Cluster Analysis

Like factor analysis, cluster analysis is also a data reduction method. It, however, focuses on respondents as opposed to variables. It reduces a large pool of respondents into groups (or clusters) based on common themes. The members of each cluster group are more similar within a group than they are between groups.

 

Examples

Imagine that a health food store wants to understand the different types of customers that come into their store. They could use cluster analysis to group customers based on their socioeconomic status, frequency of purchase, type of products purchased, and monthly spend. Here are two examples of groups that might arise from this segmentation:

  • Highly-Driven Moms: They come from high-income households ($150K+) and have 1–2 children. They work long hours and want to get the absolute best produce for their children. They go to the store several times per week and have an above-average spend.
  • Quality Produce Seekers: They have an above-average income ($60K+), and have 0–1 child. They don’t have a lot of extra money to spend on food, but organic produce is especially important to them, so they visit the store once a week to buy organic fruits and vegetables for the week. They might pick up a few unplanned items, but only if the price is comparable to a competitor retailer.

 

Applications

Market researchers use cluster analysis to divide the general population of consumers into market segments with data from surveys and test panels. This helps them better understand the relationships between different groups of potential customers and aids with new product development and selecting test markets.

Cluster analysis is also used in biology, medicine, computer science, crime analysis, and mathematical chemistry.