What is Retention analysis?

Retention Overview

Retention in Leanbase shows how many users return to your product over time. It helps answer critical questions like:

• Are new signups coming back after their first experience?

• Did a new feature or UX change improve user return rates?

• Which types of users or accounts retain best over time?

Retention insights are essential for understanding user engagement patterns and measuring product–market fit.

Creating a Retention Insight

Retention insights in Leanbase work with both events and actions.

When you create a retention insight, you define two things:

1. Start event – The event or action that includes a user or group in a cohort.

2. Return event – The event or action that determines whether that user or group was retained.

Often, both are the same event (e.g., “Product Interaction”).

Example:
You might use an action like Product Interactions for both start and return events. You can group by Unique Users and analyze users who first performed this action during a seven-week period.

Advanced Retention Configuration

With Leanbase, you can also:

• Analyze group retention – Measure how organizations, teams, or accounts retain users. Ideal for B2B products where not all company users are active.

• Define retention periods – Choose any interval (hours, days, weeks, or months). For example, “in the last 12 months” to assess year-long retention.

• Apply filters – Use event properties (device, country), person properties (job title, company), feature flags, or existing cohorts.

• Select retention type –

  • First-time retention: users’ first-ever occurrence of the start event.

  • Recurring retention: users included each time they perform the start event within the defined period.

• Add breakdowns – View retention by property (e.g., plan type, geography, browser).

• Configure advanced metrics – Like relative retention, comparing cohort changes between time periods.

Understanding Retention Insights

Leanbase visualizes retention data in two ways:

1. Retention Graph – A line or bar chart comparing cohorts over time.

2. Cohort Retention Table – Tabular format showing exact percentages and cohort sizes.

By default, all values are shown as a percentage of the initial cohort (“Week 0”).

Table Columns

• Column A – Cohort name or time range (e.g., May 20–May 26).

• Column B – Size of each cohort (unique users or groups).

• Column C – “Week 0,” the baseline where all cohort members performed the start event (always 100%).

• Column D+ – Subsequent periods (Week 1, Week 2, etc.), showing what percentage of users returned.

  • Example: 27.4% of the May 20–May 26 cohort returned in Week 2.

Tips for Analyzing Retention

• A user belongs to one cohort only (based on their first start event).

• However, a user can appear in multiple retention periods within the same cohort (e.g., both Week 1 and Week 3).

• It’s possible for a later period to have higher retention if users return cyclically (e.g., weekly usage).

• Retention insights track if a user returned—not how often. To analyze frequency, use Stickiness insights.

• Ongoing periods display a tooltip (“in progress”) and will update as data refreshes.

• Click a retention cell to open a user details modal for deeper analysis of that segment.

Retention Calculation Options

Reference Types:

• Period 0 (default) – Percentages relative to the initial cohort size.

• Previous period – Percentages relative to the immediately preceding period.

Mean Calculation:

• Weighted mean (default) – Accounts for cohort size when averaging.

• Simple mean – Treats all cohorts equally regardless of size.

Relative Retention

Two calculation modes:

1. Starting cohort size (default):
   % of users who returned compared to the initial cohort.
   → Example: Day 3 = 25% means 25% of the original users returned that day.

2. Previous period:
   % of users who returned compared to the last period.
   → Example: Day 3 = 40% means 40% of those who returned on Day 2 also returned on Day 3.

This helps identify where retention improves or declines most sharply.

Returning Time Criteria

Two options determine which users are counted as “returned”:

• On (default): Counts only users who returned during that specific time interval.

• On or after (cumulative): Counts users who returned on or after that time (e.g., Day 3 also includes Day 4+).

  • This represents cumulative retention, effectively showing the inverse of churn.

Mean Calculation Logic

• Simple: Equal weight for each cohort.

• Weighted: Larger cohorts contribute more to the overall average.

Breakdowns

Breakdowns allow you to segment retention by:

• Event properties – e.g., browser, device

• Person properties – e.g., country, plan tier

• Group properties – e.g., organization, team

Each breakdown shows average retention for that property value.
Click a breakdown row to expand and view cohort-level data.

Note: Both start and return events must share the same breakdown value.
Example: If a user starts in Chrome but returns in Firefox, the Chrome breakdown won’t count that return.

Displaying Retention Insights

When saving a retention insight, choose:

• Chart Type: Line or bar

• Display Type: Show as chart only, table only, or both on dashboards

Retention vs. Stickiness

In short:

• Retention = Are users coming back?

• Stickiness = How frequently are they engaging?