How long should D&D prep take, and where the time actually goes

What DMs actually report

The largest public figure comes from a Sly Flourish reader survey of 3,663 DMs, which asked how long they prepared for a roughly four-hour game. The answers spread almost evenly across the range rather than clustering anywhere: 10 per cent prepared for 30 minutes or less, 33 per cent for about an hour, 28 per cent for about two hours, and 29 per cent for three hours or more.

Two things follow from that. First, there is no norm to fall short of — a DM prepping for 45 minutes and a DM prepping for three hours are both squarely inside what most people do. Second, the top and bottom of that range differ by a factor of six for the same four hours of play, which is far too wide to be explained by some DMs being more thorough. They are doing different activities, not the same activity at different speeds.

One caveat, stated plainly because the number is doing a lot of work on this page: that survey ran in April 2020. It is the best-sized public data on the question and nothing more recent of that scale has replaced it, but it is several years old and it surveyed one blog’s readership rather than all DMs. Treat it as a good indication of the spread, not a precise measurement of the hobby.

The question that is actually useful

Asking "how long should this take" assumes prep is one activity with a correct duration. It is at least four different jobs, and they have very different returns.

Work out which of these your hours are going into before trying to cut them. Most DMs who feel prep is out of control are spending the bulk of it on the last two, which are the two that feel most like real work and change the session least.

• Deciding what is likely to happen next and what you will need for it. High return, usually fast — this is the prep that actually changes how the session runs.
• Reconstructing what happened last time, so the world stays consistent. Necessary, and almost pure overhead: it produces nothing new.
• Making things you will probably use — a statblock, an NPC voice, a rough map. Medium return, and the part most likely to be rewarding for its own sake.
• Making things in case you need them. Lowest return by a distance, and the easiest to mistake for diligence because it produces the most visible output.

The hour you cannot see yourself spending

The second item on that list is the one worth looking at hard, because it rarely feels like prep. It is the twenty minutes before you start prepping properly, when you are scrolling back through notes trying to remember the name of the guard captain, whether the party ever paid the smith, and how the last session actually ended.

It does not feel like a task, so it does not get counted, and it does not get optimised. But it is overhead in the strict sense: you finish it knowing exactly what you knew at the end of last session, and not one thing more. The prep that changes the game has not started yet.

It also gets worse over time in a way the creative parts do not. Session 30 has thirty sessions of continuity behind it and the same twenty minutes will not cover it any more. This is the mechanism behind the common experience of prep quietly expanding across a long campaign while the amount of actual invention stays flat.

What genuinely reduces it

In rough order of how much time they return for how much they cost to adopt.

• Prep decisions, not content. A list of what is likely to come up and what you will do about each is faster to make and more useful at the table than a written scene, because the party will not do the scene.
• Stop preparing contingencies. The branch you did not prepare is the one they take, and improvising it costs less than having prepared three branches you threw away.
• Write the recap immediately after the session, not before the next one. Ten minutes while it is fresh replaces the twenty-minute reconstruction above, and it is a better recap — see writing a session recap.
• Keep names somewhere findable. Most of the reconstruction problem is proper nouns, not plot: you remember what happened, you cannot remember what the innkeeper was called. A flat list beats a beautiful wiki you do not maintain.
• Reuse. A statblock with a new name is not cheating, and nobody at the table can tell.

Where recording the session fits, and where it does not

If the reconstruction overhead is where your time goes, a recording of the session removes the cause rather than speeding up the cure — the answer to "what was the guard captain called" stops being a memory problem and becomes a lookup. That is the honest case for it, and it is a narrow one. Practical setup is in the recording guide, and if you play online, on Zoom or Discord.

What it does not touch is the prep that actually matters. Nothing about having a transcript decides what the villain does next, or whether the party’s bargain with the knight should come back to bite them. That is the creative half, it is the half worth spending time on, and it is untouched by any tool.

It is also worth being clear that recording adds a small fixed cost — remembering to press record, and asking your table’s permission, which is a real conversation and not a formality (why it matters). If your prep is already 45 minutes and none of it is reconstruction, this trade does not pay.

A reasonable target

If you want a number to aim at rather than a framework: an hour of prep for a four-hour session is comfortably within what most surveyed DMs do, and it is enough time to decide what is likely to happen and what you will need, provided you are not also spending it reconstructing last session.

If you are well past that and it does not feel good, the diagnostic is not "prep less". It is to spend one session’s prep noticing which of the four jobs above you are actually doing. Most people find the answer surprising, and it is usually not the creative part that grew.

And if you are past it and it does feel good — if worldbuilding on a Wednesday is why you run the game — then prep is not a cost to minimise and none of this applies. That is a genuinely different hobby from the one this page is about, and it is not the worse one.