Wisdom in the Crowd: using CrowdWise consensus process

The New Economics Foundation is a wonderful organisation working practically and conceptually to enable us to rethink what our economy should do for us.  It calls itself a ‘think-and-do tank’. Amongst its many interests are participation and consensus-building as part of the renewal of democracy. It’s in that spirit that my near-namesake, Perry Walker (no relation) has developed the Crowd Wise tool:  a way of enabling groups to propose alternative solutions and find consensus using a combination of a slightly sophisticated voting system and discussion which allows people to take the aspects they like about a proposal and combine them to form new proposals. Sounds a bit complicated in theory!

It is much more easily understood when you try it out in practice, which is exactly what I did at the launch a couple of weeks ago.  You can try it out on 23rd September in London – see here - where our subject will be electoral reform.

Using a fictional example - the role of nuclear power

The launch was a mini-workshop where we were given some prepared options on the role nuclear power should play in a low-carbon, energy secure future.  (Of course, in a ‘real’ situation, we’d arrive at a discussion about a topic we had chosen to be present at and come with our own views which would then form the basis of the initial options.)

We were then asked to vote for the options in order of preference.  There’s a rather complex voting system, where you assign the options a preference (1st, 2nd, 3rd preference etc) although you are not obliged to rank all of them.  Depending on how many you rank, the ones you rank are assigned points.  For example, if you give a preference for five options, your 1st preference will score 5 points, your 2nd preference will score 4 points and so on.   If you decide to express a preference for only two options, your 1st preference scores 2 points and your 2nd preference scores 1 point.

The maths wizards may immediately see the significance of doing it this way: when the scores are amalgamated, it’s possible to see the degree of consensus.  In fact, the results are presented as a 'consensus coefficient', between 0 and 1.

In our nuclear power example, the results in the first round of voting varied between 0.19 (for an option based loosely on the views of the World Nuclear Association) and 0.59 (for an option based loosely on the views of Amory Lovins – demand reduction and a ‘soft energy’ path.  Since this was a demonstration workshop, we were then randomly assigned an option to brief ourselves about and represent.  We spent some time in small groups of (fictionally) like-minded people, understanding our option and discussing possible negotiating tactics. The groups were then mixed up and we had a chance to explain our option and discuss it with people who had different views.

Then came the negotiations!  This descended into horse-trading a bit, as we raced against time to find common ground with other groups.  In the end, the five options we began with were reduced to three.  One of these was from the original five, and two were new amalgams.  The consensus coefficients this time varied between 0.47 and 0.92.

The seemingly popular choice had elements that many of those supporting it did not like – perhaps this element of compromise is essential to consensus.  If we had had time for subsequent rounds, I think that more options would have emerged and perhaps what we would have ended up with would include a more precise understanding of the things that we really don’t agree about, as well as broader areas of common ground.

That’s a summary of the technical process.

Real-world example - AFC Wimbledon

We also had a fascinating insight into a real use of this tool as part of discussions about the strategic direction of a member-owned football club, AFC Wimbledon.  This process is ongoing.

The six options which the strategy group began with were generated by drawing on themes identified using a classic meta-planning technique, with the initial post-it brainstorm informed by gathering views from members and fans.

Options include “selling up to any sugar daddy who would build the club a 25,000 seater stadium” as well as something based more on the importance of the club as a community resource.


There was a very interesting discussion afterwards, as people who might well use this technique in practice explored its features.  We wondered whether it was in itself a decision-making tool, or a tool to inform a decision.  We agreed that the provenance of the options was important and needs to be clear.  It was also clear that the expertise and information about the detail behind the options, the nuances and assumptions, need to be ‘in the room’, in order for new permutations of options to be created and for well-informed voting.

NEF stress the usefulness of this tool in consensus-building, because of the in-built incentive to find common ground: your score only goes up if more people express a preference for your option.  This is the case even if the preference is quite weak.

In my group, I observed one person who was extremely keen on ‘winning’, i.e. crafting the most popular option.  This led to him being willing to include elements of other options which our initial option completely excluded, because this would increase the common ground.  I was uncomfortable with these ‘compromises’, but perhaps that’s because I was more committed to my (fictional) position than to finding common ground.  I’m not sure whether this is a strength or a weakness of the system!

Try it out for yourself?

Perry is running another taster session so you can try out Crowd Wise for yourself.  In conjunction with AMED and NEF, there will be a workshop in London on 23rd September, from 2.00 – 4.30.  It’s just £15 (£10 for AMED and NEF members).  Find out more here.


There's an interview with Perry on the Rhizome blog, here, and a description of Rhizome's use of the process (to help develop options for involving grassroots activists in organisational governance) here.

You can find case studies of CrowdWise in use here.