SIAM's Ongoing Search for a Perfect (Low-cost) Meeting SIteOctober 7, 2002
James M. Crowley
By most accounts, our 50th Anniversary Meeting in Philadelphia this summer was one of our best. Great people, great program, great spirit. . . . In the interest of future successful meetings, I'd like to use the space SIAM president Tom Manteuffel and I usually share to consider some of the practical issues involved in running a SIAM meeting.
First, a few general observations. In a recent survey of SIAM members (in which more than 20% of the membership participated), respondents rated SIAM meetings highly. On questions of technical quality, respondents indicated that they were very pleased with SIAM conferences. The quality of plenary speakers and minisymposia was rated as above average or excellent by more than 70% of respondents, and as below average or poor by less than 2% (with the remainder judging them average or not responding).
On questions of cost and location, the response was generally favorable as well. Fifty percent felt that the meeting locations were above average or excellent; 5% responded that they were below average or poor (another 30% found them average, and the remainder did not respond). When asked to rate conference fees, 40% indicated that they found them low or fair and reasonable; 34% considered them high but not unreasonable; and 13.4% found them too high.
These generally positive responses notwithstanding, the SIAM leadership gets a lot of questions about conferences, and at its meeting this summer the SIAM Board of Trustees asked me to report to the membership on conference fees. I decided to use a question-and-answer format, choosing some of the questions we hear most often. The questions usually refer to annual meetings, but in many cases apply to other SIAM conferences as well. Readers are also referred to an earlier article (November 2000) on the subject.
Why are the registration fees so high?
They're not. Compared with many engineering and computer science organizations, we tend to have rather modest registration fees; we also provide a lot of services for those fees. Two years ago, for example, the SIAM Board of Trustees decided that we should provide a data projector (with no additional fee) for each speaker, with the goal of promoting presentations of the highest possible quality. We also provide a great deal of food and beverages, in the form of morning and afternoon coffee and various receptions.
SIAM loses money on conferences. Or to put it another way, we subsidize conferences to some extent. We would prefer to break even.
What is it that costs so much?
Most of the costs are external ones. SIAM staff charges have been falling as a percentage of meeting costs. On average, SIAM staff labor costs about $43 per attendee. At the 2001 Annual Meeting (in San Diego), staff charges accounted for about 17% of total direct costs.
The bulk of the expenses, and especially the large increases in costs that we have experienced in recent years, come from a small number of items. The following numbers are based on the 2001 Annual Meeting:
- A/V. The cost of audiovisual equipment was roughly $75 per paid attendee. During the 1980s, by comparison, the cost of this item tended to be closer to $15 per person. We use much more equipment now, especially data projectors, which are expensive to rent, and we do not pass this cost along to the speakers. We have looked into ways to reduce this expense, including purchasing and servicing our own equipment, but so far have not found any cost-effective solution. This item represents about 28% of the direct cost of running our conferences.
- Special functions (food and beverages at coffee breaks; receptions). This item represented about 17% of direct costs, or $47 per attendee. This item increased by 50% from 1997 to 2001, in part because the number of receptions increased during that time.
- Printing and composition (program book, signs, etc.). This item accounted for 11.7% of direct costs, or $32 per attendee. (Expenses for this item were unusually high in 2001.)
- Advertising and promotion. This important item represented about 11% of direct costs, or $29 per attendee. (We have never tried an unadvertised meeting.)
- Invited speaker expenses. For an annual meeting, this represents about 8.5% of direct costs, or $23 per attendee.
- Part-time local help (at the registration desk). This item was about 3% of direct costs, or $9 per attendee.
- Various other miscellaneous costs. The costs incurred in this category, along with SIAM staff costs, represented 24.5% of the total.
Why not hold meetings on university campuses?
For smaller meetings, this may make good sense, depending on the style of the meeting and the wishes of the organizers. SIAM is certainly willing to work with organizers who wish to run a meeting on a campus. The next conference of the SIAM Activity Group on Linear Algebra, for example, will be held at the College of William & Mary (in July 2003). We have also run annual meetings at universities. During the last five years, we met at Stanford (1997) and the University of Toronto (1998).
But holding meetings at universities isn't a panacea, especially in the case of large meetings. The selection of dates is usually limited (as it was with William & Mary), and conferences are seldom a high priority for a university. This means that room assignments are often done rather late in the planning cycle, even after the program has been printed. Audiovisual support and other services are often not as strong as those we expect from a hotel or conference center.
When we did run annual meetings on campuses, we received complaints-about the difficulty of getting to the meeting from area hotels, and the difficulty in getting from one talk to another.
The direct costs mentioned earlier do not vanish when meetings are held at universities. For large conferences, many universities charge fees for use of the space (and some even have rental fees for chairs). This can have an impact on the registration fees.
Don't hotels charge for the space?
No, in exchange for our promise to fill a certain number of sleeping rooms, the hotel provides meeting room space to SIAM for free. But in signing such a contract, SIAM takes on substantial risk. If the sleeping rooms aren't filled, we are required to pay substantial penalties.
Don't you avoid this problem by going to a university or a convention center?
Not really. Because many attendees don't wish to stay in dormitories, we would still want to reserve sleeping rooms at a hotel. And while we can negotiate good (although higher than dorm) rates for this block of rooms, we would still need to make guarantees.
Imagine what would happen if we didn't reserve any hotel rooms and our meeting coincided with a major event that caused all the rooms in the city to be taken. This would have a major impact on attendance at our meeting---even if we were holding the meeting at a campus or convention center.
Why doesn't SIAM provide information about alternative (cheaper) hotels instead of just the flagship hotel?
This is really a tough question. We respect the right of members and attendees to choose the kind of hotel they want. But if too few people stay at the conference hotel, we don't meet the room block we guaranteed and stand to pay penalties for unused rooms.
For events the size of our annual meetings, the number of hotels with enough space---15 meeting rooms for parallel sessions---is very limited. We can't hold a meeting at the local Motel 6.
The senior elected leadership, moreover, has pushed for nicer venues. This of course usually translates to higher prices.
Why not try smaller cities?
Another conundrum. We tried that strategy---Charlotte in '95 and Kansas City in '96---and fewer people came to the meetings. Attendance tends to be larger in places like San Diego, Stanford, Boston, and Washington. If we look at the data for the SIAM annual meetings going back to 1983, we find a natural clustering in the attendance data. One cluster, containing venues in Boston, San Diego, Chicago, Stanford, Los Angeles, Philadelphia, and Toronto, averaged 855 paid attendees. The second cluster, with meetings in Denver, Seattle, Pittsburgh, Minneapolis, Charlotte, Kansas City, and Atlanta, averaged 549 paid attendees. A sharp contrast!
Realizing that no one venue will please everyone, we aim for variety in our choice of meeting sites. According, the '03 annual meeting will be in Montreal (in conjunction with CAIMS), and in '04 we'll be meeting at the Convention Center in Portland, Oregon.