Grouping similar rows together. k-means places k centres and assigns each row to the nearest one, repeating until stable.

ExampleSegmenting customers before anyone has defined the segments.