Time determines the relative value of other resources, and thus shapes the style of play in the game.
The "fifteen minute adventuring day" becomes a viable strategy when, outside of an immediate encounter, time is cheap. There's no need to economize on the use of spells and other daily powers, because at virtually any time you can call a halt and rest for eight hours, and all your spells are restored. What's more, cheap time also cheapens any resource which can be acquired or restored by expending spells, including hit points. You can be less discriminating about fights when you know you can retreat and camp for a day or two and completely replenish your hit points before heading back into the dungeon. And when spells are cheap, temporary magic items like potions and scrolls become both extremely precious by comparison and next to worthless because players seldom want to use them. Cheap time also makes equipment management incredibly easy. So what if it's two weeks' walk back to town for iron spikes and holy water?
A "cheap time" game will probably tend to feature a lot of combat because of a reduced incentive to conserve hit points over the long haul of an adventure, and spell-casters may overshadow fighters and thieves because of a reduced incentive to economize on spell use, which would otherwise be a balancing factor.
Conversely, when time is dear, resources which are renewed by time become much more precious. If taking an extra day to rest has important consequences (see below for more on ways to accomplish that) spell casters must conserve their spells and use them judiciously rather than "going nova" at the first combat encounter to show off or show up the non-casters in the party. Fighters must likewise be more cautious about charging into combat when their hit points must last a whole day and not just one or two encounters. Temporary magic items can mean the difference between victory and defeat when the party faces depletion of spells, hit points, and other resources but can't afford to take a break.
A game in which time is precious will probably tend toward a more circumspect approach to combat and a more systematic approach to exploration. Play may also be more cautious at first, and become more reckless or even desperate as time grows short.
There are two basic ways to make time more costly. The first is PC-centered: Make things happen that affect the PCs directly as time passes for them. This includes rolling for wandering monsters and other random events with the potential to threaten the party and whittle down their resources. One advantage of this method is that it doesn't require careful long-term tracking of time. If you roll every other turn, you only need to remember whether or not you rolled last turn.
That is, of course, a perfectly good method for raising the cost of time a little, but it has its limitations. It adds an element of the unexpected, but if the party gets a little chewed up, they can still retreat and rest with impunity. To really put the pressure on, things must happen outside the immediate sphere of influence of the player characters, external events that affect the success or failure of the party's goals, whether simply gathering loot or some more story-oriented objective. That does require careful long-term tracking of time - hours, days, or even weeks, depending on the scope of the adventure and the campaign.
Making adventures that are truly time-sensitive is a much stronger way of applying time pressure, while bolstering and complementing the PC-centric method. External time pressures are generally driven by in-fiction concerns rather than game mechanical ones. A few examples, to stimulate the imagination:
- The evil cult must perform their ritual sacrifice of a princess under a specific aspect of the stars, which occurs at a specific date and time.
- The king is gravely ill and will die soon, unless the PCs return before then with the ingredients for a cure.
- The enemy warlord is mustering an army of monsters in the caves. Each day that passes swells their numbers, and makes exploring the caves and finding the villain more difficult.
- The pirate treasure is hidden in a sea cave that is only accessible for a few hours at the lowest tide of the month.
- The vampire lord is out hunting. Find and destroy his coffin before sunrise! If you fail, he'll pack up and move to a new hideout, and more innocents will die and rise as undead.
- The PCs have a hot tip to the location of a great treasure, but so does a rival party. Who will find it first?
- If you start a raid on the orc lair but leave to rest before you finish it, you might find that the monsters have bolstered their defenses when you return.
- A PC's mentor has been framed by a rival and sentenced to death. The party needs to solve the crime before the execution is carried out.
Some of these are inherently hard-and-fast deadlines - celestial conjunctions and tides happen when they happen. Others are "soft" deadlines that may be subject to modification based on player actions and NPC reactions. For instance, if the villain finds out the PCs are coming for him, he may quicken the pace of his evil plan or divert resources from it to deal with their interference, thereby pushing back the implementation of the plan. Or, if the PCs are aware that rivals are seeking the same treasure they are, they may try to sabotage or delay them. Strategic use of other resources may be used to "buy" more of the master resource of time, while inept management of other resources may squander it.
Players should generally be made aware at the outset that time is limited, and that the degree of success or failure may hinge on their management of it. Sometimes their characters may know exactly how much time they have. Other times, a more vague idea is appropriate and sufficient to keep them moving relentlessly forward. In some cases, you might tell them a range, and then determine the actual time randomly - maybe even after the adventure is over! For instance, inform them that, "The royal healer tells you the king will die in five to eight days without the elixir," keep track of how many days they spend in search of it, and then roll 1d4+4 when they return to see if they were in time. (Sometimes it's fun to be surprised right along with the players, and to remind everybody that the most heroic efficiency can still come to naught, or a faltering effort can be rewarded by fate. Making it an open roll avoids undermining the lesson that time matters with perceptions of DM fiat.)
The simplest way to implement a deadline, either hard or soft, is to track the time in days, which puts a cap on the number of times the party may rest and sharply limits their traveling back and forth. If the stars align for the ritual in three days, then the party can afford to rest three times while exploring the ruins where the evil shrine is located. If the king will die in a week, and the journey to and from the forest where the key ingredient for the cure grows is three days each way on foot, then the players have a single day to look for the ingredient - unless they devise some way to reduce travel time.
As important events draw near, you may want to use a finer time scale, in hours or turns. (Rounds really are probably most appropriate to scenes in which the PCs are present.) It isn't necessary to track small increments of time when an event is still far off, so don't feel obliged to keep track of the Brotherhood of Chaos turn by turn if it's still twelve hours until the stars align for the ritual sacrifice. At that point, things outside the immediate vicinity of the PC party can be tracked in broad strokes. An hour or two before, though, might be a different story. You might mark down turn by turn where the major forces of the cult will be and what they'll be doing, absent party interference. When the party does interfere, and the cult responds, you can move the NPCs and monsters on the map according to their movement rates, turn by turn just as the PCs move.
Making an adventure time-sensitive shouldn't add much to your prep time, though. Make a simple time line, and mark the important events on it. As the days pass, cross them off, and note any changes in future events caused by the PCs' actions. Draw loops and arrows to indicate events which are moved ahead or back on the time line, and write brief notes as to why and how the altered timing affects the event. If the villain marches his forces against the kingdom ahead of schedule to beat the PCs to the punch, for example, note what it costs him to do so - maybe his forces are not at optimum strength, or siege engines are left behind.
Here's a very simple time line showing the three days from the start of the adventure to the "deadline" and the twelve turns (two hours) immediately preceding the deadline. Note that a longer or more complex timeline may include "blank" days between those labeled with events, just like not every turn in the short-term time line has an associated event.
Protip: Draw or print up time tracker sheets before a session. Check boxes are easy to make by using the "q" key with the Wingdings font. Make a generic time sheet with several lines of boxes for ordinary dungeon exploration - it's easy to mark the end of light source and spell durations next to the appropriate turn box, and you'll know exactly when they expire as you check off turns. Print a few more blank time sheets to use when you need to track durations of effects in rounds during encounters. (Ordinarily I'd recommend hand-writing the events on a printed time sheet, but obviously I needed a fully electronic image for the blog, and I don't have a scanner, so I typed these in.)
All of this should not be construed to imply that a plain old-fashioned dungeon crawl with no external time constraints is a bad thing and should never be done. That sort of adventure can be great fun, as decades of players and DMs have confirmed, and in fact it might be just the thing for the next session after a tense "timed" adventure. It also doesn't mean that the time limits you impose should always be tight and make for razor-thin margins for error. You don't have to be brutal with time to make time an important factor in player decisions. Often a bit of wiggle room, which allows for greater discretion in decision-making, will make for more interesting choices and trade-offs, while still making clear that time is not unlimited.
One point which I think is paramount: When you do impose time constraints on an adventure, you should stick to them, and let the players experience the consequences, good or bad, or what's the point of it at all? By all means, modify the sequence of events logically in response to player actions, but once you make a decision, don't waffle, and don't fudge to try to set up a dramatic climax or to give the players unearned help. This is D&D, not Hollywood. The story doesn't have to end with rescuing the princess in the nick of time. It could end by sneaking her out of her cell on the first day of her captivity, by rescuing her from the guards marching her through the corridors toward the altar an hour before the ceremony, or by fleeing in terror and shame from the eldritch horror summoned by her death. For time to fulfill its role as the Master Resource, the players' management of it must produce real results, not illusionary ones.
For ideas on managing campaign events between adventures, see this old post.