Consumer Behavior Analysis Models: Path, Funnel, Retention, and Lifetime Value

Most brands and merchants seek to leverage massive data to better understand consumers, enabling the creation of superior products and experiences. This process involves challenges such as ETL, data analysis, and data comprehension, requiring appropriate analytical models to integrate data from all channels, gain a comprehensive view of consumer behavior, uncover patterns, and derive refined insights.

Basic Concepts

What is behavior analysis? Consumer behavior refers to the decision processes, influences, and actions taken by consumers when purchasing goods. In e‑commerce, behaviors include exposure, click, product page view, shop page view, homepage view, search, add‑to‑cart, favorite, order, and payment. Consumer behavior analysis applies data‑analysis methods and visualization models to reveal underlying behavioral rules. Common methods are:

Behavior Path Analysis : Analyze dominant consumer paths, identify issues, and optimize the flow.

Funnel Analysis : Examine conversion performance between adjacent steps from start to final conversion.

Retention Analysis : Track how retention rates change over time after initial contact.

Lifetime Value (LTV) Analysis : Quantify per‑user revenue growth over time and assess value across segments.

Solution 1: Behavior Path

What is a user behavior path? It is a sequence of actions ordered by occurrence time. In e‑commerce, a buyer’s journey from login to payment includes browsing, searching, adding to cart, submitting order, and paying. By calculating real flows for each step across user segments, a comprehensive Sankey diagram can be generated.

Understand user behavior patterns and identify optimal paths.

Validate path rationality and pinpoint drop‑off nodes.

Detect abnormal paths and explore underlying motivations.

Compare behavior differences across sources or user attributes.

Identify typical user groups within paths.

Scenario Descriptions

Gain a holistic view of overall user paths; visualize upstream and downstream relationships.

Locate behavior steps that impact conversion and guide operational optimization.

Segment users (impulsive, rational, comparative, coupon‑seeking) for refined marketing.

Model Explanation

The behavior path model is typically visualized as a Sankey diagram composed of the 4W elements: WHO (user), WHAT (behavior type), WHERE (location), and WHEN (time). By filtering start/end points and behavior types, detailed flow maps can be examined to discover anomalies and high‑value nodes.

Case Study

In an e‑commerce scenario, a brand’s active users in June mainly placed orders after searching, indicating strong search intent but low conversion from order to payment, highlighting a key optimization point.

Solution 2: Funnel Model

Scenario Description

Funnel analysis examines conversion efficiency across sequential behavior steps, from an initial action to final purchase, measuring conversion rates between adjacent stages.

Application Scenarios

Measure overall funnel conversion rate, identify problematic steps, and improve total purchase conversion.

Compare funnels across segments to guide targeted operations.

Model Explanation

The funnel model input consists of two parts: the defined steps (e.g., Browse → Add‑to‑Cart → Purchase) and the user paths. Each user’s sequence is mapped to the funnel, and aggregated counts produce the overall funnel.

Example algorithm:

Browse → Purchase → Add‑to‑Cart => Funnel: Browse → Add‑to‑Cart

Add‑to‑Cart → Browse → Purchase => Funnel: Browse

Browse → Add‑to‑Cart → Purchase => Funnel: Browse → Add‑to‑Cart → Purchase

Add‑to‑Cart → Purchase (incomplete) => Funnel: none

Aggregated result: Browse (3 users) → Add‑to‑Cart (2 users) → Purchase (1 user).

Case Study

Comparing groups A and B with funnel steps “Exposure → Browse → Add‑to‑Cart → Purchase” shows group B has the highest overall conversion, but a sharp drop from Browse to Add‑to‑Cart, indicating a focus area for optimization.

Solution 3: Retention Model

Scenario Description

Retention analysis measures the percentage of users who, after an initial touchpoint, continue to perform a specified action over time. It is used to:

Estimate daily conversion numbers for a reached user segment.

Track conversion rate changes across channels and optimize marketing strategies.

Model Explanation

Retention calculates how many users who performed an initial behavior (e.g., exposure) later perform subsequent behaviors (e.g., browse, add‑to‑cart, purchase) within defined time windows.

Results are typically displayed in tabular form, showing retention percentages over successive periods.

Case Study

The retention chart shows that group A generally retains better than group B, with high retention in the first two days that gradually declines.

Solution 4: Lifetime Value (LTV) Model

Scenario Description

LTV measures the total value a user contributes over their lifecycle with the product. It helps operators assess the value of users from different channels (e.g., paid search vs. organic) and over various time horizons (90, 180, 360 days).

Model Explanation

The core LTV algorithm aggregates cumulative per‑user revenue from the first interaction onward, segmented by time range, interval, and dimension.

Input Table

Model Input

Description

Example

User Value Metric

Define the metric for user value

Average cumulative transaction amount

Time Range

Statistical period for calculation

180 days

Time Interval

Aggregation granularity (e.g., weekly)

Week

Dimensions & Segments

Drill‑down by dimension or segment

Dimension: ad channel; Segment: new users

Application Case

The LTV chart shows three different channel groups over 30 days; the value peaks around weeks 12‑13, indicating the optimal observation window.

About Us

The Alibaba Mom (Alimama) Strategic Data Solutions (SDS) team aims to make growth strategies more scientific and effective through data. We provide marketing insights, strategies, value quantification, and attribution services for all Alimama advertising clients.

Contact: [email protected]