In modern automated plant bakeries a large amount of data is collected on the operation of the plant. When this data is combined with product quality data such as loaf colour, appearance, consumer complaints, sales data etc. it has the potential to be used to improve processing efficiency, final product quality, and product marketability. However the huge volume of this data means it is often ignored as being too hard to analyse in any meaningful way. Data mining, which is a combination of techniques that produces information from large data sets, has the potential to be applied to this data to extract useful information. This paper describes our experience of applying data mining techniques to a plant bakery in New Zealand. The process involved setting up the systems required to extract data from the bakeries SCADA system, setting up sensors to automatically measure and record quality parameters, cleaning the data to remove faulted or anomalous results and then combining all the separate data blocks into one complete database for analysis. Data were analysed at two levels. Firstly, selected data were analysed for simple trends on an individual loaf basis which served to identify variability caused by divider pockets, tin positioning etc. Secondly, data mining techniques such as various classifiers and principal components were applied to the whole data set to find relationships between process data and product quality.