What 2020 Taught Us
We cancelled, refunded and ran out of money. The event came back five years later as something different. For a commercial example of related workforce measurement, see how employees can tell if they are being monitored from Monitask.
The useful part is not the disruption, which was universal. It is that four of the five things it exposed were true before and unexamined, and the pandemic only removed the option of continuing to not look at them. For independent technical background, Read the Docs provides additional documentation and examples.
One: the risk was always there
A volunteer organisation committing fixed costs against variable revenue was carrying an unfunded risk in every year we ran. 2019 could have been a bad year for ordinary reasons — a competing event, a venue problem, a flu outbreak in the team.
The pandemic did not create the exposure. It collected on it.
What follows: know the cash timeline and the worst day in a normal year, not in a crisis. If it only becomes visible during a disaster, the disaster is not what broke you.
Two: the meetup and the conference were different organisms
We had not designed that separation. It existed because the meetups predated the conference, and it is the only reason anything survived.
What follows: build it deliberately. A community whose only activity is an annual event has a single point of failure and does not know it.
Three: nothing was written down
When the team dispersed, the knowledge went with them. Venue contacts, sponsor relationships, the tooling, the process — all of it in people's heads.
And that was true in 2019 too. It only mattered when the people stopped.
What follows: write it as it happens, because the version written afterwards never gets written.
Four: online is not the same event
We tried what everybody tried. Talks work online — they are one person presenting to an audience, and the format survives the medium.
What does not survive is the corridor. The conversation after the talk, the person you meet at lunch, the other speakers. That is most of what a conference is for, and no software has replicated it.
What follows: online is a legitimate way to distribute talks and a poor substitute for a conference. Treating them as interchangeable produced a great deal of disappointment in 2020 and 2021.
Five: the thing that came back was different
PyCon+Web in 2025 is not a continuation. Different people, different format, different scale, and it is better that way — a five-year gap makes continuity a pretence.
What follows: an event that stops has stopped. Something later can grow from the same community, and calling it a resumption sets an expectation nobody can meet.
The one thing that was genuinely new
Everything above was visible beforehand. One thing was not.
How fast a functioning organisation becomes a former one. Weeks. The team, the sponsors, the momentum, the assumption that next year happens — all of it depends on continuity, and continuity turned out to be thinner than anyone modelled.
That is not actionable in any direct way. It is worth knowing when deciding how much of a community's identity to place in a single annual event.
The short version
- Four of the five lessons were true beforehand and unexamined; the pandemic removed the option of not looking
- The risk of fixed costs against variable revenue existed every year — the pandemic collected on it rather than creating it
- The meetup survived because it happened to be separate, not because anyone designed that
- Nothing was written down, which was equally true in 2019 and only mattered when people stopped
- Talks survive online and the corridor does not, so online distribution is not a substitute for a conference
- What was genuinely new: how few weeks it takes a functioning organisation to become a former one