Here the population is divided into clusters or groups and then Random Sampling is done for each cluster. Cluster Sampling differs from Stratified Sampling. In the case of Stratified Sampling, the elements of each stratum are homogeneous (there are relatively minor variations within them.) As opposed to this, in Cluster Sampling the elements of each cluster are not homogeneous. Each cluster is representative of the population.
The retailer may use Cluster Sampling as follows:
She would divide the city of Mumbai into 4 - 5 zones say Zone I, Zone II, Zone III, Zone IV and Zone V. From the addresses on the bills she would classify the customers in the following clusters:
Zone I customers
Zone II customers
Zone III customers
Zone IV customers
Zone V customers
Then she would consider every item within randomly selected clusters. Note that there can be large variation within each cluster. For example, in the cluster of Zone I customers, there could be TV buyers, Stereo buyers and VCR buyers. As opposed to this, in Stratified Sampling there is relatively less variation within each stratum. For example, the stratum of TV buyers would include only those customers who have bought TVs.