1st Virtual Tech-for-Good Hackathon: After Action Report

DemocracyLab is a US nonprofit empowering people who use technology to advance the public good. We have been convening tech-for-good hackathons in the Seattle area every other month since mid-2018. On March 3rd, facing mounting evidence of the impending coronavirus pandemic, we made the decision to cancel our in-person gathering occurring on March 14th and to convene our community online instead. This decision proved to be prescient, as on March 11th Seattle closed public schools and banned public gatherings. On March 13th a National Emergency was declared.

We were forced to rapidly adapt to changing circumstances and figure out how to mirror the familiar flow of our in-person events in a virtual environment. Our pivot was successful, with 145 people checking in over the course of the day to contribute to the 14 tech-for-good projects that participated. These attendees spent a total of 407 hours on video calls and made many additional contributions on asynchronous tasks.

Background

The 2020 St. Hack-trick’s Day event was the 10th bi-monthly hackathon DemocracyLab has convened, but the first to occur online. Our hackathon philosophy is a bit unconventional, eschewing judges and prizes and instead encouraging teams to focus on making incremental progress on long-term projects, and helping volunteers establish long-term relationships with their project team.

We’ve come up with an agenda for our hackathons that works well for us:

8:30 Registration and Coffee

9:00 Welcome and Overview

9:15 Project Pitches and Team Formation

10:00 Teamwork

12:00 Working Lunch

4:00 Project Presentations

5:00 Closing Comments and Adjourn

The day goes fast, but it’s amazing what project teams are able to accomplish, and how important the events have been in establishing enduring volunteer relationships.

Choosing Tech Tools

When establishing the requirements of a suitable online platform for our March 14th event, we first considered the requirements of the physical venues we’ve used:

  • A space large enough for everyone to convene
  • Enough breakout rooms and/or tables for teams to work
  • Collaboration tools for teams to use

We decided to set up a QiqoChat group for our hackathon, a platform that piggybacks on Zoom to simulate an unconference environment online by giving participants the freedom to move themselves at will between different rooms. Each room has a Zoom video call and embedded Google Docs, Post-it Notes, or other collaboration tools. We configured our meeting space with a room for the General Sessions, a breakout room for each project team, a dozen extra rooms for smaller groups to split into as needed, and specific rooms for user testing and tech support. The platform can accommodate up to 1,000 participants, significantly more than we anticipated attending.

Each project had prepared a Google Doc in their breakout room with the key information participants would need to efficiently onboard to the project, including messaging tools, code repositories, design tools, shared drives, and more. The Docs also linked to the projects’ profiles on DemocracyLab, which offered even more detail about the project’s purpose, history, and progress.

On the Monday before the event, we asked a number of project leaders and volunteers to join us online to stress test the platform. We walked through the agenda for the day, with an emphasis on the transitions from one type of activity to another. This helped us identify numerous points of friction in the user experience, and to make important adjustments to our workflows and instructions.

Hackathon Day

We “opened the doors” at 8:30 am Pacific time, and shared music and photos of past events to set the tone for the day. At 9:00 am we welcomed participants and began the program. This video recording captures the morning session, where expectations were established and projects were pitched. We gave projects the opportunity to record a one-minute video ahead of time, or to do a live pitch using a single slide that described their project, the scope of the day’s work, and the skills needed on their team.

Once the pitch process was complete, participants sorted themselves into project teams and entered the breakout rooms that had been established for them. The graph below shows how participation in the QiqoChat platform varied over the course of the day.

This chart shows how the Zoom video minutes were allocated among the different rooms. Note that some groups chose to make greater use of the video call functionality than others, so time on the video calls does not necessarily correlate to the effort or productiveness of the group.

At 4:00 pm, groups were asked to reconvene in the General Session for final presentations. Team representatives were given three minutes each to share their progress over the course of the day. Representatives used Zoom’s screenshare feature to present directly from their machines. This video captures the final presentations and the accomplishments of each group.

Lessons Learned

As a whole, we were very pleased that we were able to continue our community’s forward momentum by pivoting from an in-person hackathon to an online event in 11 days. Convening online had some significant advantages, most notably transcending geography. While we’ve had a few project leaders fly into Seattle for past events, the online event was by far the most geographically diverse, with participants from across the US, and one each from England and Germany.

The low barrier to entry for online events did, however, create a problem we didn’t anticipate. Because we’d posted links to the event online, a number of people poked their virtual heads in to see what was going on, but we hadn’t planned ahead for how to orient latecomers to the work in progress. In the future, we will make sure that we post resources during the day to better orient late arrivals so that they can catch up and contribute effectively.

The flipside of the low barrier to entry is a low barrier to exit. It’s likely unavoidable that in an online setting people will have more distractions than if they’re physically in the same space, but something we need to think more about is how we can strengthen the group’s camaraderie during the course of the event. One decision we will revisit relates to intergroup communication tools. We decided to forgo setting up a Slack space for the event with channels for each team, believing that it is more useful long-term for each project to use their own messaging tool of choice. While this is likely correct, it presented a drawback on the day of the event because there was no easy way to send messages to all attendees. This lack of connection may have hurt our participation late in the day, particularly when we asked all participants to reconvene in the General Session for final presentations.

One smart decision we made was to simulate the hackathon experience ahead of time. Our pre-event stress test helped us anticipate and avoid problems our participants were likely to have. The technology worked well on the day of hackathon, and we ended up having little demand for the technical support resources we had established. This likely would not have been the case if we hadn't performed this simulation and made adjustments based on our experience.

We anticipate that the coronavirus pandemic will likely force us to hold our May 9th event online, but with more time to plan and equipped with the lessons we’ve learned, we expect to be able to improve our execution and increase participation. Our goal is to effectively combine in-person and online events once life returns to normal. In the meantime, if you’re working on a tech-for-good project or want to find one to contribute to, please join us online in May to contribute your talents to make our world a better place.