Methodology

A DAO can look alive on its own website and be quiet or abandoned in reality. Four things, taken together, distinguish the two: whether decisions are still being made, whether anyone participates in them, whether the treasury could survive the organization's own token falling, and whether power is spread widely enough that the votes mean anything.

Each sub-score is computed only from figures that can be read from a public source. Where a figure cannot be sourced, it is not shown and the sub-score is excluded.

Each is scored 0–100 and contributes the weight shown.

The status label is derived from the score plus time-based rules, applied in this order.

• A newly listed organization is “Not enough data” until it has 90 days of history, either observed here or retrievable from a source.
• An organization with a large treasury and no governance activity is Dormant, and its page says so. Money alone never makes an organization Active.
• If a source is temporarily unreachable, the previous figures are kept and marked “unchanged — source unavailable”, rather than showing a hole or a wrong zero.
• A sub-score that cannot be computed is excluded, and the remaining weights are re-normalized to add up to 100%. The organization's page states which sub-score is missing. An absence of data never contributes a zero.
• An organization is scored on whatever can be measured, even where that is a single dimension. Where only one of the four sub-scores could be computed, the profile says so in a notice above the breakdown — the score is real, but it rests on one thing rather than four, and that is worth knowing before quoting it. Just over half the directory is in that position — 646 of 1,181 organizations — almost always because no treasury could be found and no quorum is set, leaving activity as the only measurable dimension. Where that single dimension is activity and the organization has proposed nothing in 90 days, the score is 0, and 669 organizations currently sit there. A 0 on this site means “nothing observed on the one thing we could measure”, not “measured across four dimensions and found wanting”. The notice on each profile says which.

Every figure comes from one of these. All of them are free and public, and each organization's page links to the exact record each figure was read from.

On-chain-only organizations that govern through Tally are not covered. Tally's terms of service prohibit republishing figures derived from their data without written permission, and daohaus.com does not have it. An adapter exists and stays switched off until that changes.

Every organization is refreshed once per day. Each page shows the date its figures were read, and the Status page shows whether the most recent refresh completed.

If no refresh has completed for 48 hours, every page carries a banner saying how old the figures are. Pages keep working with the last good data.

Daily snapshots are kept, which is what makes the 30, 90 and 365-day trends real rather than estimated. History is never edited except to fix an acknowledged error, which is then recorded on the Changelog.

• Participation is measured as quorum attainment, not as a share of eligible voting power. The brief for this site defines participation as the share of eligible voting power that actually votes, but no free source publishes eligible supply per proposal. What is measured instead is whether the typical proposal clears the bar the organization set for itself. Where an organization sets no quorum, participation cannot be computed at all and is excluded.
• Treasury values move with token prices. A treasury that has not been touched can still change value substantially from one day to the next.
• Only treasuries that a source knows about are counted. An organization holding assets in wallets it has not published will show a lower treasury than it has, and its page says that no treasury was found rather than reporting zero.
• Off-chain activity is invisible here. An organization deciding things in a forum, a call or a private channel looks quiet to this site, because a public vote is the only decision it can observe.
• Delegation hides real voters. An address voting with delegated power is counted as one voter, which understates how many people are behind a decision and overstates concentration.
• Vote concentration is sampled from the three most recent settled proposals, not from every vote ever cast.
• Multi-chain treasuries are only as complete as the sources' own chain coverage.
• Automated proposal summaries are generated from proposal titles by a language model, are labelled as automated wherever they appear, and can be wrong. The original is always linked.

Every organization's page has a link to report an error. Submissions become public, numbered items on the corrections tracker, and are answered within 7 days.

While a correction is open against an organization, its page shows a notice saying that figures may change. When the correction is resolved, the outcome is recorded on the Changelog.

There are no paid corrections and no way to pay for a change to a score.