K means is used in different applications like document clustering, market segmentation etc., K means is usually performed to derive meaningful information or intuition of the structure of the data. Wide variations in the behavior of different subgroups is analyzed.
Clustering is done by increasing the relation between the data points in the clusters and decreasing the inter cluster similarity with respect to the nearest centroid. The number of clusters for a particular data set can be determined only by exploratory analysis and trying different options. It may also sometimes depend on the size of the data set. The number of clusters formed is also dependent on every data set. Without enough information about the data set it is highly unlikely to determine this part of the question.