One link, no account: cutting setup to nothing

Part of a single argument: the engagement gap is a workload problem.

Participation has a funnel and it is steeper than most people running sessions realise, because the drop-off is invisible. You never meet the people who quietly did not join. They are simply not in the numbers, and the numbers look like disinterest.

Count your own steps

Write down every action between someone being told to join and actually being joined. Find the link. Open it. Install something. Create an account. Verify an email. Choose a password. Find the right session. Enter a name.

Each step loses a few people. Not many individually, which is why each one is easy to justify. Multiply the drop-offs and a five-step join can lose a third of a room before anything has happened.

The arithmetic here is worth doing explicitly, because the intuition is badly wrong. Five steps that each retain ninety per cent retain fifty-nine per cent overall, and ninety per cent per step is generous for anything involving a password.

The steps you can delete outright

Account creation, for participants, is almost always unnecessary and is reliably the single largest drop in the funnel. Anyone who has watched a room try to sign up to something has seen the moment where a third of it gives up.

Installs are worse and are very rarely justified for a one-hour session.

What should remain is a link or a short code. That is achievable, and where it is achieved it moves participation more than any improvement to the activity itself. It is also the least interesting work in this blog, which is probably why it goes undone.

Design for the worst realistic case

Someone joining on a phone, on weak signal, at the back of a room, while a presenter waits. That is the realistic worst case and in many organisations it is most of the audience.

Testing on a laptop on the office network is how joining flows end up quietly excluding a slice of every session. The flow that works for the phone at the back works for everyone; the reverse is not true.

It is worth doing this test once, properly, with an actual phone and the office wifi turned off. It takes ten minutes and it usually changes something.

The presenter's steps count double

Whoever is running the session also has a setup path, and theirs is performed under pressure, in public, with people watching and waiting.

Every step there is a chance of a visible failure, and visible failures are what teach people to stop trying interactive formats. The cost of a presenter-side failure is not the ninety seconds it takes to fix. It is the next six months of that person choosing the safe format.

So the presenter path deserves more scrutiny than the participant path, not less, even though it affects one person rather than fifty.

What good looks like

Participants: one link or code, no account, no install, works on a phone.

Presenter: one action to start, nothing to remember, no state left over from last time that can surprise you.

Neither of these is technically hard. Both are routinely not achieved, because setup is nobody's job and its cost is spread thinly across everybody.