This is part 2 in a series of 3 posts on ALGO 2009, largely repeating the organiser’s report from the business meeting.

Boring statistics

For some conference attendees, this has already become a drinking game: the part where the organiser entertains with useless statistics.

I have probably the best number ever: the total height of the ALGO attendees is 283m, discounting those that didn’t give us their measurements at registration. (We needed these to plan ALGO 2009’s social event, a short bicycle trip to Copenhagen city hall.)

Number of participants

More interestingly, in particular to later organisers of ALGO: the budget. When planning an event like this, the big open question is how many people will actually show up. Previous ALGO organisers were very helpful in giving us their numbers, but it’s hard to use them for prediction. Previous ALGOs hosted ESA, WAOA, and ATMOS, but in 2006 and 2008, ALGO also hosted WABI, the Workshop on Algorithms in Biology, which is a large conference. We did attract IWPEC to ALGO 2009 (two dozen talks), but it is unclear how many extra attendees this produces – ESA itself had several tracks of IWPECish presentations, so there is a lot of speaker overlap. In the end, we were able to match the Karlsruhe numbers, even without WABI, and without the large number of local attendees Karlsruhe presumably had.

Registrants by type

The number of attendees doesn’t give full information for the budget, because students pay far less. ALGO budgets over the last four years operate with the number 57% for the number of expected regular participants. At ALGO 2009, we had some 75 students, so the number of regular participants is in fact slighly above 60%. We tried to make student participation as attractive as possible, but it seems to be difficult to attend ALGO just for the event, without a paper. (Or is Copenhagen sufficiently attractive for the advisor, not the student, to attend?)

Length of stay

Another constant throughout the ALGO budgets from previous years is the assumption that an attendee stays for 3 days. (You need this to calculate the number of lunches.) We asked attendees at on-line registration to tell us their expected date of arrival and departure, and asked them to confirm these dates when they collected their name tags. If these numbers are to be trusted, more than half of the participants actually stayed for the full week; in particular, they gluttonously devoured all the food they actually paid for. This would be a huge success from the perspective of us wanting to organise a large, monolithic 5-day event. Of course, since we have to order all these lunches, there is very little wiggle room left in the budget.

Budget

Finally, the budget. For paying 180 participants, we budgeted with something like 500,000 Danish kroner, or roughly 65,000 euros. Most of the expenses are food. This includes lunches, coffee, breakfast, and some cakes during the day. After the business meeting on Monday, we provided snacks and drinks at ITU. The social event was a bicycle trip through Copenhagen, the main expense is renting some 200 bicycles. ITU provided the rooms for free, and most of the staff. (We had to find some help with registration.)

The only expense that could have been reduced is the invited speakers. First, ALGO had a lot. 7 invited speakers for an event with less than 200 registrants cost a lot of money per ticket. In even years, ESA “pays” for 2 invited speakers on the first three days, WABI gets the third. We decided to have 3 speakers on the first three days anyway, and 2 speakers per day for Tuesday and Friday, which hopefully transformed these days into very attractive experiences and made many people stay beyond ESA. Of course, all of this is expensive.

Should I do this again, I would probably be less generous. A possible model would be to let invited speakers pay for their own travel. (This is something that the speaker can easily find funding for anyway.) I understand that other conferences use that model, or at least put a cap on the travel reimbursement.

This is part 1 in a series of 3 posts on ALGO 2009, largely repeating the organiser’s report from the business meeting.

ALGO week

Here’s what the ALGO week looked like, from 7 September to 11 September 2009:

Without WABI, the ESA conference has the first three days to itself – an alternative would have been to give three full days to IWPEC and let it overlap, at least partially. Given the IWPEC submissions that might have been a better idea.

PC meetings

More critical is the situation before the conference proper, when all the various PCs have to coördinate their activities. It has been decided that ESA waits for ICALP’s decisions, so with Springer’s deadline for conference production, the available time slots are pretty tight. For the organiser this means that the deadline for early registration has to be pretty late, which makes it hard to produce a realistic final budget. Also, we were unable to hold hotel rooms longer than the first week of July, which was before WAOA’s notification. This turned out to be no problem, hotel rooms were still plentiful.

Still, all the PCs had to work under very tight conditions. Moving the ICALP PC meeting to the left a bit, and possibly moving the ALGO conference date farther into mid- or late September would remedy this.

Registration

One of our early decisions was to view ALGO as a monolithic, five-day event, rather than a smörgåsbord of workshops and events for which to register (and get billed) separately. This is a hard decision that has many consequences for organisation, financing, dealing with PCs, etc.

1. Student registration covers social events. The social events are an important part of the conference, in particular for students. Therefore, I’m never really happy with conferences where student registration does not cover the joint dinner or the social events. Many institutions make it much harder to get reimbursement for a dinner ticket than an exorbitant conference fee. Thus, we slightly under-priced participation for students, and slightly over-priced participation for regular attendees.
2. No separate workshop registration. ALGO 2009 participants paid the same, whether they stayed for five days including all social events, or whether they just popped by for their own talk and left the same day. This made it attractive to stay for more days, which is what we want, but probably also prevented some people from attending.

In the end, I think we made the right decision. We had lots of attendees, and many stayed for many days. However, the ATMOS workshop is probably hurt by this: it’s a 1-day event, even trying to attract relevant industry, and the full ALGO participation fee is just too high for that. In hindsight, I believe a separate “1-day, no fun”-registration, for exactly half the price of the full ALGO registration, is the proper way to go.

Venue

We decided early on to have ALGO at the university, during the semester. I like this a lot better than using a dedicated conference venue or a hotel; the atmosphere is nicer, and we’re sending important signals in both directions. The possible downside is that ALGO 2009 was crowded (which I like), and that we had to make some compromises with respect to lecture rooms that would otherwise not have been an issue.

It’s also dirt cheap.