I will introduce substitutions on compact alphabets, which naturally extend the theory of symbolic substitutions on finite alphabets to allow for infinitely many letters. Just as in the finite case, substitution may be iterated to define a language of words and associated compact subshift, there is still a notion of primitivity, and primitivity still implies minimality of the subshift. However, unlike in the finite case, there are primitive examples with non-uniquely ergodic subshifts. This was first shown by Durand, Ormes and Petite, using constant length substitutions on Cantor alphabets, constructed by a beautiful but somewhat complicated procedure applied to certain Toeplitz subshifts. I will show that there are, in fact, very easily defined primitive but non-uniquely ergodic examples, even on highly discrete alphabets such as the one-point compactification of the naturals. I will also introduce the substitution operator, an analogue of the (transposed) substitution matrix from the finite case, and explain its role in determining whether or not the substitution generates a uniquely ergodic subshift.