Tempers added to pottery have the potential to obscure geographical variation in the chemistry of exploited clay resources when analysed using bulk methods. A pilot study was conducted to assess the possibility of using LA–ICP–MS to analyse clay matrix in pottery with extremely subtle partitions between INAA chemical groups. LA–ICP–MS was used to analyse the clay matrix of 32 pottery samples from seven Middle and Late Woodland period (c. ad 200–800) sites in north-east Florida and results were compared to data from INAA of the same samples. The results reveal many of the challenges in analysing highly heterogeneous materials with LA–ICP–MS. However, the study also generated data that are potentially useful in the determination of pottery provenance. In some respects, LA–ICP–MS of clay matrix replicated the chemical variation recorded by INAA, and in other respects the results are divergent. The similarities and differences between these results are explained by specific differences between the methods and the materials analysed (matrix versus bulk analysis). Suggestions for improving the method are discussed.