocol operations, grants, and treasury management. That requires deep involvement and sustained attention, which naturally limits how broad active participation can realistically be.
Concentration does exist, that’s true. At the same time, metrics like “the top 5 wallets control 65% of voting power” are often cited without sufficient context. A meaningful share of that voting power sits in the DAO treasury and CEX wallets, which in practice represent user tokens rather than coordinated governance actors. Treating those wallets as unified power centers can be misleading.
I don’t see the system as excessively concentrated, especially when taking Dual Governance into account. Stakers have real counter-power there, which materially changes both the governance risk profile and how concentration should be evaluated.
On small and mid-sized delegates: it would be great to see more of them. In practice, however, LDO holders consistently choose to delegate to known and trusted delegates. Actively pushing treasury delegation to smaller or less established delegates would be a mistake in my view. Several DAOs have gone down that path and ended up assigning power to actors with little or no skin in the game, which weakens governance.
Ultimately, this is a conscious trade-off - to prioritize depth, competence, risk ownership, and accountability over raw numbers. A smaller set of active delegates is balanced by a high level of transparency and clear accountability to LDO holders and the broader community.
Thanks Jenya — this is exactly the kind of context that helps sharpen the analysis.
You’re right that the “top 5 wallets” metric needs nuance. Treasury and CEX wallets representing user tokens are fundamentally different from coordinated actors. That’s a gap in how we currently present the data — we should either exclude or flag those wallets explicitly.
The Dual Governance point is interesting. We’re not capturing counter-power mechanisms in the current DGI — it only looks at Snapshot voting patterns. If stakers have real veto power through a separate mechanism, that changes the risk profile significantly.
Two follow-up questions:
Is there public documentation on which wallets are treasury/CEX that we could use to filter them out?
How would you suggest we factor Dual Governance into a governance health score? Separate dimension, or modifier on the existing ones?
Appreciate the pushback — this is how the methodology gets better.
Good call — Etherscan labels are a solid starting point for identifying treasury and CEX wallets. We could cross-reference those with known multisig addresses (e.g. Safe) to build a more complete exclusion list.
For Lido specifically, would the recognized treasury/multisig addresses be documented somewhere beyond Etherscan — like in governance docs or the DAO’s own records? That would help us validate the labels and catch any wallets that Etherscan might not have tagged yet.
On the Dual Governance question — any thoughts on how that should factor into a health score? Curious if the Operations Workstream has a perspective on that.
The correlation between high quality and low participation often points to information asymmetry. Even for rational actors, the “governance debt” (time spent tracking every AIP) is high. Are there specific areas—like CSM or treasury management—where participation is notably higher? Understanding where the “Signal” is most concentrated could help us design better signaling tools that don’t require the time commitment of a full-time DAO contributor.
Great framing — “governance debt” captures exactly why participation drops even when governance quality is high. The cost of staying informed scales faster than most contributors can keep up.
To your question: we haven’t broken down participation by topic area yet, but that’s a really compelling angle. If certain categories (CSM, treasury, protocol upgrades) consistently attract higher engagement, that tells us where the community’s attention naturally concentrates — and where signaling tools could have the most impact.
This is something we could explore in the DGI methodology — measuring not just how many people participate, but where participation clusters. A “governance attention map” per DAO would surface exactly those signal-rich areas you’re describing.
Would that kind of topic-level breakdown be useful from an operational perspective?
Absolutely. From an operational standpoint, a “Governance Attention Map” acts as a Resource Allocation Signal. If we find that Treasury and CSM attract 80% of the active signal while critical infrastructure upgrades (like 0x02 migration or V3 implementation details) suffer from “rational apathy,” it indicates a dangerous gap.
Operationally, the DAO could use this to trigger “Attention Incentives”—allocating more grant budget or delegate rewards specifically to those under-discussed but high-risk areas. It moves us from passive transparency to active governance steering.
I can only speak as a NO, we discuss a lot of stuff (also concerning DAO governance) only in a private group chats or closed calls. What is seen publicly is not the full picture. As I understand the governance processes in Lido DAO there’s also direct communications between delegates and the working groups, with little or without public exposure of the results.
In my opinion this is a problem for DAOs in general. It makes it hard to follow and understand processes and decisions.
This is exactly the kind of use case that turns analytics from “nice to know” into operational infrastructure.
“Rational apathy” is a perfect term for it — contributors aren’t disengaged, they’re just triaging. But if critical infrastructure votes pass with minimal review while treasury debates get all the attention, the risk isn’t visible until something breaks.
The idea of Attention Incentives is compelling — using participation data as an input for grant allocation or delegate compensation. That creates a feedback loop: measure where attention is thin → incentivize participation there → measure again.
We’re going to build this. The Snapshot data gives us proposal-level voting patterns, and with topic classification we can map exactly where attention concentrates vs. where it drops off. Lido would be an ideal first case given the diversity of governance areas (CSM, treasury, protocol upgrades, grants).
Would you or the operations team be open to giving feedback on an early version? Having real operator input would make the methodology much stronger.
@stefa2k
That’s a really important point — and honestly, it’s one of the biggest limitations of any on-chain/Snapshot-based analysis. We can only measure what’s publicly visible, and if key decisions happen in private group chats or closed calls, our data shows the outcome but not the process.
In a way, this reinforces the case for better public signaling. If the gap between public activity and actual decision-making is large, that’s itself a governance health indicator — not necessarily a negative one, but something worth making visible. The question becomes: how much of the “real governance” is captured in the public record?
This is exactly the kind of nuance we want the methodology to reflect rather than ignore.
The problem is that delegates end up voting on every proposal, even when it’s clearly outside their scope. Picking the right delegates who have the knowledge to judge each vote, and incentivizing token holders to delegate, feels like the only option.
@BCV
This connects directly to the attention map idea — if delegates vote on everything regardless of expertise, participation numbers look healthy but vote quality might not be.
The interesting metric would be: do delegates vote consistently across all topic areas, or do they selectively engage where they have domain knowledge? A delegate who votes on 100% of proposals might actually be a weaker signal than one who votes on 60% but concentrates on their area of expertise.
That’s a hard thing to measure purely from Snapshot data, but combined with delegate profiles and topic classification, we could start distinguishing between “informed participation” and “checkbox voting.”
Your point about incentivizing delegation is key too — the current model often creates a handful of super-delegates who become bottlenecks rather than specialists.
Quick update for everyone who contributed to this thread — @mizuki, @stefa2k, BCV, Jenya_K:
You helped shape something real. Based on the discussion here, we built a Governance Attention Map — a topic-level breakdown of where Lido’s governance attention concentrates, and where “rational apathy” creates blind spots.
Here’s what it looks like for Lido: chainsights.one/lido
Some of what your input directly influenced:
Topic-level participation breakdown — @mizuki, your “rational apathy” framing became a core concept. The map shows exactly where the signal is concentrated and where critical areas might be flying under the radar.
Visibility gap transparency — @stefa2k, your point about off-chain governance is reflected in how we present confidence levels. We’re upfront that this is based on public Snapshot data only — what happens in private calls and group chats isn’t captured, and that gap is itself worth making visible.
“Checkbox voting” detection — BCV, distinguishing informed participation from delegates voting on everything regardless of expertise is on our roadmap. Your observation shaped how we think about vote quality vs. quantity.
Data quality filtering — Jenya_K, the Etherscan labels approach is informing how we handle treasury and CEX wallet identification across all DAOs.
This is v1 — we’d love your feedback on whether the categories make sense for Lido’s governance structure, and if the attention distribution matches your operational experience.
As a thank you: we’d like to send each of you a full Governance Audit (our €149 report) — on the house. Your input shaped the methodology, so it’s only fair you see the result first. DM me if you’re interested.