Get a free consultation with a marketer marketolog@adbest.ru

Description of the analytical method "cluster analysis"

Cluster analysis is used to segment consumers based on their interests in different sections of a website, or based on their opinions and attitudes toward various products, company activities, and various promotions and events held by the company. Cluster analysis utilizes several different classification methods, and accordingly, cluster analysis is divided into several types: hierarchical, k-means, and two-stage. The search for the best solution is most often conducted by using several alternative methods and comparing the results, after which the researcher selects the method that yielded the best results. A limitation of this method is the need to use quantitative variables as input data.

Cost of cluster analysis

If the values of different variables differ significantly, standardization is used. If the number of original variables is too large, factor analysis is used to reduce dimensionality.

Cluster Analysis Training

Algorithm for applying hierarchical cluster analysis: 1.Launch and obtain dendrogram. 2.Determination of size with construction of frequency tables and number of clusters. 3.Creating a variable for the desired number of clusters 4.Visual display of the result with labels indicating the segments 5. Interpretation and verification

Segmentation by website behavior