The Governance Gap: Measuring DAO Participation

Governance dashboards report turnout, proposal counts and voter numbers. These are easy to compute and close to uninformative, because they measure activity rather than influence. The gap between a DAO’s described decision process and the handful of addresses that actually determine outcomes is, in our experience, the most under-reported thing in the sector.

Why the standard metrics mislead

Turnout expressed as a percentage of circulating supply flatters almost every DAO, because a small number of very large holders can produce a high figure without any broad participation at all. Proposal counts measure activity, not consequence — a DAO passing forty routine parameter changes is not more governed than one passing four significant ones. And unique voter counts treat an address holding a thousand tokens and one holding ten million as equivalent participants, which is the opposite of how the vote actually worked.

None of these are wrong as facts. They are simply answering a question nobody needs answered, while the question that matters — who decides — goes unmeasured.

Decisive voters

The most useful single measure: for each passed proposal, how many of the largest voting addresses, removed from the tally, would have changed the result?

Frequently the answer is one or two. A proposal can show hundreds of participating addresses and still have been decided entirely by two delegates, with everyone else contributing rounding error.

Computing this across a DAO’s proposal history produces a distribution, and that distribution is the honest description of where power sits. It is straightforward to calculate from public voting records and it is almost never published, presumably because the number is uncomfortable.

Delegation makes concentration invisible

Delegation is a sensible response to voter apathy — most holders have neither time nor context, so they delegate to someone who does. It also concentrates decision-making while improving every headline metric.

Turnout measured by tokens looks healthy. The number of entities actually deciding may be very small. Both statements are true simultaneously, and only the first appears on the dashboard.

The measures worth tracking: how much voting power sits with the top five delegates; how often delegates vote together; how often a delegate has voted against the apparent preference of those who delegated to them; and how much delegated power has never been re-examined by the delegator since it was assigned.

Untested quorums

Most DAOs have a quorum threshold. Many have never approached it in a contested vote, either because proposals pass overwhelmingly or because contested proposals are withdrawn before voting.

A threshold that has never bound anything is a parameter, not a safeguard. The question worth asking is whether a proposal has ever failed for want of quorum, and what happened next — the answer usually reveals whether the mechanism has any teeth.

The filtering nobody measures

By the time something reaches a formal vote it has usually been through forum discussion, informal signalling and, very often, a private conversation among the largest holders. Proposals that would fail are typically withdrawn rather than defeated.

This means on-chain voting records systematically overstate consensus. Near-unanimous passage rates are frequently evidence of effective pre-filtering, not broad agreement — and the filtering is where the real decisions happen, invisible to any dashboard.

The tell is the ratio of proposals that reach a vote to proposals that were raised. A DAO where almost everything raised proceeds to a vote and passes is either unusually aligned or has an informal gate.

Treasury composition as a governance signal

A treasury held almost entirely in the protocol’s own token is a governance fact as much as a financial one. It means the DAO’s capacity to act is correlated with the thing its decisions affect, and that diversifying requires selling into a market that will read the sale as a signal. Boards facing that constraint tend to defer decisions, which is itself a governance outcome.

What honest reporting looks like

Decisive voter counts per proposal. Delegate concentration over time. Proposals raised versus proposals voted. Whether quorum has ever bound. Treasury composition and runway in non-native assets.

None of this requires special access — it is all public. It is simply less flattering than turnout, which is why turnout is what gets published.