A practical guided wave SHM system will be interrogated frequently and this will generate a large volume of data that must be processed. This paper discusses the use of sequential analysis to assess whether a change has occurred and whether it is of a form consistent with damage growth, rather than environmental changes. A methodology based on the generalised likelihood ratio (GLR) algorithm is introduced and applied to the guided wave monitoring of an 8 inch diameter pipe loop which is subjected to temperature cycling and the introduction of damage at different locations along the pipe. It is shown that the GLR algorithm can be used to process the signals automatically and to identify damage at specified probability of detection (POD) and probability of false alarm (PFA) rates. It is possible to detect damage corresponding to 0.25% cross section area loss in regions remote from features and to 0.5-0.75% loss at other locations. The methodology promises to be an important tool in the practical implementation of guided wave SHM.