What a supervision meeting should actually produce

Ask a professor how their last supervision meeting went and you will usually get a verdict on the conversation: it was good, it was difficult, the student seemed lost. Ask what the meeting produced and the answer gets vaguer. That gap is the most common failure in research supervision, and it is not a failure of effort or goodwill. It is a failure to decide, in advance, what the meeting is for.

A supervision meeting is not a status update. If all that happens is a scholar narrating what they did since last time, you have spent an hour reconstructing a record you could have read in five minutes. The meeting is the one recurring slot where a more experienced researcher and a less experienced one think about the same problem at the same time. That is expensive and it is scarce. It should leave something behind.

Three things worth leaving behind

Almost every useful supervision meeting produces at least one of these, and the best ones produce all three.

A decision. Something was open at the start and is closed at the end. Which dataset. Whether to abandon the third experiment. Whether the chapter goes to the committee this term or next. Decisions are what supervision is for — the scholar usually has the information but not yet the judgement to weigh it, and that judgement is the thing you have that they do not.

A correction of direction. Not a decision so much as a nudge: you are reading too widely, you are reading too narrowly, the objection you are worried about is not a real objection, the one you are ignoring will sink the paper. These are cheap to say and enormously expensive to discover alone.

A concrete next action. Not "keep working on the analysis." Something with a verb, an object and a date.

If a meeting produced none of the three, it was a catch-up. Catch-ups have their place — a scholar who is struggling may need one — but you should know that is what you are doing, rather than discovering it a month later when nothing has moved.

Set the agenda from the last meeting, not from scratch

The most reliable technique in supervision is also the dullest: the agenda for this meeting is the output of the last one.

If the last meeting ended with three actions, this meeting opens with those three actions. Not as an interrogation — the point is not to check homework — but because it is the only way to notice a pattern. One action slipping is normal. The same action slipping three meetings running is a signal, and it almost never means the scholar is lazy. It usually means the task was badly defined, or it is blocked on something they have not mentioned, or they do not know how to start and are too embarrassed to say so.

You will only see that pattern if last month's actions are written down somewhere you actually look. Memory does not do this. Memory smooths: by the third meeting you have a vague sense that the analysis has been slow, rather than the specific fact that the same task has been restated three times in slightly different words.

Say the decision out loud, in the meeting

A surprising number of decisions in supervision are made implicitly. The supervisor asks a few sceptical questions about an approach, the scholar reads the scepticism correctly, and both leave believing the approach has been dropped — or, half the time, one of them believes it and the other does not.

The remedy takes ten seconds. Before the topic closes, say the decision in a plain sentence: so we are dropping the survey and going with interviews, and you will write the revised protocol by the 30th. This feels almost patronising to say. Say it anyway. It is the cheapest possible insurance against a month of work in the wrong direction, and the person most likely to have misread the room is the one with less experience and more at stake.

It also does something less obvious: it forces you to notice when you have not actually decided. If you cannot say the sentence, the topic is not closed, and it should go on the agenda for next time rather than quietly evaporating.

Write the record before the scholar leaves the room

Write the record in the meeting, or immediately after. Not that evening.

This is the advice most often given and least often followed, and the reason is that writing up a meeting properly takes fifteen minutes you do not have between back-to-back supervisions. But the cost of skipping it is not that the record is missing. It is that the record is wrong. A write-up done on Friday for a Tuesday meeting is a reconstruction, and reconstructions preserve the parts you found interesting and drop the parts the scholar found difficult.

Two practical options. Either write the actions live, in front of the scholar, in the last five minutes — which has the useful side effect of surfacing disagreement about what was decided while you can still resolve it. Or record the meeting and draft the notes from the transcript, which removes the fifteen minutes entirely.

Either way, the record needs to be somewhere both of you can see. A supervision record that lives only in the supervisor's notebook is a record of one person's impression of the meeting, and the scholar cannot correct it.

What this looks like in practice

A supervision meeting that works has a shape that is boring and repeatable:

• Open on the actions from last time. Five to ten minutes, not an interrogation.
• One substantive topic, chosen in advance. Not four topics touched lightly.
• Close each topic with the decision said out loud.
• Last five minutes: the actions, with dates, written where the scholar can see them.

The discipline is not in any individual meeting. It is in the fact that meeting eleven can be compared to meeting three, because both left a record in the same shape. That comparison is where supervision stops being a series of conversations and becomes a trajectory you can steer.