A better way to allocate tickets?

[crtr0]

The sole purpose of any conference is to match the right speakers with the right attendees. That is when magic happens and communities truly grow and become stronger.

However, there is an issue when demand for a conference outstrips supply of tickets. CascadiaJS (like many JSConf conferences) strives to create a deliberately intimate environment for its attendees and speakers. This means smaller venues and fewer tickets.

Our current solution is to sell multiple batches of tickets during documented windows of time. The earlier you buy a ticket (before speakers and other details are announced) the cheaper they are. If you're not on your computer, attempting to buy a ticket during the prescribed window, they will likely be sold out by the time you show up. This is our attempt to allocate tickets to those that care the most about attending. We are using "attention" (attention to email, attention to Twitter, setting reminders, etc) as a proxy for desire to attend the conference.

But perhaps there is a better way. Some have suggested a lottery would be more fair. What do you think?

[beaugunderson]

One thing I missed is that today's batch of tickets was documented ahead of time (even though I thought I had been paying attention).

That said, I think the documented window approach works against those who have last-minute work emergencies, find themselves traveling during the window, or otherwise have limited access to the Internet during the window. I also realize that if you "really" want to go the conference you may be motivated to work on a contingency plan, but I also think it's important to think about the concept of privilege and how that may affect things too. If there are developers with less privilege who still want to go to the conference they may be at a disadvantage if they have greater constraints on when they can be available to purchase tickets. One option may be a pool of tickets specifically for those cases.

My original question was whether there are compelling reasons not to do a lottery. There very well might be, I just want to know what they are. :)

[grahamb]

I'm not concerned about the process, per se; first-come, first-served seemed to work well enough. I am curious why ticket prices were not published ahead of time. For those of us who need employer approval, advance knowledge of ticket prices is important if we want our request approved. I'm fortunate that my boss was OK with "it should be less than $400", but others may need the exact price.

[apike]

This has been debated by the organizers of any popular conference to great degree. Which compromises they make to allocate tickets have a big impact on the feel of the conference.

• WWDC does a seconds-long first-come first-serve scramble
• TED blends invites with extremely high prices
• SXSW just sells as many tickets as people will buy, diluting the conference to irrelevance
• Burning Man tried a lottery, and had big problems (splitting up groups, people gaming the system, etc)
• JSConf usually does a scramble, but last year did a lottery (with maybe some invites mixed in, hard to say)
• Paul Campbell (FunConf, Brio, etc) seems to take the approach of intentional obscurity: when his conference gets too popular, he retires it and founds a new one
• XOXO takes applications and invites people that fit its goals
• Çingleton opens sales to previous attendees and a few invites, and then does a first-come scramble for the remaining few tickets

Any popular conference will be criticized by some for its approach to allocating tickets. What compromises you're forced to make will go a long way to defining the conference.

[beaugunderson]

Isn't "scramble" another way to say "lottery for those who happen to be at their computer at the time"?

I also feel like it's possible to make a lottery support groups (sign up as a group, have that count as one or some number of lottery spots).

There's also Google IO, which is an hour long scramble/lottery for people who wake up at 6am (which I think is an incredible waste of everyone's time).

[apike]

Indeed, a scramble is like a lottery amongst a subset of potential attendees: people who care enough about the conference to be sure they're at their computer and are able to do so. In a scramble, more people who care deeply get tickets than in a lottery. The main price the conference pays for having more attendees that care deeply is resentment. There is also collateral damage, by way of people who care deeply who can't participate in the scramble due to being unlucky. In theory, those spots will be taken by others who care deeply, and the conference as a whole does okay.

Formal lotteries cause resentment too, especially from people who would probably have gotten a scramble ticket but end up not getting a lottery ticket. Lotteries are easier to game and harder to administer than scrambles, but that's not as big an issue for smaller conference. The longer the lottery application window, the more folks will apply for a ticket on a whim, or because they are somewhat interested in the conference, or because they haven't used their yearly conference budget and they might as well.

In the end, the length of the "lottery" (instant, hours, days, weeks) is a tradeoff between having the most passionate attendees and not being resented.

[beaugunderson]

I would amend "people who care enough about the conference to be sure they're at their computer and are able to do so" to add "and have the privilege to be able to do so" because I think lotteries are also different in that dimension.

I think requiring payment information at the time of lottery signup might mitigate people signing up on a whim.

I like the idea that there is a tradeoff between passion and resentment, but I think there are other dimensions that are worth considering.

[notmatt]

Someone, probably Chris Williams, said at JSConf 2011 that the "right people" were by definition the ones in the room. I think that's worth remembering.

So, does a scramble really work as a good proxy for interest? It's not terrible, but it's not great, and the problems with it are (I would hope) pretty evident. Just to toss out one example, imagine two people:

1. early-bird buyer, expensing ticket, travel & accomodation to their employer
2. late buyer, out of pocket, waited until speaker details were available

The current "refresh the web page" proxy for interest asserts that attendee #1 is most interested in the conference. That can't be ruled out -- having a conference-friendly employer doesn't mean they aren't interested -- but it's clear that these two people are making very different decisions, with different personal consequences, and relying on very different information to do so. Without knowing them personally, we can't actually assume anything about the amount of passion and desire to attend they actually have. In short: our proxy has errors built-in. Denying this imprecision is seductive ("the people who care enough to...."), but ultimately self-defeating.

I'd point to the JSConf "significant other track" (and even recent JSConf locations at family-friendly resorts) as a concrete way to make the choice easier for some set of people. There are lots of other solutions, addressing a lot of other problems, but what's clear is that better ways are possible.

[crtr0]

This is a great debate, but so far I've only heard two options:

• The current system (first-come, first served)
• A lottery

I may have missed another one, please let me know if I did. Are there any other systems that you all think we should be looking at?

[rockbot]

I'm sure someone will correct me if I'm wrong, but JSConf 2013 had both an early-bird round as well as a lottery. The purpose was two-fold:

1. Those who were super duper OMG I MUST GO interested were able to click like crazy during the early-bird round
2. Everyone else has an equal opportunity to get a ticket/not get a ticket.

In the conversation re: privilege, the JSConf method tackles it a little bit, whereas some other conferences will set aside a certain number of tickets for people who might not necessarily get to the computer in time*. These tickets are dispersed via a variety of methods, i.e. meetups, organizations, etc.

* Could be they're on travel/away from their computer/etc; could also be that they don't even know about the conference, but would be a fantastic fit for it.

I think a good system would involve all of these methods.

(True, some people will be annoyed/feel excluded, but there will always be folks who feel that way about, well, pretty much everything.)

[notmatt]

@crabasa - there's a list of options in @apike's post above. Any of them could be applied to all or part of the process of obtaining a ticket.

Anyway, some concrete suggestions, not all mutually compatible:

1. Early-bird tickets move to a lottery, other rounds stay the same. They are always over-subscribed, and I think that willingness to pay before any conference details are available is a good marker that someone is "interested enough" in attending, any further distinction in "passion" is probably neither measurable nor important.
2. vice-versa: early-bird is a scramble, the remainder are a lottery (I believe that was JSConf's model this year).
3. Eliminate multiple purchases - if the proxy for interest is the willingness to refresh a page at the right time, then let's make sure each attendee is actually willing to do it. Personally, I think this is a bit of an asshole move, but it would be a solid commitment to the idea that the scramble is a good substitute for real interest/passion.
4. Take attendee applications (or take a survey of early-bird buyers or some other sample). I have every confidence that the speaker selection & presentations will be objectively very good, but subjectively it's a different matter. Going through the proposals, I can imagine a "dream lineup" for me that would be totally uninteresting to many other people. Knowing the interests/background of attendees can help shape the program.
5. Demo track/poster track tickets. Pretty common in academic conferences - Not a confident speaker? Didn't get chosen? Is your material relatively stand-alone while you're not there? Willing to commit to some time to field questions about it? Get a (possibly discounted) ticket. Edit: possibly a great fit here, "TAs" for nodebot hacking.

[crtr0]

@notmatt Thanks that's what I was looking for. I was aware of @apike's list of what other events were doing, but I was more curious as to some concrete ideas for what CascadiaJS should be doing. Thanks very much!

[ghost]

I'm liking the idea of a combo of Mad Rush plus Lottery. I think it's the right combination of determination and luck.

I'm not a fan of the application process.The problem with an application process is twofold:

1: The selection of attendees could reinforce biases of the selection committee, even unintentionally, creating a lack of diversity in opinion and experience.

2: The application process is additional labor for the organizers. That either takes time from more important things or increases costs.

I'd also suggest a small set-aside for K-12 teachers and 501c3 non-profits like Code for America.

[dw2]

I like the idea of combining multiple methods together:

1. Early bird tickets - make it a lottery. If you win, hooray for saving a few hundred on tickets. If you lose, no big deal, as there are more chances for tickets ahead.

2. Scheduled batches - announce in advance the time at date that tickets will be going on sale, and make these times variable. If noon on Tuesday just wont work, that's ok, there is another batch in two weeks on a Friday, etc. This also makes planning easier on attendees who need approval from upper management in advance.

3. Reserved tickets - keep a portion of tickets set aside for both speakers, and as a discretionary reserve pool. You might run across someone who just "has to come", and the organizer could use their best judgement to make these tickets available for sale on a case-by-case basis. Otherwise, if these don't get used, just have a "last minute" batch to sell out the last of the stock.

@GregBulmash made two great points against an approval process. I'd like to add two more: 1) The difficulty in weighting and determining a potential attendee's excitement for the conference as well as their "street cred". and 2) The negative feelings that many would-be attendees would have from being rejected by the committee.

Great discussion so far. Also, @crabasa you've done a great job as an organizer. I got in on one of the last batches of tickets last year, and am glad I did. I know how stressful organizing a conference can be, so please accept this thank you and high five in advance.

[murvinlai]

I tried within 5 minutes at the early bird and right at 10:00am at the "Not quite early bird" days, and I couldn't get any ticket. I was quite disappointed. I don't know if I can get the ticket for next batch now.

If it is a big event, are 600 tickets for early & semi-early bird enough? Where is it going to host? Should it be in Vancouver Convention Center?