how to get user retention? which metrics could i use? | Mixpanel Community

Mixpanel Community Icon

APP

·

(edited)·

Mixpanel's Retention report is the primary tool for tracking user retention. Here's how to build one: [Retention Report]

1.
Define your Retention Behavior – Select two events: the starting event (e.g., "Sign Up Completed") and the return event (e.g., "Any Event" or a specific action like "Watch Video"). [Quick Start]
2.
Choose a Measurement type:

Retention Rate – Percentage of users retained within a window
Unique Users – Absolute number of retained users
Property Sum – Total sum of a property (e.g., total watch duration)
Property Average – Average property value per user

3.
Apply Filters – Narrow down by attributes like browser, platform, or geography.
4.
Add Breakdowns – Segment results by cohort, browser version, country, etc.

You can also choose between two retention modes: [Track User Retention]

N-Day Retention – % of users who return on a specific day
Unbounded Retention – % of users who return on a specific day or any time after

Metric	What it measures
Retention Rate	% of users who continue using your product over a period
Churn Rate	% of users who stop using your product (inverse of retention)
DAU/MAU Ratio	Stickiness — how often users return relative to monthly actives
D7 / D30 Retention	% of users still active 7 or 30 days after a starting event
Reactivation Rate	% of lapsed users who return to the product
Net Promoter Score (NPS)	Likelihood of users recommending your product (a proxy for loyalty)

[Top Retention Metrics] [Product Experimentation]

Tie retention to value, not just logins. Track a meaningful action (e.g., "Complete a ride", "Watch a video") that signals users are getting real value. [Retention Primer]
Choose the right frequency. A messaging app should measure daily retention; a healthcare app might focus on weekly or monthly. [Retain Your Users]
Look for "magic number" behaviors. For example, users who add ≥3 tools to a dashboard may be significantly more likely to retain long-term — use Mixpanel's correlation analysis to find yours. [Predict Retention]
Analyze by cohort. Compare retention across user groups (e.g., by signup month, acquisition channel, or geography) to find what drives stickiness. [Retention Primer]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Wellington L.
·
·
kapa.ai and how to get recorrent users?
kapa.ai
APP
·
(edited)·
Tracking Recurring/Recurrent Users in Mixpanel
There are a few approaches to identify and track recurring users:
1. Use the Retention Report (Recurring Mode)
In the Retention report, you can switch the retention type to "Recurring" (called compounded in the API). This tracks users who repeatedly return to perform an event, rather than just first-time returners. [Retention Query API]
2. Use the Frequency View
The Frequency report (available inside the Retention report) shows how often users return to perform an action. Toggle to Frequency mode and choose between: [Frequency View]
Cumulative Frequency – Users who did the event in at least X unique intervals
Non-cumulative Frequency – Users who did the event in exactly X unique intervals
This directly answers questions like "How many users came back 3 or more times in a week?"
3. Build a Cohort of Recurring Users
You can define a cohort for recurring users in the Cohort Builder: [Cohorts]
Example: "Users who performed 'Watch Video' at least once in two consecutive 7-day intervals"
Or: "Users who used the app on 5 out of the last 7 days"
Once saved, you can use this cohort as a filter or breakdown across all Mixpanel reports.
4. Lifecycle Analysis
You can also define specific recurring user types as cohorts: [Analyze User Engagement]
Retained users – Performed your value event at least once in two consecutive intervals
Resurrected users – Lapsed but came back
Power users – Performed your value event at a high frequency (e.g., 90th percentile)
ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.
Wellington L.
·
·
kapa.ai how i get engagement?

APP

·

(edited)·

Metric	What it measures
DAU/WAU/MAU	How many unique users perform a meaningful action daily/weekly/monthly
DAU/MAU Ratio (Stickiness)	How frequently users return relative to monthly actives
Feature Adoption Rate	% of users actively using specific features
Frequency	How often users return to perform an action

To monitor active users over time: [Get to Know Your Users]

1.
Open the Insights report
2.
Select a meaningful event (e.g., "Watch Video")
3.
Change the count from "Total Events" to DAU / WAU / MAU
4.
Customize your date range

Inside the Retention report, switch to Frequency view to see how often users return and perform actions within a time window (e.g., unique days in a week or month). [Frequency View]

You can build cohorts to classify users by engagement level: [Analyze User Engagement]

Power users – Perform the core value action at or above the expected natural frequency
Core users – Perform it consistently but less frequently
Casual users – Perform it rarely or sporadically
Dormant users – Have not performed it within the retention window

Focus on meaningful actions, not just logins or app opens. Track events that signal real value (e.g., "Complete a ride", "Send a message"). [Retention Primer]
Match frequency to your product type — a messaging app expects daily engagement, while a healthcare app may expect monthly. [Retention Primer]
Strategies to improve engagement include adding use cases, increasing feature breadth, increasing visit frequency, and increasing session depth/intensity.

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Wellington L.
·
·
kapa.ai how can i make a funnel that step one could go to step two or three, not necessary follow get a sequence.
kapa.ai
APP
·
(edited)·
Creating a Funnel Without a Fixed Sequence
You have two options depending on your exact use case:
Option 1: Any Order Funnel
If you want users to complete all steps but in any sequence, use the Any Order setting: [Ordering]
1.
Open your Funnel report
2.
Expand the "Advanced" menu in the Conversion Criteria
3.
Toggle on "Any Order"
With Any Order enabled, users can complete the funnel steps in any sequence and still be counted as converted. You can also anchor specific steps by clicking on the step number — anchored steps must occur at a specific position, while unanchored steps (marked with *) can happen in any order before the next anchored step.
Example use case: A job application form where users must fill in their name, email, and job title — but can do so in any order.

Option 2: Custom Events (Step 2 OR Step 3)
If you want Step 1 → (Step 2 OR Step 3) — meaning either event counts as the next step — use a Custom Event: [Custom Events]
1.
Open the Metrics menu and select "Create New" → "Custom Event"
2.
Combine Step 2 and Step 3 into a single custom event (e.g., "Step 2 or Step 3")
3.
Use this custom event as Step 2 in your funnel
This allows your funnel to look like:
Step 1: Event A
Step 2: Event B OR Event C (custom event)

Which Should You Use?
Scenario
Best Option
Users must complete all steps, any order
Any Order Funnel
Users can go to Step 2 OR Step 3 (not both required)
Custom Event
ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.
Wellington L.
·
·
kapa.ai how is the difference btw data view users and project users?

APP

·

(edited)·

Project Users are all members who have access to the entire project and its data. They are assigned a project-level role (Owner, Admin, Analyst, or Consumer) that determines what actions they can perform. [Project Roles]

They can see all events and properties in the project by default
Their permissions are governed by their role (e.g., Analysts can create reports, Consumers can only view)
Managed via Project Settings → Project Users

Data View Users are project members who have been granted access to a specific Data View, which acts as a filtered subset of the project data. [Data Views]

They only see the events and properties defined by the Data View filters
They still retain their project role permissions, but applied within the filtered data scope
Managed via Project Settings → Data Views → Add User

	Project Users	Data View Users
Data access	All project data	Only filtered data in the Data View
Scope	Entire project	Specific Data View
Default	"All Project Data" view	A restricted subset
Who manages	Project Owners & Admins	Project Owners & Admins
Plan required	Any plan	Growth plan or above

Every project has a default "All Project Data" Data View with no filters, which is equivalent to standard project access. [Data Views Overview]
A user can be a project member and be assigned to one or more Data Views simultaneously.
You can also add entire Teams to a Data View instead of individual users. [Manage Visibility]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Wellington L.
·
(edited)·
kapa.ai But I only can add a data view user if i added he/she as project user
kapa.ai
APP
·
(edited)·
Yes, that is correct. A user must first be a project member before they can be added to a Data View. As stated in the Data Views documentation, when you click + Add User inside a Data View:
"The modal shows only users that are part of your project."
[Data Views] So the workflow is always:
1.
First, invite the user to the project (via Project Settings → Project Users)
2.
Then, add them to the desired Data View (via Project Settings → Data Views → Add User)
Alternatively, if you manage users via Teams, you can add a team to a Data View, and all project members within that team will automatically get access to it. [Teams]
ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.
Wellington L.
·
·
kapa.ai if i have one project and into that project there are five properties where i need to give permission in a specific propertie. How can i do it?
kapa.ai
APP
·
(edited)·
Restricting Access to a Specific Property
You can use Data Classification to restrict access to a specific property within your project. This feature allows you to mark a property as "classified" so only certain users can view it.
Steps to Classify a Property
1.
Navigate to Lexicon in your project
2.
Go to the Event Properties or User Profile Properties tab
3.
Select the specific property you want to restrict
4.
Click Mark Classified
5.
Confirm the action in the pop-up [Data Classification]

Then Grant Viewing Access
For individual users:
1.
Go to Project Settings → Project Users
2.
Find the user in the Current Users table
3.
Check the Can View checkbox in the Classified Data column [Data Classification]
For teams:
1.
Go to Organization Settings → Users & Teams
2.
Select the team
3.
Enable the Classified Data checkbox for the relevant project [Teams]

Important Notes
This feature requires an Enterprise plan [Data Classification]
Users without Can View access will:
See the property is classified in Lexicon, but cannot use it as a filter or breakdown
See a warning when opening reports that contain the classified property
As of April 2026, classified data access is managed through Custom Roles under the "Classified Data Access" permission, rather than a separate per-user toggle [Custom Roles]
ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.
Wellington L.
·
·
kapa.ai What mean the filter 'is set' when I apply that in an event?
kapa.ai
APP
·
(edited)·
The "Is Set" Filter in Mixpanel
When you apply the "is set" filter to a property on an event, it means you are only including events where that property has a defined, non-null value. In other words, the event must have been tracked with that property present and containing a value — it excludes any events where the property is:
Not sent along with the event
Null (intentional absence of a value)
Undefined (property was not included in the event at all)
[Falsy Values]
Practical Example
Say you have a Page View event and you filter by Campaign ID is set:
✅ Included — Page View events where Campaign ID = "summer_promo" (property exists and has a value)
❌ Excluded — Page View events where Campaign ID was not sent (shows as (not set) in reports)
❌ Excluded — Page View events where Campaign ID = null

When Is It Useful?
To remove "(not set)" or null values from your analysis and focus only on events with complete data
To filter out organic traffic and focus only on events that came from a campaign (where UTM parameters are set)
To ensure your breakdown or aggregation is based on clean, complete data only
[Falsy Values]
ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.