The Internet of Things is connecting an ever-increasing number of devices. These devices often require access to personal information, but their meagre user interfaces usually do not permit traditional modes of authentication. On such devices, one-time pairing codes are often used instead. This pairing process can involve transcribing randomly generated alphanumeric codes, which can be frustrating, slow and error-prone. In this paper, we present an improved pairing method that uses sets of English words instead of random strings. The word method, although longer in terms of character length, allows users to pair devices more quickly, whilst still maintaining the complexity necessary for secure interactions.