A key requirement for a well-functioning DAO is user rights. Irrespective of what the tokenholders, delegates, or other governance decision makers do, users must always retain rights.
While there are some attempts like Optimism’s Citizen House, this is not enough. Another interesting proposal is Lido’s dual-governance: the users of Lido, i.e. stETH holders, participate in governance as much as LDO holders. Users of public protocols and blockchains are largely pseudonymous, and most people don’t really want to participate in complex governance protocols. They simply want to use the protocol, and if the governance misbehaves, have sufficient protections.
The key ingredient is mass exit or mass migration systems, in addition to timelocks. Users of a protocol should be able to exit at any time. To make this process easy/cheap, there should be batched exits to the host blockchain, or preferably, move to an alternative protocol. For example, if you don’t like what Aave V4 is doing, you alongwith thousands of other users can exit directly to Compound V4 or whatever. Likewise, with Arbitrum to zkSync Era. Currently, this isn’t really possible, but crypto users should collaborate to build consortiums and standards to make mass migrations possible. If this collaboration does not happen, at least the protocol should have mass exit systems where exits of many users are batched to keep costs as low as possible.
Another possibility is user vetos. So, users of the network can veto upgrades timelocked by governance, but simplified. For something like Lido it’s easy to be sybil-resistant - simply gauge by stETH holder usage. For something like a smart contract rollup, it’s a bit more challenging, but possible per historical usage patterns like fees spent etc. It’s important to keep this process as simple as possible - just offer users exit or veto rights and that’s it. If a proposal is vetoed, the governance simply collects feedback and we get a fixed proposal.
Needless to say, user protections will vary a lot depending on the type of protocol, but it’s imperative we push for efficient mass exit and/or veto rights. Once sufficient user protections are in place, it also lets us minimize governance further with less pressure. Minimizing or preferably eliminating governance is after all the key selling point of a DAO - otherwise it’s just another type of cooperative or public corporation.