Title: Cohorts Overview
Locale: en
URL: https://sensorswave.com/en/docs/user-operation/cohort-overview/
Description: Understand the core concepts, types, and use cases of Cohorts
Cohorts help you group and manage users based on specific conditions, creating meaningful user sets. With Cohorts, you can identify key groups such as high-value users and users at risk of churn, and apply cohorts to analysis, Feature Gates, and Experiments to enable targeted user operations.

## What Problems Do Cohorts Solve?

### From "All Users" to "Target Users"

In product analysis and operations, you often need to focus on specific user groups:

- During analysis: Only look at paying users' behaviors, excluding interference from test users
- During rollout: First open to internal employees, then gradually expand to production users
- During operations: Send re-engagement messages to users at risk of churning

Cohorts save "filter conditions" as reusable "user sets", allowing you to quickly locate target users at any time.

### From "One-time Analysis" to "Continuous Tracking"

In analysis scenarios, you may need the same set of users multiple times:

- Today: Analyze behavioral characteristics of high-value users
- Tomorrow: Compare differences between high-value and regular users
- Next week: Track the retention of high-value users

Cohorts standardize user definitions, ensuring consistency of analysis subjects across sessions and supporting long-term tracking.

### From "Manual Selection" to "Automatic Updates"

Dynamic Cohorts calculate users in real-time based on rules:

- The "Active users in the past 7 days" cohort updates automatically every day
- Users who newly meet the criteria are automatically added; those who no longer qualify are automatically removed
- No manual maintenance needed — always stays up to date

## Static Cohorts vs Dynamic Cohorts

Sensors Wave supports two types of cohorts:

### Static Cohorts

**Definition**: Saves a snapshot of users at a specific moment; the user list is fixed after creation.

**Characteristics**:

- The user list is determined at creation time and will not automatically update
- Suitable for scenarios requiring a fixed set of users
- Can be created via manual upload or user selection

**Use cases**:

- **Campaign participants**: Save users who participated in a campaign for subsequent performance analysis
- **Specific date users**: Save users who registered/paid on a specific date for comparative analysis
- **Externally imported users**: User lists imported from CRM or other external systems
- **One-time operations**: Send one-time notifications to a specific group of users

### Dynamic Cohorts

**Definition**: Saves user filter rules; the user list is computed in real-time based on the rules.

**Characteristics**:

- The user list changes dynamically based on rules
- Suitable for scenarios requiring continuous tracking
- Supports complex combined rules

**Use cases**:

- **User lifecycle**: New users (registered within 7 days), active users (logged in within the past 7 days)
- **User value tiers**: High-value users (cumulative spending > 10,000), low-activity users
- **Behavioral characteristics**: Search users, users with items in cart but unpaid, frequent visitors
- **Ongoing operations**: Re-engaging at-risk users, VIP-exclusive offers

### Comparison Summary

| Dimension | Static Cohort | Dynamic Cohort |
|-----------|--------------|----------------|
| **User list** | Fixed | Dynamically updated |
| **Creation method** | Filter and save / manual upload | Configure rules |
| **Update mechanism** | No automatic updates | Automatic updates |
| **Computation cost** | None (pre-computed) | Computed on a daily schedule |
| **Use cases** | Snapshot analysis, external imports | Continuous tracking, automated operations |

> **Our recommendation**: For most scenarios, we recommend using Dynamic Cohorts unless you specifically need a fixed user snapshot. Dynamic Cohorts automatically stay up to date, reducing maintenance costs.

## Core Capabilities of Cohorts

### Rule Engine

Cohorts support powerful rule configuration capabilities:

**Supported rule types**:

- **User properties**: Membership level = Gold Member
- **User behavior**: Spending in the past 30 days > 1000
- **Existing cohorts**: Belongs to the "VIP Users" cohort

**Supported rule combinations**:

- **AND**: All conditions must be met simultaneously
- **OR**: Any condition must be met
- **Nested combinations**: Complex logical combinations

See [Cohort Rules](cohort-rules.mdx) for details.

### Size Estimation

When creating a cohort, you can estimate the number of matching users:

1. Configure the cohort rules
2. Click the **Estimate Size** button
3. The system calculates and displays the estimated result
4. Adjust the rules based on the result until satisfied

### Cohort Applications

Created cohorts can be applied to various scenarios:

| Application Scenario | Usage |
|---------------------|-------|
| **Analytics models** | Use as a Segment By condition to analyze specific user groups |
| **Feature Gates** | Use as a targeting condition to enable features for specific users |
| **Experiments** | Use as an allocation condition to run experiments on specific users |
| **User export** | Export user lists for outreach through external systems |

See [Cohort Management](cohort-management.mdx) for details.

## Typical Use Cases

### Scenario 1: User Lifecycle Management

Create cohorts by user lifecycle stage:

| Cohort Name | Rule Definition |
|-------------|----------------|
| Newly registered users | Registration date within the past 7 days |
| Active users | Has login records in the past 7 days |
| Silent users | No login in the past 30 days, but has login in the past 90 days |
| Churned users | No login in the past 90 days |

**Applications**:

- Compare retention differences across lifecycle stages in Retention Analysis
- Enable onboarding features for new users
- Conduct re-engagement campaigns for churned users

### Scenario 2: User Value Segmentation

Create cohorts by user value:

| Cohort Name | Rule Definition |
|-------------|----------------|
| High-value users | Cumulative spending > 10,000 or spending in the past 90 days > 3,000 |
| Medium-value users | Cumulative spending between 1,000 and 10,000 |
| Low-value users | Has payment records, but cumulative spending  10 |
| Browse-oriented users | Products viewed in the past 30 days > 50, but search count  5 times in the past 7 days, but no payment |
| Frequent visitors | Active days in the past 7 days >= 5 |

**Applications**:

- Optimize search experience for search-oriented users
- Push promotional messages to cart hoarders
- Analyze usage paths of frequent visitors

### Scenario 4: Gradual Rollout Control

Create cohorts by user type for gradual rollouts:

| Cohort Name | Rule Definition |
|-------------|----------------|
| Internal employees | Email suffix = @company.com |
| Beta testers | Belongs to the "Beta Program" cohort |
| Android users | Device type = Android |
| iOS users | Device type = iOS |

**Applications**:

- Open new features to internal employees first
- Gradual rollout by platform
- Invite Beta users to try new versions

## Prerequisites

Before using Cohorts, you need to:

- Complete [SDK integration](../data-integration/client-sdks/javascript.mdx) and start receiving user data
- Properly implement [user identification](../data-integration/user-identification.mdx)
- Configure user properties in Data Center
- Have permissions to create and edit cohorts

## Next Steps

Now that you understand the basic concepts of Cohorts, you can:

1. **[Cohort Rules](cohort-rules.mdx)**: Learn more about configuring cohort rules
2. **[Cohort Management](cohort-management.mdx)**: Learn how to create, manage, and apply cohorts
3. **[Best Practices](best-practices.mdx)**: Reference cohort design and naming conventions

---

**Last updated**: January 19, 2026