Most cryptoenthusiasts don't understand that decentralization is not always the best answer to everything. They want to decentralize everything just for the sake of it.

I think there is a good case for centralizing the decision making process of decentralized/distributed autonomous organization (DAO).

The existence of a DAO has to be decentralized. There must be lots of copies of the blockchain in different computers around the world. This makes it resilient and impossible to stop. But this is completely different thing than decentralization of decision making power.

The decision making power of DAO can be and in many cases should be more or less centralized to make it work more effectively.

I know some purists want to emphasize that a DAO needs decentralized governance, so from now on in this text the D in DAO stands for distributed.

Learn from Bitshares

I consider Bitshares 2.0 as the most advanced DAO so far. It's not perfect but it has a lot of great innovations.

Good:

• Delegated proof of stake (DPOS). In a normal proof of stake (POS) system those who have most core tokens can produce most of the blocks. But in DPOS block producers are employees of the blockchain and shareholders have to vote who are hired to do the job. This has some great advantages like scalability. Blocks can be produced much faster than in proof of work (POW) and it's better for the environment (POW wastes huge amounts of electricity to produce blocks).
• A blockchain should have a way to pay for it's own development. Bitshares does this by having workers who are hired by shareholder voting. If somebody wants to do something for the blockchain (like to develop a new smartcontract, fix bugs, write documentation, develop libraries, etc) they publish their worker proposal in the blockchain and let shareholders vote for it. If shareholders feel that the proposal is worth paying for, they vote for it and it gets funded. I'm not very familiar with Dash but I think it has similar system.
• The reserve pool for funding. It's highly undervalued innovation. Bitshares doesn't fund everything with inflation nor make it's shareholders pay for everything with their own money. Instead it has a fund of BTS (the core token) from where all the employees are paid, including witnesses (block producers). Bitshares earns BTS from customers as transactions fees. Earned BTS goes straight to the reserve pool (instead of going to block producers, like in many other cryptocurrencies). So it's is like a company account – when it gets bigger, company is making money – customers are paying for the services more than it costs to produce them. When it gets smaller, company is losing money. Although at first it will of course lose money when the products are still under development and there are only a few customers.

Bad:

• Everybody has equal say to decisions (according to their balance). Bitshares has too many low quality shareholders which hinders it's decision making process. Too much time is spent to discuss how to achieve consensus and too little to actually make services that customers want and are ready to pay for. There is no way of nullifying somebody's voting power.
• Voter apathy. Only a small minority is actually interested in daily activities of Bitshares. Those who don't care to vote actively shouldn't have any voting power.

My proposal for a more efficient DAO design

One big inspiration for this text is the success of Steem. Currently Steemit and the developers hold most of the steem. This has made them able to do hardforks whenever they want. The development has been amazingly fast when no time is wasted to ask everybody's opinion and then discussed everything thoroughly. Consensus of the broad community is an overvalued concept.

It's pretty clear for me that a DAO should work more like a traditional company with a board of directors. Shareholders can influence only in the annual general meeting and the rest of the time directors can act quite freely as they see best. Shareholders of a company don't need to know everything that the company is doing. Likewise, most of the shareholders of a DAO shouldn't need to care about day-to-day decision making.

Probably a DAO doesn't need to copy company structure fully. For example, it's hard to see any use for the role of CEO.

The goal of a DAO is not to make everybody feel that they have a fair say to everything. The goal is to make a great product that people want to use. A product that will answer to a demand. A product that actually works.

Many people in the cryptosphere have been too fixated to thinking that every shareholder or user should have equal right to take part in the decision making process. This is dangerous because having too much discussion usually is just a waste of time. Not everybody is qualified to have a say. Too many times we have seen uninformed or badly motivated individuals to spam the discussion with their useless comments.

Decision making power has to be in the hands of a small elite

There should be two different core tokens. Or maybe it's better described as one core token with two states: active and passive.

The passive core token is like a bitcoin or similar cryptocurrency. It can be transferred to others without restrictions but it doesn't have any voting power.

The active core token is like Steem Power in Steem. It's nontransferable (the blockchain prevents it's transfer from one account to another) and it gives the right to vote in the blockchain.

Core token owners can be called as shareholders and active core token owners as the executive team.

Initially all core tokens are passive. Users can change them to active (similar to powering up in Steem). This should probably cost a little bit to discourage users to change them just for fun. Only those who seriously want to have a say in things should have active shares.

Changing shares from passive to active can be slowed down if necessary. The executive team might want to avoid drastic changes to keep decision making process more stable and predictable. When a passive shareholder wants to become active, he will go to a queue. The queue is cleared gradually by changing the shares of the top account. How fast this is done can be adjusted as a blockchain parameter.

Active shares can be reverted back to passive. User can do it himself, for example if he feels he doesn't have enough time to take part in decisions. Majority of the active shareholders can also decide to change somebody's active shares to passive if they feel that a particular shareholder is causing unnecessary conflict in the executive team, is no longer trustworthy or acts harmfully in some other way.

To make active shareholders to have more skin in the game, their shares should stay nontransferable for a while (maybe something like a week or two) after they have been changed from active to passive. This creates an incentive to consider more of the longer term prospects of their actions.

If the executive team does something that passive shareholders and customers don't like, the price of shares will go down. Having some extra time in the process of changing the shares from active to passive makes active shareholders to think more of the consequences of their actions. They will have to let passive shareholders to sell first.

It's important that those who have power to make decisions have also more skin in the game. If something goes wrong, they will suffer more than those who don't have the power. If somebody has more ideas how to make powerful shareholders to have more skin in the game, I'm very interested in hearing them.

If active shares aren't used for voting regularly, they will be automatically converted to passive shares. This makes sure that only active people have power. If somebody dies or loses private keys, the DAO won't become dysfunctional.

Votable things for the executive team:

• How the blockchain is spending money: block producers and other employees, positive contributions for the ecosystem (like documentation and bug fixes).
• Blockchain parameters: block size and time, transaction fees, how fast passive shares can be changed to active.
• To change somebody's active shares to passive.
• To drop somebody at the bottom of the queue of users who want to become active shareholders.

It's possible to let passive shareholders to vote too. For example a DAO could do something like Steem is doing, which requires ordinary users to vote how to allocate content creation rewards.

Owners of passive shares can enjoy the price going up without responsibility to make any decisions and with freedom to sell whenever they want. If people want to invest, they can just buy the core token and focus on other things – voter apathy is not a problem when you don't have voting power.

Funding and employees

In my opinion, DPOS is the best way to decide who will produce blocks. It's a job that a blockchain is paying for, so the blockchain should have a say who can do it. If a block producer does a bad job or is in some other way untrustworthy, blockchain should be able to fire him. This will also make sure that there is enough decentralization among block producers.

Funding for employee salaries comes from the reserve pool. The reserve pool should be large in the beginning to make sure that DAO has enough resources to fund initial costs of development and user acquisition.

When hiring employees is done by the executive team, employees can be sure that their funding is stable and isn't cut for stupid reasons in the middle of the project.

Salaries are paid in passive shares. If a worker is already an active shareholder, he could receive his salary also in active shares. He can define this in his worker proposal.

DAO can also use a referral system. When user brings in a new customer, he could get a share of the profits that the new customer creates for DAO. It's best to pay only based on actual profits, but of course it's possible to do a referral system which is funded from the reserve pool. For short term campaigns it might be easier way.

Initial coin offering (ICO)

ICO (or maybe it should be called ISO, initial share offering?) should be only for selected people who have a good reputation. That way it can be made sure that DAO will get frictionless start.

If a DAO wants to be successful, it need to has a core team that has an ability to make decisions fast. When the core team can work together effortlessly, has a great vision for the DAO and technical ability to actually do everything that needs to be done, it has all the chances for success. I don't think this can be achieved if there is public offering of shares for everybody. If the offering is public without any reputation requirements for buyers, it will open the doors for low quality shareholders. This is dangerous for a new DAO.

In the beginning, most of the shares should be in the reserve pool. This makes the system fairer because the people who take part in the ICO will get only modest profits when the price goes up. Most of the shares are allocated from the reserve pool to people who actually do something useful for the DAO. Probably at first many of the active shareholders will act also as block producers and employees.

Resiliency

Most of the cryptoenthusiasts seem to have an opinion that more decentralization means more resiliency. But by limiting the amount of decision makers, DAO can be more efficient and faster to adapt. I think that is very undervalued feature of a DAO. If it can't adapt to changes fast enough, it's not resilient.

Most important thing for resiliency is that the existence of blockchain is distributed widely. DAO can help this by paying more for block producers and nodes. Incentives for block producers and node maintainers should be something that DAO can easily change when necessary. If there is not enough of them, it can pay more. If there is plenty of them, it can pay less.

Potential usecases

• Financial platform (like Bitshares was supposed to be). All kinds of useful smartcontracts that create a lucrative ecosystem. Optimized for scalability and speed. The executive team has to make sure that smartcontracts are actually high quality and useful, and drop everything that doesn't have demand on the market. Having an efficient management team helps to make changes fast to answer the demand. So far Bitshares has been too slow to make changes to satisfy the needs of customers.
• File sharing platform for magnet links (or some other suitable technology, like IPFS). Lets users to decide what is rewarded like it's done in Steem (by upvoting). Users can also comment and give feedback that creates reputation score for content publishers.
• Anonymous currency. As far as I know all current anonymous cash projects are proof of work (Zcash, Monero, Dash) which means they are environmentally destructive and will fail eventually. With DPOS it's possible to create fast payment network for anonymous cash. Having a dedicated and powerful executive team makes it easier and faster to handle all security issues that might pop up.

If you want to create a new blockchain project, you should ask first: What the market really wants and needs? Then design the product or service and after that design a DAO that can make it happen.

Majority of blockchain projects currently are done in the other way around: first the blockchain is created, then some service is designed in top of it, and then it's pushed into the markets and wished that there is somebody who is willing to to pay for it. I think we have seen too many failed projects already, it should be clear by now that a new kind of approach is needed.

"This is a scam because it's so unfair!"

If you think the organizational model of the DAO is unfair, then don't use it. It's the same thing that happens every day in normal business. If you don't like the organizational model of some company, you don't invest to it or buy products it makes. Of course customers have usually zero interest in the organizational model. They just want to buy great products.

When the model that I have described here will be used for a DAO, it will be much easier to see whether or not it's a scam. First thing should be the service itself – is the DAO going to make something that market actually wants? Second thing should be the organizational model – does it function effectively so it can actually produce the service it promises?

DAOs should be thought as businesses where the results are the most important thing, not as hippie communes where the most important thing is to listen to everybody and respect everybody's opinion.