Experience report and lessons learned
Due to You-Know-What, the 15th European Conference of Technology-Enhanced Learning (ECTEL) in joint cooperation with DELFI conference (ECTEL+DELFI=ECTELFI) could not be held, as planned, at Heidelberg University of Education, Germany. Instead, it took place fully online, last week from 14th to 18th September 2020. In contrast to some conferences earlier this year (e.g. LAK20), we had ample time to prepare a virtual conference. And because conferences are not only about the presentation of research, we wanted to try to also transport the social experience of a conference into the online space. To do this, we used gather.town, a location-based videoconferencing too. It worked like a charm. We experienced some genuine conference moments and people really seemed to like it. In case you’re also thinking about how to enhance the social aspects of your online event, here’s some information as to how we implemented gather.town as well as lessons learned.
What is gather.town?
Gather.town is a location-based videoconferencing tool. This means that you have an avatar which you can navigate through an online space. Once you’re in proximinity to somebody else’s avatar, videoconferencing pops up and you can start speaking. As you move away, videoconferencing disappears. Socially, this means that as you walk around the map, you might bump into people, known or unknown, stop for a brief chat before heading to the next presentation, ask the keynote speaker some questions after their presentation, or just hang out in a group and enjoy an evening drink. Aesthetically, it is low-fi. The avatars and the online spaces are designed in 8-bit pixel art (see picture below). You can choose from many different pre-configured maps, design your own map (mapmaker), or you can commission custom maps. Gather.town is free for up to 25 persons at one time, for larger capacities of up to 2000 people, you can book premium packages of 2hrs, 8hrs, 24hrs and one-month.
How did we implement gather.town?
Introducing the tool. ECTELFI20 was split into a pre-conference, consisting of workshops and doctoral consortium on Monday and Tuesday and a main conference, from Tuesday evening to Thursday. We decided to use gather.town only for the main conference, as this is when we expected to see the most traffic. Also, this would give us a chance to introduce the tool at the opening of the main conference, where most participants would be joining. Accordingly, we booked two 24hrs premium slots. At the opening session on Tuesday evening, conducted via Zoom, we introduced gather.town by taking a tour of the conference map and demonstrating its main features.
Custom map. Although gather.town has an increasingly large array of preconfigured maps, we felt that none of them was exactly what we wanted. For example, the “Professional Conference”-Map had a really nice and slick design, but seemed much too big for our purposes. Also, according to conference program, we needed exactly six session rooms and one keynote room. So, we commissioned a custom map for the conference. Our map designer, Phillip, was friendly and quick, usually implementing small changes within 24 hours. In total, the map work was done within two weeks. Below you can see the map request we first sent to Phillip as well as the final map that we used for the conference.
Zoom integration. Although the videoconferencing of gather.town appears to be very stable and smooth, it was clear from the very beginning that ECTELFI20 would be using Zoom as main videoconferencing tool, through which all sessions, keynotes, and other events would take place. This was because, over the course of the last months, Zoom has become a de-facto standard and we did not want to take any chances by messing with this recipe. Fortunately, the developers of gather.town got this figured out with their “external call” integration feature. In our custom map, upon entering the session rooms and the keynote room, participants were prompted to join the Zoom call by pressing X. This pauses and mutes the avatar while participants are in Zoom. Once the session is over, participants can close the Zoom tab and be transported right back to the conference map in gather.town.
Think smaller. In total, we had about 250 registered participants visiting the conference. As registration remained open, we had trouble estimating how large the conference would actually be in the weeks leading up the conference. Moreover, we had absolutely no clue how many of these participants would actually join the social experience in gather.town. Although we planned on strongly encouraging the use of gather.town at key points during the conference, participants could just as well directly navigate to Zoom sessions by clicking on the URL’s provided. As a result of some deliberation, our custom map was then designed to comfortably accommodate approx. 100 simultaneous participants. This turned out to an overestimation. At the most busy times, that is after keynotes and in the breaks between parallel session, there were not more than 45 participants present at any given time. Yet, this was enough to allow for a truly social experience as was intended. For example, participants bumped into the keynote speaker Linda Castaneda after her presentation, where the discussion continued and they took a virtual selfie together (see pictures below). But even at these times, large parts of the map remained empty. For example, we had meeting rooms on the left-hand side of the map and session rooms on the right-hand side. As most traffic happened by navigating to the sessions, the right-hand side was where all the action was happening, while the left-hand side felt somewhat deserted. So as general advice, think small and plan for much less people than you would first intuit so that you’ll create a social space that feels lively and busy.
Incorporate into program. We first introduced gather.town during the opening of the main conference. Over the course of the conference, we repeatedly prompted participants to join us in gather.town during breaks and in the evening social events. Although this seemed to have worked to some extent, we may have just solicited the participants who would have joined anyway. To make sure that more resistant participants at least try it out once and see the benefits, we could’ve incorporated it even more into the program, for example by scheduling a Q&A with keynote speakers that can only be joined through gather.town or holding a virtual race through the conference map with a price for the winner. The more this tool is incorporated into the program, the more participants may feel that participation in gather.town is central to the virtual conference experience.
Prepare for minor awkwardness. Although the place-based videoconferencing worked extremely well on a technical basis, some psychological challenges of approaching other people and joining conversations may still remain. More than once, I noticed somebody’s avatar hovering right around the area where videoconferencing pops up, seemingly unsure about joining the conversation. In these cases, just like in real life, these issues are easily solved through a hearty “Hello” and a smile. In another situation, you might have your gather.town tab open but you’re actually in the kitchen cooking coffee. Another participant will then see a non-idle avatar as he/she approaches but will be looking at an empty chair instead of another human as videoconferencing pops up. However, these minor quirks are to be expected when attempting to transport a social experience into the virtual space.
Beware of switching Zoom-URL’s. Each of our session rooms had one static Zoom URL, so that at any given time, if somebody walked into the session room and clicked X, he would either join an ongoing session or see that it is currently closed. But in all cases, it was essentially the same Zoom room. This means that once these URL’s are implemented in the map, there was no need to adapt these during the conference. However, for keynotes this situation was slightly different as we had different Zoom URL’s for technical reasons, yet they were all scheduled to take place in the same room in the conference map. As a result, URL’s needed to be swapped several times during the conference. To do this, we needed to access the mapmaker and change the URL in one of the “external call integration” objects, an invisible feature in the map. Changes are implemented instantaneously and once figured out, it was really easy. However, for larger conferences this type of double-booking may quickly result in a lot of work and a major source of error. So keep in mind that, ideally, each Zoom-URL that will be used in the conference should have a dedicated virtual space with as little switching around as possible.
All in all, gather.town was a great choice and served our purpose well. The feedback was overwhelmingly positive, although we wished that even more people would’ve joined. I hope the above lessons learned help you if you’re planning your own virtual event with gather.town. Also let me know if you think I’ve missed something important!