In my (now long-running) D&D Next/5e game, the players explored a dungeon called the Monastery of the Blessed Scroll (layout based on Fountains Abbey in Yorkshire). I run very few dungeons that take much more than 2-3 sessions at most to explore, so this post documents what I did, and what parts of that worked or did not work. Players who were involved in some of those eleven sessions should feel free to chime in with comments. A lot of this post concerns the unusual-to-me format of this campaign, so I'll start there.
In the Aurikesh campaign, the players belong to a mercenary company called the Gallant Shields of Chardecum, so that there will never be a shortage of adventure hooks and I will never have to spend a lot of time resolving why these characters have teamed up to go adventuring. The campaign has never had a particularly stable roster - we just play with whoever can make it that night, as long as we didn't end the previous session in the middle of an adventure. There have been 24 characters played to date, among 15 PCs (several players have two or more characters). To make sure new characters can hang with more experienced ones, the campaign has an XP cap that I occasionally increase - this also encourages players to create and play alts. Gameplay has alternated between one-session adventures and arcs of three or so sessions; eleven sessions in the same dungeon represented a major shift.
When I started the Monastery of the Blessed Scroll, I knew that all of the action would be relocated to the area, as a large group expedition. I didn't want to have one group of characters at the monastery and another group back in the city, because (among other things) interspersing the dungeon exploration with ongoing adventures within the city would stretch out exploration over... probably years. And, well, I don't need to compete with myself for adventures to hold the party's attention. So the mercenary company sends a team of PCs to establish a base of operations with a supply depot that will sustain the whole expedition.
The supply wagons held a certain amount of generic Supplies, expressed in a silver piece total, that the PCs would use to pay their weekly upkeep during the expedition. I had imagined that, presented with a colossal amount of supplies, the PCs might pay higher weekly upkeep values, as Aurikesh grants modest benefits for paying more than the minimum upkeep. They did not; in hindsight it is obvious that they saw the supply total as a timer on the whole expedition, and with no way to know how long the expedition might need to reach completion, they didn't see the appeal in increasing the risk of having to go home in defeat. Completely reasonable! (Also I probably need to juice up the rewards for upkeep a bit, but that's another story.) The Supply did also pay the expensive material components for spells (identify, we're looking at you here) and copying spells from one wizard's spellbook to another. Toward the end of the adventure, the PCs pretended to use the remaining Supply as a bargaining chip in dealing with a large force of partisans (this, not this).
The whole expedition was carried out at the behest of a priestly order within the city, which had controlled the Monastery some fifty years earlier before it rather mysteriously cut off communication with the rest of the order and apparently fell into ruin. The order had not previously had the money or influence to mount the necessary expedition, but now they wanted someone to recover the Blessed Scroll from the heart of the monastery. The good news - and what made this kind of an unusual adventure - was that they had a detailed map of the monastery from before its ruin (the map above), and gave the PCs a copy.
Now, the map doesn't have a ton of practical information - after all, the PCs knew that there was a half-century gap between the map's creation and the present day, so anything could have happened. There were also underground chambers not included in the map. Still, there are some reasonably evocative names and a general shape of the layout. The benefit I reaped from this was a bit different than I had been imagining - I had not really examined how little practical information is there. On the other hand, with a constantly shifting player roster (including two players who joined the game after the company arrived and left the game as the company departed, so they never saw any of the campaign that was not the monastery), it was extraordinarily useful to have a quick way for more experienced players to explain what was going on and what area they planned to explore next.
The other great thing was watching the players mark up the map with notes on what they found in various locations. This is a scan of its final form:
As you see, the players used the map prop as a living document of their discoveries. I love that the characters still have the map afterward, as visible evidence that interesting things happened earlier in the campaign. Even things they weren't personally there for give them a sense of connection to members of the company they never met (I hope).
The shifting party roster meant that each foray from the base camp into the monastery ended with a return to the base camp at the end of the session, with the exception of a two-session exploration of Ravin's Tower. On one hand, this was necessary; I didn't want to tackle the complications of introducing new characters to a party deep within the monastery. On the other hand, it limited the scope of danger, the "hand in the cookie jar" feeling that you may have overcommitted to this particular foray and will have to fight your way out. It also coexisted awkwardly with my desire to make the dungeon feel "alive" by having NPCs repopulate (really, keep patrolling) areas where the PCs had previously had a battle or two.
For example, the Cloister of the Sword was the site of three or four combats, and because the players weren't always laser-focused on gameplay (lots of table talk, etc.), sometimes that battle was the majority of the session's action. When this happens for two sessions in a row, it's easy to give the impression that the adventure as a whole isn't advancing, even though from my behind-the-screen perspective they're ticking down the number of NPCs who are in the whole complex. Since the players were the ones deciding what part of the monastery to explore in each session, I largely absolved myself of making sure they saw something new each time, but I still could have handled this in a more satisfying way - after all, it wasn't obligatory for there to be a patrol at just the time the players passed through. (My thinking at the time was that greater resistance from NPCs would communicate something interesting to the players. This did not exactly come across as I had hoped.)
I should have come up with more conversational NPCs to populate various areas of the monastery, though I had a bit of a tough time justifying that. One of the better ones was the wizard Ravin, who the PCs found in his tower. He had been turned into a lemure by a warlock's curse, and he was dripping down the inside wall of his tower over the course of decades. The PCs managed to communicate with him, and eventually to free him from the curse by transferring his soul from the lemure-glop to a tree. There was also a trip into the Ghostlands (Aurikesh's Shadowfell, more or less) version of the monastery, where the got to talk to the souls of the clergy who had once lived there. There were also fey who took an interest in the PC warlock's deeds, a spirit who helped them pass into the Ghostlands in the first place, and a few partisans who were willing to talk. Still, I should have spent more time on interesting reasons for NPCs looking for something other than a fight to show up.
The story of the infernal warlock cabal that used the monastery for their base of operations was a pretty good one, I think. The journal of one of their recent victims delivered a lot of useful exposition as to the nature of Hell in Aurikesh and the cabal's particular function. More focus on the cabal members as individuals would have been good, but doing so probably would have dragged the action out over still more sessions, which would have been unfortunate given my aforementioned move to another state.
Over the course of the PCs' time there, the warlock cabal did something Bad that caused a minor earthquake and released a new threat from below the monastery. A new set of more freaky-weird monsters started showing up after that, and the PCs fought them a few times, but I didn't communicate something or other about this well enough, so the PCs largely shrugged and assumed that these monsters were normal around here, or otherwise just weren't all that interested. The monsters were just something else standing between the PCs and the Blessed Scroll, and I think by that point they knew they were getting close and didn't want to get distracted from that prize. Having a hook or weird thing get ignored isn't a big deal - they still had a good time and took interest in things, and I'm not attached to the thing they glossed over. Maybe it'll show up later in the campaign, but if it doesn't, no big loss.
Helping Ravin was one side-quest I introduced in the course of the expedition. Another was the partisans: they were commanded by Master Sapphire and Lady Sapphire, who the bounty hunters in the group recognized as incredibly high-ticket bounties (500 gold apiece is a staggering amount of money in Aurikesh), who were further famed for wielding particular magic weapons. The PCs did take interest in that, because a pile of money like that is not to be overlooked, but they didn't make it a priority until they had the Blessed Scroll and were ready to leave. The PCs were badly outnumbered by the partisans, but when they actually engaged them in battle, they had a tactical advantage and some help from Ravin's spirit, resulting in a victory snatched from the jaws of defeat. That segment of the overall adventure worked out pretty well, I would say.
In conclusion, there were definite good and bad parts of the whole run of the dungeon. On the plus side, players mostly seemed to have a good time and engaged with some parts of the setting that had not previously come up in-play, such as Hell, the Ghostlands, and the partisans that live in the wilderness just a few days from the company's home city. On the down side, I can clearly see places I could have made the experience better if I had put more into it - more talking, more varied fights, a richer story. The experience didn't wind up feeling all that much like a dungeon, per se; I should have found more ways to emphasize a sense of threat.
I think the map helped a great deal in making the complex feel concrete and consistent, and I want to experiment further with giving the players a lot of information up-front with what they will face, so that they can plan, and those plans can be subjected to contact with the enemy. It'll be awhile before I run another extended dungeon, I think. My style of DMing and the episodic way I run Aurikesh mean that urban intrigue gets to be the majority of the campaign's content.