A consumption bundle is a set of goods that a consumer may choose to consume. Suppose the only goods available in the world are tea and coffee. Then a consumption bundle is any combination of cups of tea and coffee that the person could choose, and you can write
(tea, coffee)
For the bundle containing one cup of tea and one cup of coffee, the bundle would be written as
(1 tea, 1 coffee)
Now imagine that the items in the brackets can represent any goods whatsoever. Call them xi, where i is an index identifying the good in which you're interested. You can then rewrite the bundle as follows:
(x1, x2, . . . xn)
Here n denotes the number of all goods possible to consume.
Although this bundle of n goods is realistic (in that at some level, every good competes for your wallet with every other good), it's also cumbersome. Instead, for simplicity economists often use two goods: the one they're interested in and everything else, which you can think of as money. If you're interested in doing so, you can eventually generalize the simple model to all goods.
The two-good layout leaves the consumption bundle as being the following:
(x1, x2)
Here x1 is usually plotted on the horizontal axis of any graph or space, and x2 on the vertical axis.
Consumption bundles have to follow the normal rules of preference, which means that, for instance, with three bundles, A, B, and C, and where A is preferred to B, and B to C, A must be preferred to C.
Provided that your tastes satisfy the rules for well-behaved preferences (completeness, reflexivity, and transitivity), any consumption bundle can be associated with an indifference curve, and an indifference curve describes all the consumption bundles that yield an identical level of utility. If you were to take any bundles on the curve (call them P and Q for the moment), you'd be indifferent between them and can write the following:
P ~ Q