cases are selected by a sampling process in which the researcher identifies new cases that are similar to previous cases. When these cases no new insights, the process is repeated with newly selected cases that yield different insights, again until no new insights are noted.
Here are some suggested steps: