Dungeon creation

Players & Game Master (GM)

Before you even begin with any of the dungeon elements themselves, you'll want to consider what your players enjoy to play, and what you enjoy to run. If your players only enjoy a hack and slash dungeon, there's no point in creating a story driven, puzzle filled dungeon. At the same time, if you don't enjoy running a dungeon filled to the brim with combat encounters, you'll obviously want to go a different route. If you're not enjoying yourself, it'll affect your players, and vice versa.

Of course, if your players' and your own wishes don't align, you'll have to meet somewhere in the middle. This is easy enough to do, as dungeons generally allow for a wide range of different rooms and encounters. Plenty of chances to have a moment for both sides to enjoy, or perhaps even for every player if there's a wide range of interests within the group.

Main goal

Once you've figured out the desires of the group, it's time to focus on the main goal of the dungeon. The dungeon itself can be seen as a single encounter for this purpose, even if the dungeon probably has dozens of separate encounters. The main goal will act as the main foundation for everything else from the dungeon layout to the monsters within it, and from the story elements to the non-player characters (NPCs) inhabiting it.

The goal could be something like "Retrieve the ancient sword of Azrak", "Find and kill the galactic drug lord", or "Figure out the secrets of the Lancarsians". It's whatever the main quest of the party is or will be. This will also give you the ideal starting point for your dungeon, as you know you'll definitely need a location to resolve this quest, whether it ends up being a successful mission or not. After that you simply have to add to that room, add corridors, side rooms, guards, and anything else that may be applicable. Of course, that's not all there is to keep in mind.

Monotony

Dungeons can quickly become bogged down with elements too similar to each other, making everything seem the same, even when some of the minor details are different. This goes for everything from room types to monster types, from loot to traps, and even includes the pace of the game. I'll go over each of these elements in a second, but I first want to make a point of how important it is to avoid monotony. Monotony will quickly kill the fun of a campaign or game session, which is of course the last thing you want. But it'll also cause players to miss potentially crucial elements as their attention isn't fully with the game, so they might not pay attention to story elements you spent ages creating, and one or more players might become disruptive as a result of this boredom, further ruining the fun for the others.

Some dungeons are meant to be monotonous though, so how would you solve the issue there? The key is to make the dungeon monotonous only to the player characters, not the actual players. Change the pace, spice up some of the monsters, or add in some odd loot. There's still plenty to play around with in a dungeon designed to be clean, symmetric, and otherwise monotonous. But more of this will be covered in each of the individual parts below.

Rooms

You'll already have the final room, the part where the main goal is resolved. This doesn't have to be an actual room, but for simplicity's sake I will refer to all dungeon locations as rooms. From this main room you can build the rest of the dungeon, but some of you might struggle with the types of rooms you might want to add. This depends, to some extent, on how big you want your dungeon to be, and on the overall background of the dungeon.

Let's take my first main goal example, and use it to build a dungeon. We know the party has to find an ancient sword, and let's say the sword is buried with a long dead king and guardian of the realm. This means the sword will be in a tomb of sorts, and we'll say the tomb, as well as the entire system around it, was built underground, deep underground.
So we have our tomb room, which will be elaborately decorated. Now, this sword is a sword of legends, a powerful sword sought after by many people, including the king's enemies. So, clearly there needs to be some security down there. There will be two doorways on either side of the tomb room, each leading to a corridor with other rooms. Initially the tomb room was protected by the king's forces, so we know we'll need sleeping quarters, supply rooms, a mess hall, perhaps a kitchen, maybe even some jail cells, and bathrooms of sorts will be required as well of course. There will also be two paths leading out of the dungeon, as the people within will need to be resupplied from time to time, and it's a terrible idea to only have one way in and out.

Right, now where do we add all of this? Again, this can be answered by taking into account the purpose of this dungeon, which is to keep the sword of Azrak safe. This means you don't want one long corridor, it'd allow enemy forces far too much range, and once you're pushed back you'd be pushed back constantly until the end. So, twisting corridors it is. But what about the rooms? Since I imagine the king (or any military general for that matter) would want all the guardians to be on standby at all times, it'd make sense to have all the rooms close together, that way all guardians can act swiftly should there be an enemy or other event that needs taking care of. You don't want guardians to have to run for 10 minutes to get to a meeting point in case of an emergency.

So, all rooms will be close to the tomb room, but the most vital ones will be on the inside (storage, sleeping quarters, etc.), while the less vital ones will be on the outside. If an enemy does show up, you don't want them to be able to destroy your storages.
The dungeon will also have 2 or three levels. This keeps it more compact as rooms can be stacked on top of each other, and keeps travelling time shorter (think a fire station), and it means there can be more tactically advantageous areas.
Having different levels also means the player characters could potentially blast their way through a floor to go down, should the players think of this.

From here you can keep building and building, and this method is the same for all types of dungeons. But there's more than just adding rooms, there's story and history you can add as well. So let's briefly continue the already lengthy example.
The tomb room will have statues, murals, and other decorative pieces depicting the king's conquests. The hallways will have paintings and statues as well, the mess hall and sleeping quarters will have mounted heads of local creatures to show what this part of the universe has to offer, and to hint at the players what else they might encounter in the future. The other rooms will have similar small details reflecting the culture of the king at the time the dungeon was created.

Room monotony

Having "useless" rooms, like bathrooms and plain storage rooms, allows you to make the dungeon itself less monotonous, but make sure you skim down on their descriptions when players encounter them, or else they'll spend time in rooms they won't find anything in, which can slow down the pace of the game, and, to some players, make the dungeon seem boring.

Encounters

Of course, no dungeon is complete without some encounters. You probably already have the main encounter, assuming your quest ends with an encounter, but you'll probably want some more. The amount, type, difficulty, and other factors depend on your group, hence why you needed to consider your group's desires at the start. But after you figured out what your group wants you still need to give it in a fitting way.
The group might love social encounters for example, but when a tomb is abandoned this might not be as viable an option as combat or puzzles, but it's still an option with the right choices.

First and foremost you want your encounters to fit the dungeon. I've already written about this in the encounter specific guides, so I won't delve into this too deeply in this guide, but let's quickly add some encounters to the previous tomb example.
Those statues I added, they're optional encounters. They can be turned into animated suits of armor if I think the game needs some combat encounters. The murals, paintings, and other decorations could be used for puzzle purposes, perhaps they contain clues through story telling, or perhaps a password required to enter the tomb room is hidden within a painting somewhere.

If the players prefer a social encounter, and assuming the tomb has been abandoned for centuries, the spirit of the king or some guards could still offer such an opportunity. They might wish to see if the adventurers are worthy of entering, or they might need convincing to give up the sword for a good cause. Alternatively, perhaps the tomb's been claimed by other beings as their home, and they might need to be convinced to allow the adventurers access, or perhaps a trade can be made for the sword.

Encounter monotony

There are (almost) always multiple types of encounters you can use in a dungeon, so try to be diverse with what you pick. A specific kind of monster might seem like a perfect fit, for example, but if your group has encountered them several times before already, it might be wise to go with a different monster instead. It'll make this dungeon stand out more, and it'll make the encounters more interesting too. Even altering those same monsters could do the trick, perhaps making them look a little different, and giving them different abilities, or perhaps a dire, shadow, spirit, energy, or other type of version of that monster. There's plenty that can be done to spice things up.

Pace

Something to keep in mind too is the pace of the dungeon and overall story telling. If the focus is on exploring, mystery, and figuring out what the dungeon is, you'll want to avoid adding many combat encounters as they'll take away from the pace for mystery. If the focus is on trying to escape the dungeon, or some other kind of time pressure based event, many combat encounters can definitely be the answer, as having to deal with them takes up precious time and resources, which adds to the urgency to get out of that dungeon.

You can alter the pace differently as well, by taking longer or shorter when you describe the rooms for example. If your dungeon has a lot of rooms, but you need the pace to be quite fast, don't spend a lot of time describing all the rooms in full detail, especially not the less important ones.

At the same time, pace can help make the dungeon and all of its parts less monotonous. Did you go for the same monsters again after all? Describe their attacks differently, maybe even make them attack differently. Describe parts of them you haven't described before. Have them interact with the player characters in individually specific ways. As long as you add something the players didn't already know, these enemies, or any other elements for that matter, will seem less repetitive.

Loot

What's a dungeon without loot? Still a dungeon, but perhaps a disappointing one to some. Whether your dungeon needs loot or not will again depend on your group's desires, but the way you add loot can offer a lot of potential gaming fun too. Sure, giving loot in the form of carried items dropped by whichever enemies are killed is fun enough in and of itself, and there may be some loot stashed in chests here and there as well, but there's far more you can do with loot in a dungeon.

Loot can be hidden in secret compartments or rooms, loot can be hidden in plain sight (valuable paintings and other art for example), loot can be given freely at one point, only to be taken away at a later point. Loot can be false (fake gold meant to trick thieves for example), loot can be cursed as much as the dungeon itself is, and loot can tell a story.

The entire dungeon itself is part of a story, so why shouldn't the loot fit this as well? Gold's always welcome in most parties, but thematic loot can offer some fun backstory opportunities should the players desire them. My tomb example could have loot in the form of paintings depicting legends and stories of the tomb, the king, and anything else important at the time it was built. The legends themselves could prove incredibly valuable to scholars. Gold could come in an old and outdated currency, the statues might have valuable ornament weapons, and perhaps some of the storage rooms hold some utility items that could prove very helpful in some specific upcoming encounters. The value of loot doesn't always come in monetary value after all, a healing potion during an intense boss battle will be more valuable than a chest of gold, because what good is gold when you're dead?

Map creation

Most dungeons (and players) will benefit greatly from having a visual representation of the dungeon you've created, especially if it's big or if you're not that great at describing things. A visual representation will also help with keeping track of the party's location, which can be especially helpful between gaming sessions. Creating a map doesn't have to take much effort either, it could be as simple as crude drawings on a piece of paper. If you do want to make it look better, try getting hold of some graph paper or a similar type of tiled paper. The tiles will help with keeping things to scale, and at the same time can serve as actual movement tiles if your game uses them.

Alternatively, if you want to go a step further, try out the dungeon creator on this site. There's a whole range of icons and doodads available already, but I'll be adding more periodically in the future.

You can take map creation another step further though. You can have various versions of the same map, some with more information showing than the other. You can have layers, shifting parts, rotating parts, hidden compartments, and so much more. All of this can be done easily with paper too. Shifting parts require nothing more than separate pieces of paper. Rotating parts require nothing more than a pin to keep a rotating part in place, and hidden compartments can be as simple as sticking a bunch of small pieces of paper to the main map, which can then be folded back to reveal whatever's hidden. This does work best if there's lots of them, perhaps even the entire map, just to make sure players don't meta game and go for the obvious taped part first. This is also where the value of having multiple versions of a dungeon map comes in handy. Show the players only the one without the hidden latches, then when they do find them you can reveal them.

Spicing it up

There's still more you can do with dungeons. Traps are an obvious one, as are hidden compartments and passages, but you can add further mysteries and events than that. Shortcuts, side quests, more story lines, more history, expansions built by others at a later time, creatures who have taken over part of the dungeon, lost adventurers or other NPCs, remnants of previous battles or adventurers, and other, similar elements are the easiest to add, but will help make sure each dungeon stands out more.

But the dungeon itself could be part of the encounters too. Rather than simply being a background setting, the dungeon could be a crucial part of combat encounters. Small corridors, different height levels, broken bridges, rotting wood, crumbling stones, and all sorts of other elements could either make combat more difficult, or offer opportunities to use the dungeon against the enemies. A fragile wall in a corridor would be easy to take down, which could cause that part to collapse, which in turn could mean an easy escape.

Need even more spice? How about turning a dungeon from the ordinary to the extraordinary? Think outside of the box, think bigger about every element. Location? In the sky, or across several different planes, or underwater, or all of the above. Monsters? Unique to the dungeon, shaped by its elements, or perhaps oblivious to the world outside of the dungeon. Loot? Tainted or otherwise altered by the elements of the dungeon, or perhaps unusable outside the dungeon, or priceless and worthless at the same time, as nobody might know what the items really are, and therefore don't know their real value.
Every element can be turned into something really unique and special, but it's usually best to focus on only one or two of these, that way the dungeon overall is still grounded in the familiar, and it leaves you with some creative ammunition for future dungeons as well.

Bulking it up

Is your dungeon too small for your liking? Simply bulk it up with random fluff. Storage rooms, extra corridors, more sleeping quarters, more religious shrines, more studies, more libraries, and so on. There are always rooms or features you can add, which add to the overall size of the dungeon, but don't require too much in terms of describing and/or planning out. If you've covered every type of required room, try expanding to entertainment and/or utility rooms. The dungeon from the tomb example may have all the necessary rooms to provide the guards with a decently comfortable life, I could still add rooms for entertainment purposes if I want the dungeon to be bigger. A hot spring pool, a room with a stage for bards or plays, an archery practice room, a game room, a library, and so on.

Depending on the person or people who built the dungeon, you could also add a whole bunch of completely redundant rooms. Think about how huge mansions often have a whole range of bedrooms, multiple bathrooms, multiple dining rooms, multiple kitchens, and so on. Nobody really needs that, but they're still being built, as some people do desire them. The same can work for dungeons.