Oh right! Everyone has a different definition of "idea." Do I mean big ideas or small ones or what? I actually mean any of the above. The definition has to be flexible for this exercise. So here is a more practical description of how I do this exercise. (Feel free to adapt it at will.)
First, you define what you're going for. What problem are you trying to solve? This method doesn't just have to be about writing after all. It can be ways to market a book, or ideas for your sister's birthday party, or 101 uses for a kumquat. I admit, I don't usually use this method for solving problems in a well developed story. A developed story is already full of ideas. A method like this is like pulling out a Howitzer to swat a fly. But sometimes at least starting this process will help a smaller problem.
For instance: I am 90 percent of the way to solving my MacGuffin problem, but if I were still stuck, here's how I would use the Magic of 100 to deal with this: Define the problem. I have to have a MacGuffin that affects three different characters in three different ways -- each not aware of the others. That's complicated.
So I might define three separate problems: What would Character A want? What items could Character B possess and not know the value (given that B is a smart cookie)? What sort of things would Character C keep silent about?
And with normal thinking I'd get myself lost in that: "A wants money, but B would know that's important so it can't be money, so maybe it's a secret, but how would B have that in his possession in a way that A could be seeking to get it from him without tipping his hand...?" and pretty soon I'd be lost in ideas.
With the brainstorming method you never do "it can't be that." You throw in everything and keep going. So I would start by just seeing how far I could go with just the first problem -- what does Character A want?
2. Suitcase of money.
3. bearer bonds.
4. bank passbook.
5. key to money
6. a map
7. a map out in the open that nobody knows what it means.
8. a map B is using as a coaster for his beer, and the rings from his glass throw off the markings, and that can be a complication later when nobody can read the map....
9. a map hidden in the piano
10. a key hidden in the piano.
11. a map etched in the metal of the tankard B is using to drink, and he thinks the coaster is the map when A tries to steal his drink.
12. a key hidden in the piano music as a code. (what kind of code?)
You'll notice that even though I was supposedly working on just the first problem, I was already getting into the second. Because I defined the questions up front, they are influencing what I do. And frankly, if I ran out of one line of thought, I could move on to others -- a deed, diamonds, his mother's diamond ring with sentimental value....
The other thing you might notice is that some of the items are objects, and some are situations. Some are even plot twists. This is another value of high numbers -- you have the freedom to go off on tangents. You could start listing ideas as to which characters are associated with the tankard or the piano, and that could give you more ideas because you suddenly realize that the thing A wants is the trust of the bartender! The bartender knows where the money (or key or map or suitcase) is and the bartender will not talk to A, so A wants to use B to get at that info.
Which leads me to another value of high numbers -- extra ideas. Let's say the trust of the bartender doesn't work for my story, but the idea is cool. I just got myself another story idea for free.
And that's really where this method shines -- new ideas. When your just starting a project, this really can open up your possibilities. You still have to define your problem up front, though. Say you want to write a flash story or a short story. You have a theme from a contest or just an idea, say from a previous brainstorming session - like the bartender idea. Or you've decided to play with new approaches to a cliche - say The Butler Did It.
Then you start listing whatever comes to mind.
1. the butler is a thief
2. the butler is a cop
3. the butler is an out-of-work longshoreman - kind of the opposite of My Man Godfrey.
4. Beautiful employer
5. what did the butler do?
6. employer in bad romance, butler breaks it up
7. butler seduces the employer
8. rival butler, each trying to undercut the other
9. servant pranks (switch salt with sugar)
10. rival butler murdered, our butler has to clear himself
11. our butler did do it after all
And when you get stuck on the ideas about the butlers trying to kill each other, you can go back to listing pranks butlers can play on each other, and then motives for pranks. You are free to go off on tangents as long as they seem interesting. (As a matter of fact you should go off on tangents - that's what this method is all about.)
I usually find that as I work through a list like this, my items get longer.
36. Okay, butler is longshoreman, and that's why he can lift the trunk with the body that his beautiful employer accidentally killed with a prank....
Or short cut ideas which refer back to an earlier idea:
74. like #27, only with turkeys instead of tigers.
Sometimes a story will take off before you get to 100, and that's great, although you can still use the brainstorming to come up with more details for the developing story, which may no longer involve a butler at all, but may involve turkeys, tigers or turtles which play pranks on beautiful employers.
The point of the 100 is to get outside of the box.
What if your style of ideas makes 100 too easy to get to? What if, say, you find yourself listing "turtles, turkeys, tigers, kumquats, ostriches" each as separate items? Wouldn't that be cheating?
Nope, you're not cheating. That's just how your mind works. Keep working that way. Just raise the number of your goal. You get to go for 200 or 500 or 1000. (And always keep checking back with your purpose, or at least with earlier list items to bring yourself back into focus if you find yourself just listing pointless words.)