Usually you will need to define a separate .xsl file for each bit that you want to extract from the XML result. It will help to use the feature of the HTTP binding that caches the page returned, so that each item does not require its own HTTP GET over the wire.
Thanks for the reply and clarification @watou that what I suspected. Instead I decided to use the HTTP binding with the JSONPATH transform to extract the data in a rule. At least that way I have less files to open!