A first conference is mostly a social problem wearing a technical costume. The talks are the easy part: somebody speaks, everybody listens, and nothing is required of the audience. Everything that makes a first conference worth the ticket happens in the gaps between them, and that is the part nobody explains in advance.
Before the day
Three things are worth doing in the week before, and each takes a few minutes. Read the schedule and pick a first choice and a second choice for each slot, so the day does not begin with a decision. Find two or three names of people attending, from the speaker list or from wherever the event is being discussed. And check the practical details, which are covered on attending a conference.
What not to do: prepare questions. It reads as sensible and it produces a day spent waiting for the moment to ask a rehearsed question rather than listening.
Arriving
The single most useful decision is to arrive for the opening rather than for the first talk. A room where everybody is standing about with coffee is far easier to walk into than a room where everybody is already seated in rows and a talk is under way.
Registration is the natural first conversation, because the person on the desk has to talk to everybody and is not deciding whether to. Asking them what they would go to is a real question with a real answer, and it starts the day with one exchange already behind.
Arrive early. An empty room is easier to enter than a full one.
The talks, and how to choose between them
Choose the talk closest to the thing already being worked on, not the most advanced one. A talk about a problem already met lands completely; a talk two levels above lands as a list of words to look up later, and the looking-up rarely happens.
On not understanding a talk: a substantial part of every audience does not follow a substantial part of every talk. This is invisible because nobody says so, and the belief that everybody else understood it is the most reliably false belief in the room.
Sitting near an aisle is worth more than it sounds. It makes leaving a talk that turned out to be wrong a small act rather than a public one, and leaving a talk is entirely normal.
Lunch, and the rest of the unstructured time
The second peak on the chart, and the one with the simplest fix: join a table where somebody is already sitting rather than looking for an empty one. Asking what somebody works on is a question everybody at a conference is prepared to answer, and it is the standard opening for a reason.
The conversation that follows is frequently more useful than the talk that preceded it, because it is about the specific problem in hand rather than a general account of somebody else’s. That is the whole argument of the hallway track.
Afterwards
The day produces two things worth keeping: a short list of things to look up, and two or three names. The list decays within a week unless it is written down on the day, and the names decay faster unless something is sent within a few days.
What to send is not a request. A message saying which part of a conversation was useful, with a link to the thing that came out of it, is enough and is answered far more often than a request for help is. That is the mechanism by which a first conference turns into a second one with people to find at it.
For the practical side of getting there and back, attending a conference covers it. If a first conference makes standing up look appealing, first-time speakers is the next step, and the newcomer track is what an event should be doing from its side.
Questions
What should somebody do at their first conference?
Arrive for the opening rather than the first talk, ask two people what they work on before lunch, choose the talk nearest to their own current work, and write down the list of things to look up before leaving the building.
Is it rude to leave a talk partway through?
No, it is normal and organisers expect it. Sitting near an aisle makes it a small act rather than a public one. A talk that turns out to be aimed elsewhere is not worth forty minutes out of a day that has a fixed number of them.
What if the talks are too advanced to follow?
A substantial part of every audience does not follow a substantial part of every talk. Nobody says so, which is why the belief that everyone else understood it is so common and so reliably wrong.
How do people start conversations at a conference?
By asking what somebody works on. It is the standard opening because everybody present is prepared to answer it, and joining a table where somebody is already sitting is easier than finding an empty one and waiting.
What is worth doing after the event?
Write the list of things to look up before leaving the building, and send two or three short messages within a few days saying which part of a conversation was useful. A note that asks for nothing is answered far more often than one that asks for something.