SIG Governing Board Minutes: April 29, 2022
Cherri M. Pancake
*List is not all-inclusive
SIG Overhead (Pancake)
Palsberg provided an outline of the agenda and ask to begin with the SIG Overhead (OH) agenda item. He introduced Cherri Pancake for presentation and discussion. Per Palsberg, the big picture is that one way or the other, we are headed toward a vote of the SIGs on the overhead formula.
Pancake began her presentation by providing an outline of why the SIGs were meeting. They are doing well financially but the OH calculation method isn’t covering SIG related costs. That means that other ACM groups have had to subsidize the SIG operations. The ACM President charged the SGB to update the calculation for FY’23. Pancake referenced the 2 back-up documents that were sent and indicated that this was the final step in a multi-year effort to improve the processes so that finances are clearer for volunteers. The TF was charged with reviewing how SIG OH is handled and to make recommendations for adjusting the process. The timeline began in January/February with recommendations being presented to the SGB for questions and discussion today. The expectation is that the SGB will vote on a new method in early May and in early June, the ACM EC will report it to Council as part of the final FY’23 budget discussion.
Pancake thanked the members of the TF who she indicated did an incredible amount of work. It’s been hundreds and hundreds of hours over the last 4 months. In every discussion everybody has been trying to consider the impact on all SIGs. SIGS and conferences account for 1/3 of ACM’s total revenue and expenses. She provided expenses from FY’21 of what she said was a remarkedly consistent picture. SIGS are 1/3 of revenue and exp but that’s just part of the picture because SIGs have reserves. At the end of FY’21 SIGs controlled 70% of the operating reserves. 2/3 of SIGs have been in a trend of accumulating wealth, not on purpose, fund balances end up growing during lean years and some SIG funds continue to grow at an incredible rate. Pancake did one-on-one with every SIG Chair while gathering information for the TF. All leaders indicated they were transitioning back to in-person events with a virtual component. Their biggest challenge was recruiting members. Especially recruiting people to help as volunteers. Many were interested becoming more inclusive. Most said they were worried about how to spend funds for community rather than earn. SIGS are on the road to recovery. The SIGS are doing well but its not the full picture. SIGs and conferences spend about $13M in direct costs – expenditures with receipts but there are also indirect operational costs that ACM incurs on behalf of SIGS – staff and financial services, membership services, IT infrastructure, financial and legal indemnification, SGB participation, ACM wide boards and councils, elections, ethics, plagiarism, harassment, sanctions, etc. See the back-up document for more information.
Operational costs add up ad a detailed analysis was conducted last fall. FY’21 related costs were $4.091M. To cover those costs, ACM charges OH to SIGs and conferences. The existing method hasn’t been updated since 2001 and doesn’t generate enough dollars to cover the SIG related costs. It fell short 5 times in the last 10 years and the worst case was FY’21 with a $2.050M shortfall. There is something called and OH reserve fund in years where there is a shortfall, that is tapped to cover the expenses. It had $400K but is depleted. The ACM Council asked EC and the SGB to address this for FY’23 budget.
First, the TF wanted to know why it was happening. They looked at SIG related costs which have gone up less than inflation. As an annual percentage it’s 2.3% over 20- year period. They determined it wasn’t inflation. What has changed is the pattern of SIG spending. Turns out that 10% of the SIGS are now accounting for half the overall expenses of all the SIGS and those SIGS pay OH at a low effective rate due to original method set up. 20% have what you would call minimal expenses and they don’t pay much in OH other than 10K minimum which is actually covered by the minimum form the DL. Each year ACM is recovering less OH in dollars even while spending is increasing. Clearly a problem with method calculation. Is it a blip or trend? Because of the factors discussed we are ended up with an overhead shortfall. It’s not getting better on its own so needs to be fixed. The TF got everyone to same level of understanding, brainstormed, looked at 15 methods including tweaking and a blank slate. They had 4 rounds of discussion and did 2 types of analysis. They did a retrospective analysis looked at what would happen going forward using each method including a year or two of recovery and secured input from Chairs on the outlook for each SIG.
The TF is now recommending how the SGB can solve the OH issue. Their written recommendation will be sent following the meeting.
- Clarify the function of the OH reserve fund
- Improve the transparency of the OH process
- Update the minimum OH fee
- Update the OH calculation method to solve the problem
Leaders were told they’ll decide on the recommendations after the call. The first recommendation was updating the minimum fee to $25K. The minimum fee is a call for a minimum level of SIG activity. Setting it too low provides incentive for an inactive. Pancake then discussed choosing the new calculation method and indicated consensus was to tweak the formula. One method keeps the previous method but applies a 1.6 multiplier to rate scale. The other method changes the increment to $250K instead of $125K so multiplier is 1.26. A calculator will be issued following the meeting and will allow you to play with these methods. Before that, think of the concept of fairness. Conceptually simple, each SIG should pay a fair share but in practice it’s impossible to be equally fair to all. SIGs are just too heterogeneous. The TF also tried to look at correlations but there aren’t any. No way to identify a typical SIG and no way to say what happens to a typical SIG under these 2 methods. Each SIG is unique. Grouped SIGS into cohorts on the 5- year history. Impact varies under 2 methods. Original method favored SIGs with high activity. Also significantly favored low levels of activity. The first method is even-handed in terms of the increase that each cohort sees in their OH bill. The 2nd method shifts burden significantly and favors lower-level activity SIGs. See slides for details.
It's important that you consider not only the impact on your SIG but the message being sent to SIGs with different impacts.
Questions/Comments Moderated by Palsberg
What does other “ACM’s groups” mean?
All the Boards and Councils, ACM wide (pubs, education, DEI, etc.).
Please provide input on slide 13 historical data.
Last time an in-depth study was conducted was 2001 so we’re trying to document this being done on a regular basis. Time consuming to do a detailed analysis of costs so it is not practical every year. The TF recommends every 5 years.
What were shortfalls in other 4 years?
Aggregated info – see slide 25.
Things could change after conferences go back to in person – comments?
Doing forward projection 22 and 23 are recovery but by 24 the TF is hoping to be back on track. There will be a more frequent review in the future.
What if you pay more than you need in reserve?
The money is distributed back out to SIG s in proportion to what was paid.
How much impact do virtual conference have on trend?
VCs seem to have made more than projected – We were reasonably conservative. Look carefully at recommendations which stipulate what if SIGs pay too much.
Can someone list the 10 SIGs that count for half expenses and the overlap with representation on the TF?
COMM, DA, IR, PLAN, CHI, GRAPH, HPC, SOFT. CHI, GRAPH, HPC, PLAN, and KDD had reps as well as CSE and MOBILE.
Was there a deficit before pandemic?
Shortfall occurred 5x over past 10 years.
Was there minimum DL in 200?
No didn’t start until about 2009 or 2010.
Should we not index minimum cost by inflation instead of fixed?
Interesting point talked about other ways of indexing but said for planning purposes you need to know what it is going to cost. Saying wait until end of ear might not be a good idea. We’re hoping that by guaranteeing this is reviewed regularly if inflation goes wild it may take 3 or 4 years before the minimum fee is adjusted but it will be.
Which years does this less than ½ apply to?
SIG OH covers less than ½. In FY’22 it was mentioned specifically. The first in 21 the SIG OH was less than ½ amount needed. In 22 it will still be less than ½ of the total. Pancake showed the calculator attendees will receive following the meeting. Looking at old method there was a percentage applied to each $125K steps starting at 16%. The 2 methods looked at the target to determine the multiplier and steps.
Is ACM as a whole losing money? I would argue SIGs contribute more beyond OH.
There are many volunteer led programs in ACM. When you talk about ACM, ACM is us. The SIGs are one branch of the tree. Please review the slides to put this in perspective.
SIGGRAPH has 3,000+ members are these the only options available?
Looked at 15 methods these are the ones the TF agreed were the 2 best ways to do it after running 15 methods and a huge number of simulations.
Palsberg asked other members of the TF to comment.
TF member Lawrence indicated that moving toward an even-handed method was less equitable because it perpetuates status quo where richest SIGS pay most. The SIGs with largest expenses pay less in percentage but pay more in final amounts and those SIGs with higher expenses are being taxed on their conference related contracts for outside services.
Andrew Kun thanked Pancake for leading the TF as it was a lot of work. He thanked his colleagues sitting on the TF for their support and all others on TF. Echoed what was said about costs. ACM costs are real and we benefit from what is provided. If there are costs, they need to be paid. Not surprising that costs are rising. Look at 1 vs 2 and we should care about fairness. TF was assembled not to look at fairness of status quo but to look at shortfall and how to we change that. Its reasonable to say a SIG with large financial resources should pay. But each SIG is made up of individuals and OH is charged to conferences so fees are paid by those people in attendance. Individuals will pay a larger fee for conferences and are directly affected by our decisions. Since the task was to address the shortfall. CHI will pay a larger portion of pie but if spread equally explanation is easier to pass onto constituency.
John West indicated that Andrew gave a fantastic explanation of the right way to look at this. Not much to add. Reminder about the method chosen being fair in passing charges on to serve the community as a whole.
Pancake indicated that the SGB should not make the mistake of thinking cohorts reflect fund balance. There is no correlation there. If arranged in terms of fund balances those cohorts would have been very different there are SIGS in cohorts 2 and 3 with large fund balances available. That’s not a criticism but easy to assume these aren’t as rich and isn’t true.
If I pay $100 for membership does it go to SIG or only sig membership?
Only SIG membership payment.
Is there an analysis of proportion of SIG OH due to each cohort?
There is no way to do that. You’d have to ask every person to track every transaction and assign to SIG to get a handle and don’t know how you’d partition IT and other costs.
Is there a way to proportion a SIG costs?
Impossible because each SIG is so different and things change from one month to the next: new conference, activity, conference goes away, etc.
Jortner asked if the size and activity directly correlate to increased cost? SC or SIGGRAPH spend contracts for services that ACM isn’t in a position to support. The other piece to keep in mind is this is a tax on expenses. SIGGRAPH pays a lot more because I have a large conference. Doesn’t mean I have more, I just have a large conference. Something to look at. Even if we implement this, we should do this short term and revisit every year while this is happening. We may want to readjust. Pancake explained that the TF We indicated at least every 5 years and there’s an annual adjustment because of the way the OH fund is specified. Too much work for anyone to do this every year.
Multiple people mentioned $400K in OH reserve?
To make fair choice we should look at SGB expense and SIG expenses over last 5-10 year can you provide this information?
These are the financial details here based on averages over the 5 years. You have the details for your SIG in budget planning spreadsheets. Not for each but overall development. Please review the slides following the meeting to see all the details that were gathered prior to making recommendation.
Multiple questions on which other parts of ACM are revenue generating
Ryan indicated that membership, DL, publications, and advertising are the primary sources.
For fair distribution, can we look at SIG related costs at ACM to reduce them to a lower deficit?
Not task at hand. Agree that it may be possible to agree to stop doing something but that is not in scope of this work or discussion. Decision of new method cannot be deferred because this problem needs to be addressed and perhaps that’s another issue for another TF. Your SIG may not use all but somebody is so it is not a just a question of my SIG but all.
Concern proposal is based on FY’21. Need better understanding of last 5 years to make informed decision.
It is based on average of FY 17 to 21. Nothing is based on 21 alone except from what was reported from cost analysis.
Palsberg called for other questions.
An SGB members asked if they could delay decision until next year.
It was suggested that they vote just for next fiscal year now. Pancake indicated the problem of solving this was discussed at Council and EC and they gave a timeline of the current year to come up with a method. Vote needs to be done on a method now because of the seriousness of this. It’s a problem that has to be solved and clear that SIGs have money to solve it. The SGB needs to make it work at least for FY’23 budget to be approved. Whether you could revisit it later is an SGB decision. You need to tell council we’re solving in this way.
Which behaviors of SIGs being incentivized by the new method?
That is an individual decision not only of each SIG but for each SIG to make about each conference. You’re the ones with the money in the bank. Decide to spend reserve so not to increase rates – its your decision.
An SGB Council Rep indicated 5 out of 10 years the SIGs underpaid.
Reasonably the remaining 5 years SIGs overpaid so in past ten years there’s a balance of $400 K reserve so immediate problem is for FY22. If that’s the case then we should only try to solve FY 22 problem because there isn’t a bigger problem. Extra amount goes to OH reserve and used in a year when OH doesn’t meet cost.
Is action needed for a blip?
In 5 out of 10 years it went up a little but started with money in there and only reason there was anything was because FY 20 was a good year. This is a trend not a blip. First 9 or 10 years worked ok and it’s not working anymore. FY’22 is not an issue, SIGs received another grace period for that. It’s FY’23 and beyond that needs to be addressed.
West indicated we have new audit and we know we’re guaranteed to fall short every FY going forward if we don’t make a change.
Palsberg indicated there were two discussions going on in chat - make shorter term and revisit next year or go with a decision. We’re not planning to revisit anytime soon and the other thing is whether ACM should have another task force on cost that ACM has to service sigs. That’s more like something for the future. For the present we can do something short term or longer term. There is no time for SGB EC to discuss at this moment so propose that the SGB EC have a chance to discuss and we may have another meeting. The SGB EC will collect its thoughts and propose which way to go.
Pancake indicated it is important to show the SGB is capable of making decisions themselves. Maybe vote on a method for FY23 and perhaps revisit next year.
Jortner asked what happens for conferences with approved budget. Could impact SIGGRAPH if applied immediately. Pancake indicated that each SGB leader needed to look at their OH agreements with each conference.
Palsberg indicated that the SGB would hear from the SGB EC soon about this. They’ll look at it and plot a good way forward with a vote in the future.
Hybrid Conferences (Rajan, Craig, Ferdman, Lunt)
SPLASH 2021 (Hridesh Rajan)
- SIGPLAN track at the conference where invited authors from prior virtual conferences could come and present their papers in person
- Hybrid coffee table – attendees from in person and virtual conference could meet (future suggestion: have more than one hybrid coffee break rooms)
- Use of Discord to bridge in-person and virtual participants
SIGCSE 2021 (Michelle Craig)
- Used Pathable but likely won’t use them again in the future (not to do)
- Had professional production support on site to help manage each of the streaming rooms
- All papers and posters submitted a video ahead of time which was available to all attendees
- Workshops and demos only in person (not accessible to remote attendees)
- “Coffee break” was highly successful and was led by a charismatic person managing a fun presentation in the exhibit hall during scheduled coffee breaks and simultaneously interacting with in person and remote participants
- All in-person attendees had full access to the virtual system
- Undercharged for virtual attendees (not to do)
- *Plan to make the conference fees the same for in person and for virtual for 2023*
- Committed to hybrid for 2023, 2024, and probably long term
ASPLOS 2022 (Mike Ferdman)
- It is possible to run a well-attended international conference now
- Avoid “hybrid” if you can
- Held off announcing virtual (Zoom) option as late as possible to encourage people to attend in person
- MASSIVE on-site staff to manage A/V
POPL ‘22 (Barry Lunt)
- Airmeet platform (for those who registered) 24 in-person talks + 47 remote talks (none pre-recorded, all live)
- Real-time captioning (hired professional help)
- Student volunteers were great, but were asked to do too much, not sustainable (Perhaps ACM can help with this)
- Virtual option is here to stay
- ACM should set up clear guidelines for registration fees
- ACM should set up professional service to manage virtual component
- Remote attendees must accommodate time zone requirements when the presentation is synchronous
- Strongly recommend visit hosting facility, put hosting facility contact on speed dial
- Have a dedicated technology person in each room
- Arrange equal registration fees to encourage in-person attendance
- Always allow some degree of hybrid participation
EIG on Reproductibility Towards SIGREPRO (Bonnet)
- 455 members
- Mission is to foster an inclusive and broad community around the issue of reproducibility
- SIG should be a forum for reproducibility practitioners so there can be an exchange across communities
- Conference working group + practice working group
- Newsletter’s first issue in the summer
- Discussion around contributing to ACM curriculum if it is relevant
- Ultimately focused on working towards SIG status
- Action item directed towards SGB: contacting Philippe with your champions (reproducibility award winners) and artifact evaluation
Social Responsibility & the SIGs (Schuler)
- SIGCAS is interested in exploring ideas related to collaborative projects that they could do with the other SIGS.
- Upside and downside to computerization of the world
- Maybe we need a Cross SIG event on social responsibility
- Issue-based working groups on, e.g., Climate Change, Fair Algorithms, or Civic Technology
- There has been no ACM level Elections Committee as required by the bylaws for some time
- SGB nominating process has not been following the bylaws for some time
Jeanna Matthews: I really hope that we can meet a little more often. We can really focus on continuity of the issues. What services? (As it relates to ACM) ACM should be doing more virtual services.