Data-driven Decision Making for Autonomous Organizations
Decentralized Autonomous Organizations (DAOs) present a new opportunity for amorphous communities to organize and collaborate around shared goals and interests without the need for intermediaries. Accessibility of blockchain tech has allowed us to quickly prototype and experiment with organizational governance. The incredible digital connectivity we now enjoy has increased our ability to interact as well as to gather and process data, but our governance systems still rely on antiquated and centralized systems that have evolved for stability and predictability. While these legacy systems developed over hundreds of years to effectively manage assets and people, they are ill-suited to adapt to the technological revolutions of today.
How can we channel the wealth of data and augmented connectivity towards effective decision making among an amorphous community, such as a DAO? Several potentially powerful methods already exist, one of the most interesting and enduring is prediction markets. Prediction markets are effective at inducing people to acquire information, signal that information via trades, and collect that information into consensus prices to motivate a wider audience. Friedrich Hayek, the Austrian economist, described prediction markets as “mechanisms for collecting vast amounts of information held by individuals and synthesizing it into a useful data point.” Combining prediction markets with governance through creating markets for trading in decision making was proposed by Robin Hanson with a system he called ‘Futarchy’. In his proposal, elected officials would define measures of national wellbeing and prediction markets would be used to signal which measures would likely have the most positive impact. The tagline for Futarchy is ‘vote on values, bet on beliefs’.
An interesting aspect of prediction markets is the darwinian-effect it has on itself — those who are certain of their beliefs are willing to stake some value towards it, and if they are wise they will be rewarded and that value increases. Those who are not wise will see their value reduced and, eventually, will be selected out.
In the context of a DAO, and more specifically rLoop, a very simple version might look something like this:
- Members of a DAO create proposals that might introduce new governance policies for the organization, modify existing governance policies, introduce new projects for the DAO to allocate resources towards, or modify the development path of an existing project (to ‘evolve’ the project — more on that later)
- For each proposal, market predictors can indicate whether they believe that proposal will beneficially raise some (or several) metric(s) of the organization (revenue, profit, visibility, impact, effectiveness, sustainability, resilience, etc.)
- When those predictive markets clearly signal the beneficial characteristic of a particular proposal, that proposal is adopted and implemented by the community.
There is a commonly referenced example of the info successes of prediction markets in the case of the explosion of the Challenger space shuttle in 1986. Following the accident, there were no clear answers from the experts as to the cause, and the following day the Financial Times suggested “it will be months rather than weeks before NASA has any real answers to the question — What went wrong with the Challenger?’’. There were four main firms that were responsible for the manufacturing of the shuttle project — Rockwell International (the maker of the shuttle and its main engines), Lockheed (the manager of shuttle ground support), Martin Marietta (the manufacturer of the shuttle’s external fuel tank), and Morton Thiokol (the maker of the shuttle’s solid fuel booster rocket). While stocks in all four suffered immediately following the accident, none were as severe as Marton Thiokol which had a 1-day return of -11.86%, more than six standard deviations greater than the average daily stock return for the 3 months prior.
And while the other three firms tracked or outperformed the market in the following months, the decline of Marton Thiokol was sustained. The stock market had effectively identified the guilty firm in the Challenger crash by noon that same day (within 21 minutes of the accident itself), while it took more than 4 months of study by engineering experts and renowned scientists to definitively identify Morton Thiokol as the sole culprit.
A prediction market allows individuals to speculate on the outcome of real-world events. They of course are not perfect — but these markets do very well when compared to other information institutions. And they may provide the best way for decentralized organizations to channel our new-found wealth of data and augmented connectivity towards effective decision making.
