Last night Paul Mison, Chris Heathcote and myself talked about the Barcamp London 6 ticket debacle. The alleged 55 first round tickets were snapped up in 40 odd seconds. This meant that unless you were sitting primed at 3pm exactly you stood no chance of getting a ticket. Luckily I'd had sufficient warning and was ready, but a few seconds later and I'd still be waiting for a chance to get a ticket.
I'm not sure what this process will lead to in terms of a conference, but I know it has led to various upset people over the process. So, on my train ride home last night I started thinking about the advantages of foo over bar camps and vice versa. Foo has a roster of great people, it is hard not to have an interesting conversation at a foocamp, I'll admit bias as I've been a couple of times. Barcamp can be a great experience too, they are open lively events and give a real freedom to talk about something that interests you. However the ticketing process has become worse than first day of the christmas sales, particularly for the London event.
I believe the selection model has some benefits and that there is space for another model for un-conferences alongside the completely open barcamp and the private invite foocamp. There are many other models for normal programme based conferences, of course. The programme committee choosing the programme from submissions, or inviting speakers are both popular models for paid conferences.
Drawing on all of that, I propose the seed16 model for unconferences. The basis is a seed invite list, like a traditional programme committee. Each of these people can invite three attendees, each of these attendees can invite two people, each of these can invite a further one person. Each seed person leads to a total of sixteen people, three directly and a further 12 indirectly invites, plus themselves.
| seed invites | 3 who can invite | 2 who can invite | 1 each |
| 2 | 1 | ||
| 3 | 2 | 1 | |
| 2 | 1 | ||
| 3 | 2 | 1 | |
| 2 | 1 | ||
| seed + | 3 + | 6 + | 6 + |
| Equals 16 per seed |
It offers much of the benefits of the barcamp model and some of the selection inherent in a foocamp, but distributes the invites amongst the participants. Nearly two thirds (62.5%) of the participants get to nominate someone to attend, which feels pretty inclusive.
It would need a simple application to manage the invite process, so that someone did not get multiple invites. Invites should time out too, so that a full leaf of 16 can get invited, should people not respond to the invite. Linking to a service like eventbrite or similar for actual ticketing would make sense.
Managing the conference in blocks (a leaf) of 16 means that it is possible to grow the conference by increasing the size of the seed pool. Also it should be possible to bring in different communities by asking different people to start. Four ruby on rails programmers will result in a technical conference, a mix of web designers and developers would give a broader mix. You get the idea. Taking a theme for the conference helps give it focus, I'd also expect these to be small close knit events for 30-160 people or so. I can imagine seed16 London or seed16 UX NYC, both regional and subject based events or even a mix.
The change here is purely in the setup of the participants, by getting the participants to largely select themselves you will get a cohesive group of people if you start with mutual friends, but by starting with a more dispersed set of people you'd get plenty of scope for serendipity. From there it would run as a normal unconference with people nominating themselves for sessions. The reboot model of voting up sessions appeals too and would be easy to add.
I'd be interested in trying this idea out later in the year with four or five seed people (64 or 80 attendees) and see how it works in practice. Fixing a date and a venue upfront would be important and to make the event self supporting I'd suggest charging an admission fee of say 10 pounds, with any spare money going to a post event bar tab. No date and venue as yet, probably in the summer when it is a bit warmer.
I'm sure seed16 has plenty of issues with the model, but running fast and cheap single day conferences should be easier than it is. Interesting and similar events show it is possible to run single day conferences successfully. I think the spreading invite model will avoid the zero day rush for tickets and the invite a friend model means it is less exclusive than the foocamp model. Distributing the invites also lessens the burden on the conference organisers.
Interested?
Building Social Web Applications by Gavin Bell.
I love this idea! Going back the your insiders versus insiders point from last night: I wonder how it would change the social dynamic? Everyone there has been personally proofed by someone no more than a couple of steps removed from them. Or would this model limit diversity? I suppose you could reduce the invites per person and increase the total allocation to generate depth... Either way you should definitely try it out.
Hi Gavin
I agree, that's a nice setup.
@terry thanks
@Mike, the scale would probably avoid the them and us feeling that happens at larger conferences. Certainly playing with the invite parameters would be possible to deepen the tree, as a 2 2 2 1 for example.
My problem with this is that you're going to end up with the same "A List" people being invited over and over again. It's fine if you're on that list or on the periphery I suppose except that's going to make the idea pool stagnate and the conversations moribund.
The problem is that you need a certain proportion of known "faces" in order to attract people. And also, at least for me, the choice of whether to attend an event is highly dependent on which of my friends is going to be there.
It reminds me of the same problem Glastonbury has - how to make it fair (for which is the obvious solution is a lottery) whilst also acknowledging that people wish to travel in packs.
For Glastonbury I played around with a system by which people could bid as a group - the size of the group affected the chance that they'd all get tickets. I played around with a few factors - should the size of your group increase or decrease the chance of you all getting in? Should it have no factor at all? Does having more large groups on site give a better vibe to the festival or does it discourage inter-group personal interactions?
I gave up on my simulations when I started writing genetic algorithms but you get the point.
Simon
You are right in that is a potential weakness. However I think it is manageable through careful seeding, an open acceptance list and encouraging people to not invite just the usual suspects. I'd also hope that this could be a regular event rather than a couple of times a year which is about as rapid as most barcamps get.
This is trying to find a middle ground between a fully selected invite list and a free for all. Mind you I might have picked something akin to the House of Lords reform and that has been going ten years.
Also picking different starting points will result in a quite different conference, say you, Mike, Terry and Chris Thorpe would produce a different list of people to Matts Jones, Biddulph, Webb and Patterson. There would be some overlap and that is what you are getting at I think, but it is sort of unavoidable if you want the faces.
Maybe some modelling is the way to explore this or just trying it out.
thanks