I’ve recently put some thought into how tickets are managed for
Minnebar and
Minnedemo. I’ve been very close to
these events for a long time, and on the board for several years now,
while our ticketing process has remained relatively unchanged over
that long time.
The Problems
Why spend cycles thinking about the ticketing issue? I think there are
some large problems with the way that tickets are managed.
Timeliness
The ticketing process encourages people to act immediately upon issuance
and get tickets since they will all be taken immediately. Often people
will take a ticket before they have even made sure they can attend. This
timeliness requirement unfairly penalizes people that are occupied and
not watching Twitter at that exact moment.
Understanding Churn
The current process reduces our ability to understand how many people
have stopped going to our events. We routinely see a high number of
“first-time” attendees, and along with that a high amount of churn. Is
the churn because people weren’t ready the second that tickets came out,
or because they don’t want to come?
Redemption
Many tickets go unused. We manage around this by predicting what
percentage of tickets will be unused, but it is more art than science.
The rushed nature of ticketing may decrease redemption rate as people
just jump on tickets as quick as possible.
Quantify Demand
All of these events “sell out” (to the extent a free event can sell out)
within minutes. This removes our ability to know the true demand for the
events. If 800 tickets are picked up within 15 minutes of release, how
many people wanted to go? 900? 3,000? That is anyone’s guess.
Limitations
When considering how to better manage tickets, it is important to
consider some limitations.
Identity Is Unknown
There are numerous ways to manage ticket redemption that would focus on
reputation. Mainly this is a means to solve redemption and put rules in
place such as not being allowed a ticket if you don’t use the last three
tickets.
Cannot Be Complex
We’ve entertained some options that avoid identity and reputation but
encourage careful assignment of tickets. The most commonly referenced
one is to use some form of a cash deposit. For example, you would pay
$5, $10 or $20 to get your ticket to the event and that money would
be refunded back to you when you show up at the event. This would
certainly reduce the number of unused tickets but at the cost of
managing a significant amount of complexity and risk.
Requirements
Before looking at the proposal let’s recap from above what our
requirements are:
- The solution must not require managing identity. You cannot know a
ticket holder from one event to another.
- The solution would ideally allow us to understand how many people
would like to come to an event.
- The solution would be able to manage preferential access for certain
groups.
- The solution would allow for improving the precision of “no shows”.
It is important to note that while “no shows” are something to
consider, I believe it is a lesser problem than understanding the true
demand for the events, churn and making the ticketing process more
equitable.
Proposal
Taking these objectives into place, I propose the following ticketing
process.
Phase 1: Get in line!
First thing, get in line. Rather than distributing tickets in groups we
propose everyone get in a line (or list) to attend. This line opens up
at a prescribed date and time, and anyone and everyone that wants to
come to the event would then get in line. The line will close at another
prescribed time, but as long as the line is open more and, more people
can get in line.
This line dampens the urgency to take immediate action. Please note
though that Phase 3 does reward people for getting in line early. It is
valuable to the organization to get a sense of demand quickly, so
getting people in line sooner is a good thing. However, no matter when
you get in line you may get a ticket.
Phase 2: Priority Access
Once the line is complete and closed we then deal with priority access.
There are some groups that get priority access to the events and are
guaranteed a ticket:
- Volunteers: Anyone volunteering at the event.
- Community Supporters: Individuals (not companies) in the
community that have donated $100 in the last 12 months.
- Sponsors: Packages get different numbers of tickets. If you are
in line under one of those slots you get a ticket.
- Students: We look to encourage high school students to attend
our events, and wish to give them priority access.
Once these special groups have priority access, we’ve used some of the
tickets and now have a smaller line.
Phase 3: Ticketing Groups
The remaining line is now divided into groups. This is where some
further math and heuristics could be applied to better manage tickets
over time. For now, just assume that all these numbers are potential
variables and you could optimize the solution over time.
Take the remainder of the line, let’s say there are 1,000 people still
left, and for easy math put them in 5 groups of 200 people.
Each one of these groups is then ticketed. Remember that these groups
are in the order they got in line, so the early people are in the front
and the last person to get in line is at the end in position 1,000.
We also at this point know how many tickets were assigned to priority
access groups and can make a determination about their “no show” rate.
Likely those groups have a very high redemption rate, so assume most are
used. The remaining tickets will be assigned to the five groups of 200
people in a decreasing percentage. All of the people in group 1 get
tickets. 70% of the people in group 2 get tickets. 40% of the people in
group 3 get tickets and so on until we get to something like 10% of
people in group 5 getting tickets.
Over time these groups may be able to provide more accurate data on
redemption, and we may know that Group 1 people redeem tickets at a
higher rate than Group 5 so we can assign more tickets in Group 5
knowing more will go unused.
Additionally, it is important that any tickets that get returned before
the event, when people realize they cannot come and notify us of that,
should be returned to the group it was assigned to. So, if a person in
Group 3 returns there ticket, it should go to another random person in
Group 3. This could be managed by creating different classes or types of
tickets for each group. If there are no people left in that group
without a ticket, like Group 1, then flow the returned ticket to the
next group in line.
Summary
This may seem complex at first, but I think the implementation could be
reasonable. Eventbrite could still be used to get in line but getting a
“Line Ticket”. The list of registered email addresses could then be
manipulated with a small program to do the rest.
I do think that this would meet our goals of:
- Understanding true demand for the events.
- Allowing special access to some groups.
- More fairly assigning tickets to the remaining people.
- Minimizing the urgency around ticketing.
- Optimize “no show” management over time.
Very importantly this would also help us understand people that stop
coming to events. If you get in line we know you want to come, but if
you don’t get in line many times in a row you are deciding our events
aren’t worthwhile and we’d like to know that.
If you have suggestions or comments contact me. If you would like to
help solution this that would be great too!
I’ve been using 
