After base64 decoding we get (hex)
5d f8 be 87 82 2b ea 5e f3 5c 23 fe 37 81 0f bd which has a size of two blocks.
Of your small word-list, only marketing has so many letters that it needs two blocks:
m a r k e t i n as the first,
g 07 07 07 07 07 07 07 as the second (or another padding, but this is a common one), and so can correspond to this ciphertext.
A word like
foo fits in a block, eg:
f o o 05 05 05 05 05, and so would give an 8 byte result. Not that we do use that ECB mode is used (no extra IV data like for CBC, a block goes to a block, and we need padding, unlike CTR mode). So if the plaintext was chosen from among the word-list, marketing is the only candidate.