Data Sheets

Persona Model Quick Guide

Issue link: https://resources.custora.com/i/1079925

Contents of this Issue

Navigation

Page 0 of 0

Q U I C K G U I D E S A L E S @ C U S T O R A . C O M Predictive Product Personas Overview Methodology By analyzing and comparing each customer's unique purchase history, we can predict which items they are likely to purchase in their lifetime. We call this bundle of items and their associated likelihood of being purchased a "probabilistic basket." Use Cases Consumer Insights Understand your customer base and discover what makes each persona unique by analyzing along additional attributes such as customer lifetime value, purchase frequency, and acquisition channel. Email Personalization Personalize new arrival, markdown, or welcome series emails by versioning content to align with each persona's preferences. Ensure coverage of the entire customer base by using all five personas. Targeted Acquisition Create lookalike audiences based on your most valuable or strategic personas for use in digital acquisition channels. Product Markdowns Curate the customer's onsite experience using content aligned to their persona preferences. population and "assigns" every customer to their best-matching persona group. Custora's persona segments are determined by applying item-level transactional data into the probabilistic model known as the Dirichlet Multinomial Mixture Model. Once each customer has an observed "basket," we identify 5 distinct probabilistic baskets that best explain the purchase distribution across the entire population. We call these "personas". For each customer, we find how closely their individual basket matches with each
 of our five personas and we generate the probability of them being in each persona. Each customer is assigned to the persona with the highest probabilistic match. Customers with a probability greater
 than 50% are assigned to "true" personas. Customers with a less than 50% likelihood for a persona are assigned to the persona for which they have the highest probability and classified as "leaning" towards that persona. 1. 2. 3. 4. For advanced detail on how our models work,
 we recommend https://journals.plos.org/plosone/article?id=10.1371/ journal.pone.0030126 Probability 50% True Leaning Custora defines a persona as a set of customers that tend to purchase the same types of items. Persona segmentation is most commonly used to curate content to customers based on their observed and predicted product preferences. Custora's personas are based on statistical patterns within transactional data, which differs from traditional retail personas based on demographic or psychographic attributes. For each retailer, Custora identifies the 5 most distinct persona types within a customer

Articles in this issue

Links on this page

view archives of Data Sheets - Persona Model Quick Guide