When consensus is dangerous

consensus-animateImagine a group of people who need to make a decision about something. As the size of the group increases, the chance that everyone will be happy with what is decided falls exponentially. Unless there’s unanimous agreement, the group will use — either explicitly or tacitly — some kind of rule that determines whether a specific decision is acceptable. Groups often use tacit rules when the consequences of the decision are minor [“Harry, you feel strongly we should do this but Kerrie & I don’t care either way, so let’s go with your approach”] or when “consensus” is only a pretense [“Well, I think we should do this. Any objections? OK, it’s decided.”]

So, you may wonder, if a group wants consensus, where consensus is defined by an explicit decision rule, then what decision rule should be used?

Danger, Will Robinson!


The moment we start trying to define consensus with a rule that tells us whether we have got it or not, we diverge from the core reason to seek consensus.

The value of consensus is in the process of seeking it — not a “yes, we have consensus!” outcome as defined by a decision rule.

There is no magic formula that will create the maximum likelihood that consensus, however we define it, will be obtained.

The best we can strive for is what Hans, Annemarie & Jennifer Bleiker, who have trained over 30,000 public-sector professionals over the last 40 years, call Informed Consent, which they define as follows:

Informed Consent is the grudging willingness of opponents to (grudgingly) “go along” with a course of action that they — actually — are opposed to.

The concept of consensus becomes dangerous when we use process that forces a fake “consensus” outcome on a group. An example of this is what is sometimes called 2-4-8 consensus, as quoted here:

2, 4, 8 consensus is an excellent tool for prioritising in large groups. This exercise will take time, but will help a group reach a decision that everyone can live with! It’s usually best to impose tight time limits at every stage of this discussion!

  • Draw up a list of proposals in the whole group.
  • Form pairs. Each pair discusses the list of possible proposals and is asked to agree their top 3 priorities (it could be any number, but for this example we’ll use 3).
  • Each pair then comes together with another, to form a group of 4. The 2 pairs compare their lists of top 3 priorities and, after discussion, agree on a joint top 3.
  • Each group of 4 comes together with another to form a group of 8. Again, each group takes its 2 lists of priorities and reduces it to one list of 3.
  • Repeat until the whole group has come back together and has a shared list of just 3 priorities.

—from Consensus in Large Groups

There is nothing wrong with using a decision process like this to pick top priorities in a group. But picking a group’s top priorities is not the same as reaching consensus.

You can’t please everyone
Seeking consensus, however you define it, is difficult for large groups. Techniques like Roman voting can help us determine how close we are to informed consent, and can pinpoint who cannot go along with a proposed decision and why. The journey towards informed consent is what we should concentrate on if we are to reach a “consensus” that everyone can live with.

Participatory voting at events: Part 2—Low-tech versus high-tech solutions

GMIC 2014 crowdstormingIn Part 1 of this series I defined participatory voting and we explored the different ways to use it to obtain public information about viewpoints and participants in the room, paving the way for further useful discussions and conversations.

There is no shortage of high-tech systems that can poll an audience. Commonly known as ARSs, Student Response Systems (SRSs), or “clickers,” these systems combine an audience voting method—a custom handheld device, personal cell phone/smartphone, personal computer, etc.—with a matched receiver and software that processes and displays responses.

Here are three reasons why high-tech ARSs may not be the best choice for participatory voting:

  • ARSs necessitate expense and/or time to set up for a group. No-tech and low-tech approaches are low or no cost and require little or no preparation.
  • Most ARS votes are anonymous; no one knows who has voted for what. When you are using voting to acquire information about participant preferences and opinions, as opposed to deciding between conflicting alternatives, anonymous voting is rarely necessary. (An exception is if people are being asked potentially embarrassing questions.) When a group of people can see who is voting for what (and, with some techniques, even the extent of individual agreement/disagreement), it’s easy to go deeper into an issue via discussion or debate.
  • Participatory voting techniques involve more movement than pushing a button on an ARS device. This is important, because physical movement improves learning. Some techniques include participant interaction, which also improves learning.

That’s why I prefer no-tech and low-tech techniques for participatory voting whenever possible. No-tech techniques require only the attendees themselves, while low-tech approaches use readily available and inexpensive materials such as paper and pens.

Wondering what no-tech and low-tech techniques can be used for participatory voting? Here’s a list, taken from a glossary of participation techniques covered in detail in my book The Power of Participation: Creating Conferences That Deliver Learning, Connection, Engagement, and Action.

Body/Continuum Voting: See Human Spectrograms.

Card Voting: Provides each participant with an identical set of colored cards that can be used in flexible ways: typically for voting on multiple-choice questions, consensus voting, and guiding discussion.

Dot Voting: A technique for public semi-anonymous voting where participants are given identical sets of one or more colored paper dots which they stick onto paper voting sheets to indicate preferences.

Hand/Stand Voting: In hand voting, participants raise their hands to indicate their answer to a question with two or more possible answers. Stand voting replaces hand raising with standing.

Human Graphs: See Human Spectrograms.

Human Spectrograms: Also known as body voting, continuum voting, and human graphs. A form of public voting that has participants move in the room to a place that represents their answer to a question. Human spectrograms can be categorized as one-dimensional, two-dimensional, or state-change.

Idea swap: A technique for anonymous sharing of participants’ ideas.

One-dimensional Human Spectrograms: Human Spectrograms where participants position themselves along a line in a room to portray their level of agreement/disagreement with a statement or a numeric response (e.g. the number of years they’ve been in their current profession.)

Plus/Delta: A review tool that enables participants to quickly identify what went well at a session or event and what could be improved.

Post It!: A simple technique that employs participant-written sticky notes to uncover topics and issues that a group wants to discuss.

Roman Voting: Roman Voting is a public voting technique for gauging the strength of consensus.

State-change Human Spectrograms: Human Spectrograms where participants move en masse from one point to another to display a change of some quantity (e.g. opinion, geographical location, etc.) over time.

Table Voting: A technique used for polling attendees on their choice from pre-determined answers to a multiple-choice question, and/or for dividing participants into preference groups for further discussions or activities.

Thirty-Five: A technique for anonymously evaluating participant ideas.

Two-dimensional Human Spectrograms: Human Spectrograms where participants position themselves in a two-dimensional room space to display relative two-dimensional information (e.g. where they live with reference to a projected map.)

And what are public, semi-anonymous, and anonymous voting? We’ll explain these different voting types and explore when they should be used in the third part of this series.