Introduction
The Cohorts dashboard lets you track user behavior over time by grouping users by the date they installed your app. Use it to analyze retention, lifetime value, and campaign return on investment (ROI) so you can understand long-term user performance and make informed investment decisions. The dashboard uses the Install Cohort Activity data source.
To learn more about Analysis, read our Dashboards Overview article. To learn about all available data sources, read our Analysis Data Sources article.
Access
Cohorts appears in the Analysis menu only when it's turned on for your account.
When to use this data source
Use the Install Cohort Activity data source when you want to understand how groups of users behave after they install your app. This data source ties each user's later actions back to the date and source of their install. That makes it useful for questions about retention, payback periods, and long-term campaign quality.
Common use cases include:
- Retention analysis: See how many users from a specific install week still open your app over time.
- Lifetime value (LTV): Track cumulative revenue and purchase events per cohort to see how much value different acquisition sources generate over time.
- Campaign ROI: Compare cost against later revenue, return on ad spend (ROAS), and ROI to see which campaigns pay for themselves and how quickly.
- Ad partner benchmarking: Break out cohort performance by Ad Partner Name, Ad Set Name, or Creative Name to see which partners and creatives bring users with the strongest long-term engagement.
Unlike the standard Activity data source, this data source has a Users measure and day-range versions of most measures, such as Revenue A and ROAS B.
Caution
The Install Cohort Activity data source supports only install cohorts. Re-engagement cohorts aren't available in Analysis.
How cohort data is organized
Every row in the Install Cohort Activity data source is anchored to an install cohort, which is a group of users who installed your app during the same time period. Dimensions such as Ad Partner Name, Campaign Name, and OS describe that install. Measures such as Purchase A, Open A, and Revenue A reflect what those users did after installing.
In this way, Install Cohort Activity differs from the Activity data source, which reports on events independently of when the user was acquired.
Use Cohorts
To access Cohorts, go to Analysis > Cohorts in Branch. If the page opens in the legacy version of Branch, the new Cohorts dashboard isn't turned on for your account.
The dashboard measures cumulative user behavior over three day ranges after install, named D Range A, D Range B, and D Range C. It starts with D Range A set to 7 days, D Range B set to 21 days, and D Range C set to 14 days. To change a range, select its parameter at the top of the dashboard and enter the number of days.
To choose the currency for revenue and cost, use the Currency parameter. It's set to USD by default.
The Install Cohort Activity data source requires a rolling date window, such as "Last 30 Days." It supports only day-level granularity.
Example
Imagine you set the ranges to:
- D Range A = 7
- D Range B = 14
- D Range C = 21
In this case, Revenue A is the total revenue from the first 7 days after install, including the install day. Revenue B covers the first 14 days, and Revenue C covers the first 21 days.
Use Ivy for insights
You can use Ivy to ask questions about your cohort data. The Ivy button on a widget appears only when it's turned on for your account. To make sure Ivy uses the Install Cohort Activity data source, select the Ivy button on a Cohorts widget. Prompting from the widget gives Ivy the correct data source context automatically.
When you use Ivy with the Install Cohort Activity data source, try prompts like:
- "Show me 30-day purchase revenue by install cohort for the last 3 months."
- "Compare ROAS across my top 5 ad partners for users who installed this quarter."
- "What's the retention rate (opens) for iOS versus Android cohorts over the past 60 days?"
- "Which campaign had the highest install-to-purchase conversion rate last month?"
- "Break down cost versus revenue by ad set for my Meta campaigns."
Tip
Be specific about time periods in your prompts. Because cohort analysis compares user behavior across different install windows, saying "users who installed in the last 30 days" is more useful than "recent users."
Fields reference
Most measures in Install Cohort Activity come in A, B, and C versions, such as Install A, Install B, and Install C. The A version includes only events within the number of days set in D Range A, counted from install. Versions B and C work the same way with D Range B and D Range C. Revenue A, B, and C include only revenue from Purchase events.
In a new widget, all three day ranges start at 0, so the A, B, and C measures don't count any events until you set the ranges.
Cost Default and Users don't use a day range, so they have only one version. Users is the only measure that counts distinct users. All other measures count events. For example, Purchase A counts Purchase events, not users.
For the full list of fields, see Analysis Data Sources.
FAQ
How do I access the Cohorts dashboard?
Go to Analysis > Cohorts in Branch.
What do the A, B, and C suffixes on measure columns mean?
Measure columns such as Revenue A, Revenue B, and Revenue C correspond to the three configurable day ranges (D Range A, D Range B, and D Range C). Each value is cumulative. For example, if D Range A is set to 7, then Revenue A is the total revenue from the first 7 days after install, including the install day.
Does Cohorts use rolling or fixed date ranges?
The Install Cohort Activity data source requires a rolling date window, such as "Last 30 Days."
What kind of data granularity is supported?
The Install Cohort Activity data source supports only day-level granularity.
Why is Ivy returning unexpected results when I ask about cohort data?
Ivy might be using a different data source. To make sure Ivy uses the Install Cohort Activity data source, select the Ivy button on a Cohorts widget.