My hybrid conference wishlist
My first conference was the IEEE Symposium on Visual Languages in September, year 2000. It was a first in many ways: it was my first academic conference, the first time I’d given a conference presentation, and the first time I’d been to Seattle for professional reasons rather than play. I was just a sophomore in college and I was simply terrified. I had to buy a new outfit to look “professional” (but I don’t know what that meant), I had to prepare a presentation on print transparencies for an overhead projector (because digital projectors didn’t work reliably back then, nor did laptops), and I had to somehow pack for a week away from home, traveling alone (which I had never done). I was nervous, but also buzzing with excited anticipation.
It was inspiring in so many ways. There were more than a hundred people that shared my interests, all smart, curious, and welcoming, and there to connect. I met other students (mostly graduate students), who helped inspire me to pursue a Ph.D. I fell in love with the University of Washington campus (where the conference was held). I fell in love with research travel, learning about new places, new people, and new ideas all at once, in an intense week of social discovery. I even met the professor who would eventually become my Ph.D. advisor, after asking him to lunch after attending his tutorial. Those three days of conferencing, plus the two days of travel, transformed my professional and personal life, all for about $1,000 of hotel, food, gas, and registration.
Of course, like any academic conference, it wasn’t perfect. Many of the talks were long and boring. I didn’t really know anyone, and there weren’t any structured activities to welcome me into the community. Despite only having more than 100 attendees, I felt like I left only having met a few people. I mostly spent three 8 hour days sitting in uncomfortable conference chairs, wishing I could talk to people, see the city, learn about their paths into academia, and collaborate with them. Instead, I felt like I mostly sat through one long three day lecture.
I’ve had the same experience over and over for the past 20 years. Conferences across CS are largely long days of presentations, sometimes single track, sometimes parallel track. Structured networking comes in the form of formal receptions. Some conferences creatively integrate networking into talks (e.g., the ACM ICER conference includes roundtable discussions after talks before Q&A, where attendees meet each other, share expertise, and develop good questions). But attendees slowly learn over repeat attendance that the best way to attend a conference is avoid the talks altogether: stay in the hallways, having serendipitous conversations; go on informal outings with attendees to explore the region where the conference is set; find a space to collaborate on that next grant or paper. We all come to learn that the unplanned, interactive networking is the most valuable part of the experience.
This past year of virtual conferences have mostly ignored this hard won insight, partly because of the difficulty of recreating meaningful networking online. Most of the conferences I’ve attended have been long series of pre-recorded videos with short Q&A slots. There’s occasionally structured networking, such as video chat roulette features, or happy hours with video chat social gatherings. At ACM ICER 2020, we tried to build in some serendipity through presentation “viewing parties”, where a small group of attendees watches a pre-recorded presentation then discusses it and other topics for 30 minutes. We also had scheduled breaks in Discord, where people could notice a friend or colleague, and say hi. All of these designs preserve the essential but least impactful part of conferences—quickly learning the latest discoveries through presentations instead of reading the proceedings—but mostly abandon the most impactful parts—building strong professional relationships through intense unplanned activity in exciting and unfamiliar parts of the world.
Despite their flaws, virtual conferences have offered several thing that in-person conferences haven’t:
- Attendees with disabilities who cannot easily travel, parents with young children but without childcare, academic leaders with hard scheduling conflicts, researchers without substantial travel funding, and even people who are just burned out on travel, but still want to stay connect—all of these groups have traditionally been excluded from participation, aside from following a hashtag on Twitter. Virtual conferences have opened these doors, allowing all of these groups to participate much more deeply than a tweet affords, even if not as deeply as being there in person. As much as they have taken away a critical aspect of academic community building, they have also broadened participation by removing structure barriers created by hard to correct inequities in child care, funding, and the built environment.
- Virtual conferences have led to valuable, reusable pre-recorded content that can be shared and viewed for years, rather than just experienced once. This has also enabled less practiced and experienced speakers prepare more polished presentations; videos are also captioned, improving accessibility of presentation content.
- Virtual conference platforms have made it far easier to learn people’s names, pronouns, and affiliations.
- The elimination of travel has dramatically reduced the academic carbon footprint by greatly reducing air travel.
As we slowly step back into our post-pandemic social worlds, how do we keep this accessibility and equity, mitigate our climate impact, while restoring—and perhaps even amplifying—the serendipitous in-person community building? Is such a conference design possible? Is there a way to do it without further burdening the unpaid volunteer communities that run these events?
I think so.
Here’s what not to do. We should not try to design hybrid conferences that synchronously weave together remote and collocated attendees. Anyone who has tried quickly realizes just how infeasible it is to seamlessly synchronous audio and video in physical and virtual space. Even the best live television production teams—take Saturday Night Live, for example—don’t try to offer the same experience to its live studio audience and its television viewers. And they still rely on highly experienced video production teams, long established sets, and an immense investment in rehearsal. I can’t see any of this feasibly happening in volunteer-run, low-budget academic conferences. The result would just be poorly produced experiences that are subpar for remote attendees and highly disruptive for collocated attendees.
While I think “synchronous hybrid” designs like this are doomed to fail, I actually think that other hybrid designs that alternate between remote and in-person could be quite successful, and even superior to the pre-pandemic conference designs of the past. Here’s a sketch of this idea:
Daily short virtual conference sessions. Begin each day with a short, synchronous, time-zone inclusive, late morning, 2–3 hour virtual event. Include ample breaks to avoid video conferencing fatigue. The point of this short virtual conference is to include everyone in a quickly learning the field’s latest discoveries. Talks would be pre-recorded with live Q&A and optionally other forms of networking like the roundtable discussions I described above. Remote attendees would attend like they do any other virtual conference, while on-site attendees might attend from their hotel room, or perhaps even in a shared space in which attendees can back channel by chatting in-person rather than in the chat platform. This portion of each day would feel like connecting with content and the broader community.
Lunch time. After the virtual conference period of the day would be a networking meal. These might be catered or they might be out in the host city. On-site attendees would get the usual benefits of meal outings, networking, exploring, and learning, but without the usual pressure to rush back to sessions. In fact, some might even just keep networking for hours and find other impromptu activities to do. On-site attendees would be encouraged to post on social media so that remote attendees can get a sense of the activity; some might even be encouraged to bring remote attendees along for the meal on a tablet, passing them around the table to participate in conversation. (I did this once at a doctoral consortium I ran; it works fine).
Afternoon events. After a generous lunch break would be a diversity of afternoon social events. The possibilities for how to structure networking and community building are endless:
- Newcomer rooms. In these rooms, newcomers to the community can meet each other as well as regular attendees, building those much needed strong ties. There would be no agenda, other than chatting, snacking, discussing the content from the morning. Attendees might even bring remote buddies on their smartphones, helping them connect as well.
- Workshop, doctoral consortiums, and birds of a feather rooms. Attendees could organize pre-planned or impromptu workshops on topics of interest, gathering attendees around research areas or ideas that emerged in the morning virtual conference session. Student volunteers could stream discussions and activities for remote attendees who want to participate.
- Adventures. Site chairs would organize a set of daily adventures. For example, for ACM ICER 2021, which I’m currently program chairing, we brainstormed things like particular tourist attractions, places to exercise together, interesting shopping destinations, walking destinations. Anything that might fit into 3 hours—or even the rest of the day—would be suitable. And while on paper this might sound like “academics go on vacation”, everyone knows that no matter the reception destination, academics can’t help but talk about research. Remote attendees could optionally participate by being paired with in-person attendees, or having in-person attendees live stream the events.
- Collaborations. Rooms could be set up to allow groups of collaborations to work together on papers and new grants, with snacks, drinks, and whiteboards. Many attendees already do this, they just have to get creative about using space, reclaiming a corner in a hallway or an unused room. This would provide explicit support and encouragement for collaboration. Many in academia already have experience with hybrid remote collaboration, allowing others to join on video calls to contribute.
Dinner events. After the afternoon sessions, attendees would get a short break (for introverts to recharge—extroverted people would probably just keep talking, as they often do before dinner). The evenings would be a range of activities, including traditional receptions or even more adventurous planned outings to restaurants, nightlife, and more distant parts of the conference destination. They would vary in structure: some events might be watching a conference themed movie, others might be delivery and board games, others still might be quiet reading rooms with box lunches and small group chats. The point of these is surprise and serendipity, catalyzing new relationships, emergent ideas, and new community.
If all of this sounds like a two-tier system, that’s because it is: attendees would choose between paying substantially less for a virtual conference experience and opt-in remote participation in some social activities, or substantially more for a rich in-person event. To compensate for inequities in people’s ability to travel, the would be several options. Most academic conferences have seen large increases in attendees due to reduced attendance barriers; much of these excess revenue could be used to subsidize travel for first-time and student attendees, ensuring that newcomers and students can attend in-person to build strong ties with the core community. Ultimately, this would be a win-win: those that can’t attend in-person due to travel or time constraints can still participate at a certain level equally with other participants, but those that can travel can get added value from the rich, serendipitous social interactions, and even more of it than past conferences offered. The calculus would be clear: $100 for a nominal remote experience, and a $400 registration upsell, plus travel costs, for the irreplaceable benefits of collocated networking.
There are of course existing policies and inequities that would need to change. For example, would universities and funding agencies that subsidize travel allow their faculty to travel to get the added benefits of social interaction? Would advisors do the same? Organizations and advisors will have to decide what the point of conferences are: purely to gain access to content, or also to build networks? Clearly it’s both, but many institutions might resist this basic fact. Some institutions also have conditions on funding, like the presenters need to “present”: what it means to present in this model is murkier, since content is pre-recorded and Q&A is short. At a minimum, policies would have to shift to view in-person travel support as an incentive for publishing. Better yet, they would shift to viewing travel support as an investment, help a researcher build the relationships they need to succeed. And conferences would reciprocate, offering maximal benefits of in-person networking, rather than drowning attendees in content.
The other exciting thing about this design is that it deemphasizes content. I personally think that the long term future of CS is in journals; conferences as venues for archival content are just a bad idea waiting to die. This conference design creates a way to accelerate the end of this broken practice, allowing us to easily shift the point of conference from “present the papers we all just wrote and reviewed” to “quickly learn about all of the great research that’s happened in the past 6–12 months in our community.” I would love to see a program built out of multiple journals, even content from adjacent disciplines. The 2–3 hours each day would just be one big curated, compact rich interactive summary of discoveries, and we can let the research papers be the ground truth documents that they were always meant to be.
Since ACM ICER 2021 is virtual again this year, I won’t get a chance to explore these ideas. But I do hope that steering committees and conference chairs looking to 2022 will strongly consider some of the ideas above, preserving the benefits of virtual conferences, while also dramatically reconsidering how we use our collocated time. It’s never been more obvious how precious our time together is, and how we’ve squander it pre-pandemic times with presentations instead of interaction.
Would you want to attend a hybrid conference like this, remote or in-person? Why or why not? @ me on Twitter or write me an email and let’s chat.
