About half-way through the lifespan of Dragon and Dungeon magazines, a company called Paizo was commissioned to take over publication. After the Dungeon and Dragon and Polyhedron magazines were discontinued, Paizo obstinantly continued to publish adventures for D&D. And from magazine to bound book, they're all really good adventures. You don't realise how good they are until you try some RPG system that doesn't have published adventures at all, or not of the same quality, and then you really see it. And the funny thing about Paizo adventures is that they're all an illusion. They just happen to be very, very effective illusions that every Game Master, regardless of what system you run, can learn everything from.
There are a few different plot types in RPG, and you probably instinctively know them if you've played video games or tabletop games. We have terms for many of them, like "fetch quests" and "dungeon crawls" and "escort missions" and so on.
Within a Paizo adventure, there's usually a three or four act structure around those common plot points.
Do this two or three times, each emphasizing one of the traditional game plot types (fetch quest, dungeon crawl, and so on), and you have a complete story!
The nice thing about Paizo adventures is that they each fit together, and yet they're also completely independent of one another. It's a simple party trick. It's like the "exquisite corpse" game, where one person draws a head and another person draws a body and then they stick it together to see what wacky mismatch they've created. And it works brilliantly. As long as one adventure book starts with a brief mention of the previous book, you've got an adventure path! It works with anything. Take two phrases and string them together with "Now it's time to..." For example:
"The goblin horde has finally been vanquished! Now it's time to go to Mirewood Grove and fetch the magical rose, but beware the evil that guards it!"
Admittedly, the goblin horde didn't have anything to do with the Mirewood Grove. The Mulch Grove probably didn't exist when the goblin horde adventure was written, but nobody knows that, and so it works.
Paizo takes it a step farther than that. They plan this stuff out, presumably writing a six-point outline of the adventure path before actually writing it. That way you get little hints in early books about what's to come. For instance, maybe there have been murmurings about an evil lich in the region, and by book 3 it turns out that stealing the lich's rose in book 2 has triggered severe retribution. That's an optional touch that isn't necessary. Even one of the most beloved adventure paths, Rise of the Runelords doesn't actually mention (to the players, at least) the big evil baddy until the third adventure. (At least, so far as I recall. I have a hard copy only, so I can't do a Ctrl+F on the text to be sure. I could be missing a mention, but it can definitely be played that way even if the text does mention him earlier than I think it does.)
The links between adventures makes it feel to the players that they're playing an extended, sprawling, epic campaign where one plot point leads seamlessly to the next. But the tenuousness of the links between one adventure and another has the opposite benefit, too. You can pick up any adventure, regardless of where it falls within the larger campaign, and just play it, with no knowledge of what came before or what's meant to come next.
I've cobbled together a level 1 to 20 campaign using a mix of Paizo and Frog God modules, and none of them were meant to connect to one another. They were the hard copies I had on my bookshelf, and so they were what I used. Never once did a player stop the game to ask what the dragon over here had to do with the giants over there, or what any of them had to do with the evil wizards over in the west. It's all connected because evil begets evil and the friend of your enemy is your friend. It literally doesn't have to be any more complicated than that.
I've said that Paizo adventures are illusory. In the sense that the events within each adventure doesn't actually matter, and yet they still connect magically to create an epic story, they are illusory. But they're also illusory because, as every Game Master knows, no story element is stable when exposed to players. An NPC asks player characters to go investigate a suspicious cave-in, and the players immediately arrest the NPC for being the obvious perpetrator trying to throw them off track. As a result, you end up running a courtroom drama in which the players prove beyond the shadow of a doubt that this erstwhile NPC is guilty. The point is, many Game Masters, deep down (or not so deep down), don't really need somebody else writing plot points for them. In a pinch (or by default, for some) it's easy to construct a story in an RPG. You present players with some condition they find displeasing, and then you follow their lead as they attempt to change that condition to something that pleases them. Simple.
Then again, there are at least 5 reasons to run a published adventure, and at least one important principle to learn from Paizo's adventures. Paizo adventures demonstrate that, maybe counter-intuitively, adventures aren't flowcharts. If anything, they're dungeon crawls. I'll explain the difference.
A flowchart is a chain of dependencies. To get from point A to point B, you must have satisfied some condition. At first glance, that seems like a logical format for an adventure. You don't really want your players barging into the monster's secret hideout before they go to the massacre site and find the monster's tracks leading to the hideout. Except, maybe you do.
In a Paizo adventure, players absolutely can go to the secret hideout before finding the tracks leading to the secret hideout. You can pick up the book set in the hideout without ever even knowing there was a massacre. Some NPC at the start of the book will tell you "Now that the massacre has been solved [by some other group of adventurers before you], you can finally go to the monster's secret hideout by following the tracks [some other group of adventurers found] and slay the beast." Your adventures need to be that robust, too, and conveniently that means your adventures can be simpler than you might think.
I've done it several times myself. I've designed adventures that expect players to go to location A so they find the clue pointing to location B, which leads them to the secret hideout. Bad idea. It's not that it's "on rails" (although, it definitely is on rails), it's that lacks breadth. It forces players to have tunnel vision. They can't look over there because they haven't found the clue that unlocks that area yet. And that means they can only look here, which may or may not be very exciting, but either way it breaks the explicit promise of an RPG that you can look anywhere, go everywhere, and do anything.
If you're not going to do an open world adventure where players create the plot themselves, which definitely isn't for everyone, then take notes from Paizo. Design adventures where a friendly NPC or some catastrophic event points them to the plot. When they go to the location where the plot is happening, there's a displeasing condition there. An heirloom's been stolen, somebody's been murdered, the land is infested, a house is haunted, a king has been deposed, whatever. It's a condition that is obviously not right. Let the players solve it.
And now that that problem has been solved, the players can finally tackle the next problem.
Keep doing that until you reach level 20. Then pass the Game Master hat to a different person, roll up new characters, and start again.