Why do we need a diagonal matrix?

Apart from simplifying matrix powers, why do want to diagonalize a matrix? Do they have any appealing application which can be used to motivate to study diagonal matrices. Thanks for any answers.

-
In addition to answers, see also some suggestions at math.stackexchange.com/q/1072459 – Alexander Konovalov Mar 22 '15 at 13:44

As an example, consider for instance that the analysis of the eigenvalues of a $3\times3$ orthogonal matrix leads immediately to the classical result that a rigid motion of the $3$-dimensional space that leaves a point fixed is a rotation around some axis.
One important application of this is if you have a vector differential equation $X'(t) = AX(t)+B$. Here, if you can switch to a basis that diagonalizes $A$, the equation decouples into independent differential equation for each coordinate, which are easily solvable. (This is even more important because higher-order ODEs in a single variable can be rewritten as a first-order vector equation and solved by the same process).
I cannot add anything here because, @Henning note almost all you need. But, I can't imagine what would we have if the identity matrix was lost suddenly. A world without $\text{id}$??? – Babak S. Dec 16 '12 at 16:32