Other granular analysis – descriptive

[jahid12]

Manual clustering
This technique is useful for grouping elements with similar characteristics, such as segmenting our URLs by type, redesigning GA4 channels, classifying users by purchasing behavior, grouping products by category, or classifying marketing campaigns by effectiveness.
Results : It allows grouping elements into categories canada phone number list or segments with similar characteristics, which facilitates analysis and decision-making by simplifying the total number of elements we observe in a few cases.
Dimensions and metrics : We use the metrics we want to analyze (traffic, sales, conversions, etc.) and group the data by the dimensions we want to use for clustering (purchasing behavior, product category, marketing campaign type, etc.). We use manual clustering techniques, such as pattern observation and creating classification rules (usually with simple rules (contains, starts with) or regular expressions).
Example : Segment our natural traffic by the type of landing pages they arrive from, classifying them into a few groups such as "Home," "Listings," "Product Sheets," etc. We can also segment users by purchasing behavior, creating groups such as "high-value users," "recurring users," or "inactive users."

Analysis of the distribution of metrics at a granular level.
Analysis of the contribution of individual elements to key metrics.
Analysis of data segmentation at a granular level.
Granular Comparative Analysis
16. Analysis of the deviation from the mean
This technique is useful for quickly identifying items that deviate significantly from the mean in a specific dimension. It allows you to detect outliers and understand the causes of these deviations.
Results : Helps to identify elements with exceptionally good or poor performance, and to understand the factors that contribute to these deviations.
Dimensions and metrics : We use the metrics we want to analyze (traffic, sales, conversions, etc.) and group the data by the dimension we want to analyze (web pages, products, marketing campaigns, etc.). We calculate the average of the metric for the dimension, and we identify items that deviate significantly from the average.
Example : Analyze the performance of different web pages in terms of conversion rate and identify pages that have a significantly higher or lower conversion rate than average. This can help identify pages that need to be optimized or pages that perform well and can serve as a model for other pages.