You start a poll to ask members if they would like to expand the community’s field of interest to include new (relevant) trends and receive 51 replies. 26 say yes, 25 say no.
What should you do?
The Problem With Consensus
How you process this information depends very much upon the social architecture you’ve designed to help groups make decisions.
Most of the time, we default to majority regardless of whether that’s the right approach. You have four broad approaches here:
1) Uniform/unanimous consent.
Everyone has to agree (or at least not disagree) for the group to move forward. This is great for having everyone buy-in to decisions. It can also be incredibly tedious to get things done (United Nations and EU bodies are good examples).
This works best in smaller groups. If you’re deciding what movie to watch or which restaurant to go to, you usually aim for something nobody hates. Not everyone gets his or her favorite choice, but no-one specifically disagrees. Most juries are required to give a unanimous verdict, for example.
2) Majority consent.
A majority is 50% + ‘X’. You can set the ‘X’ to whatever number you like.
There are different types of majority consent. The most common is a simple majority +1 (51%). This is great for fairness and getting things done, but it can upset a large number of the group and often prejudices against minority groups or perspectives. Would you want to move forward with an idea that 49% of your group don’t support?
For the bigger decisions (and for smaller groups) you probably want to set ‘X’ to a higher number than 1. A good example here is a supermajority. This is where 2/3rds (or X = 16.67%) support a decision. For example, the senate and congress need a supermajority to override a presidential veto.
The upside of a supermajority is less people disagree with an important decision. They’re less likely to leave, create their own groups and ferment dissent. The downside is it becomes harder to reach the barrier to take action.
However, imagine if your community has 2000 members and you receive only 51 votes. That represents 2.6% of the group. What if one-side campaigns covertly among their members (or outsiders) to vote for their favorite option? Ron Paul supporters are notorious for rigging online votes in favour of their candidate.
This is why sometimes we use an absolute majority. An absolute majority requires a majority of the total membership to vote in favour of the action. Most unions have a voting threshold to take action. The problem is this makes it much harder to take any serious action due to voter apathy on most issues. The upside is it reflects the entire group.
3) Plurality consent.
Plurality occurs when there are more than 2 options. Here you go with whichever option has the most votes. The upside is speed, simplicity, and fairness. The downside is most popular opinion can be split between two similar options – thus the option the most people want the least can sneak in.
To tackle this problem, some systems (e.g. deciding the Olympic host city) use the exhaustive ballot. If no option has a majority (51%) after the first round of voting, the least popular option is removed and another round of voting begins. This ensures the winner reflects the view of most people; however it requires several rounds of voting which quickly becomes tedious.
A similar option is the instant-runoff vote. Here voters rank the options in order of preference. If no option wins a majority, the least popular option is removed and those votes are transferred to the next order of preference repeatedly until you have a winner. People now only need to vote once, but it becomes more difficult to manage and explain the system (although some software exists).
Other systems (e.g. electing a pope/nominating a party’s presidential candidate) don’t remove the least popular option, but repeatedly poll members until one achieves a majority (or for a papal conclave, a supermajority). This tends to be a tedious process. The 1924 Democratic National primary, for example, featured an impressive 103 rounds of voting. The longest papal conclave lasted almost 3 years.
4) Executive authority.
This is where a single person (hopefully) listens to reasoned arguments and then makes a decision. The benefit of this should be obvious. You can move quickly to action. It also lets you take actions that need to be done but are unpopular. You can’t host a poll and then not take the result.
Most business leaders work this way. They don’t poll the entire company for their view. They listen to arguments and make an informed decision. This occurs more in times of difficulty too. Some constitutions allow for heads of state to get emergency powers in times of war.
If the upside is speed and efficiency, the downside is unpopularity, abuse, and prone to errors. The decision-maker might make unpopular decisions, only seek opinions that validate her pre-existing bias, or use their influence for patronage. They might lack the ability to challenge the advisors and make the right decision.
p.s. read Christopher Allen’s terrific Spectrum of Consent Post for a fuller, more detailed, and much better list than mine).
The Social Architecture You Design Changes The Outcome
The process you use changes the entire outcome.
If your answer to the opening question was to go with the majority, you now have 49% of the group who are unhappy. They might leave or start their own community (the biggest threat to most communities is usually from within). This could be a bigger danger than the declining sector. Worse yet, their popular opinion is the wrong option. They’re condemning a community to death.
The challenge then is to create the right decision architecture to get the right outcomes.
An alternative approach is to create a range of options. For example:
- Yes, expand to all relevant fields.
- Yes, but only expand to one new field.
- No, don’t expand at all.
Now the ‘yes’ vote will be split among two options. If option 1 receives 10 votes, option 2 receives 16, and option 3 gets 25, option 3 wins – regardless if it’s what most people want. This is also an unfortunately effective way of rigging a vote in favour of your preferred option.
Instead you might start a discussion seeking reasoned arguments for and against it. Then you review the arguments and you decide which has the greatest merit. Here your group feels consulted, although you might still pick the unpopular option that is best for the group. This only works if members of the group can bring new, unique, expertise to the table – not opinions on whether they like or dislike the idea.
In situations like these where you know the change has to be made (you have unique access to trends, legal advice, resources, or an ethical viewpoint), then asking people if they want the change or seeking reasoned arguments is a waste of time. You have to take the decision for the benefit of the group.
This isn’t bad. You just need to ride out the criticism. Facebook makes a lot of unpopular decisions (i.e. creating the wall) that pay off over the longer run. They have access to insights their audience doesn’t.
In these situations, don’t ask people if the change should be made; let them make decisions on how to make the change. Which sectors should you expand to first? Should they intermix among existing discussions or be involved in new ones?
This is presumptive decision-making. You begin the discussion after the key decision and engage people in the process to gain ownership.
Which Option Should You Use?
If all of the above sounds rather complicated, here’s a simple decision architecture cheat sheet.
Most of your decision making today is far more informal than this. Yet, even informally, the process matters a lot. In your mind, in any group discussion, are you informally seeking unanimity, majority, plurality, or just opinions so you can make a decision for the group?
If you figure that out, you can probably work out the next steps.
p.s. We’re now just 43 days away from our FeverBee SPRINT event. If you want to learn a lot of advanced community skills from 14 world-class experts and ourselves, I hope you will join us at: http://sprint.feverbee.com.