In Amplitude, the default level of reporting is the individual user. What this means is that, unless you specify otherwise, your Amplitude charts and analyses will all be based on data drawn from individual users.
Often, this is enough. But in some cases, Amplitude users need reports built around an aggregated unit of measurement—say, accounts, order IDs, or charts.
The Amplitude Accounts add-on enables you to do exactly this, by giving you analytical capabilities at the group level.
A group is an object that a set of users might belong to—such as a company of customers, a team of users, or a playlist with listeners. This will help you understand how specific accounts interact with your product, instead of just seeing the individual users in those companies.
This article will describe how to set up group properties, help you understand how to look up accounts and users, and show you how to set up this feature when using certain Amplitude integrations, like Segment or Salesforce.
To instrument account-level reporting in Amplitude, see our Help Center documentation.
Before you begin
- This feature is only available to customers on Plus, Growth, and Enterprise plans who have purchased the Accounts add-on.
- Changes to account groups and group properties will be applied to new data moving forward.
- Historical data cannot be retroactively updated to include account groups and group properties.
- You are able to instrument up to five different group types per project. If you try to instrument more, they will not appear in the UI. You can remove groups via Amplitude Data.
- Account-level properties have a limit of 1000 per project.
- See this Help Center article to instrument account-level reporting.
Group-level reporting: an overview
The Accounts add-on is built around the ability to instrument groups. Once you've set up groups, Amplitude will include them in a drop-down list in the Segmentation Module. It's there that you'll tell Amplitude to report at the group level instead of the individual user level.
An analysis using group-level reporting performs its counts by distinct user property groups. This allows you to analyze not only your product's individual users, but also the different groups your users are part of.
When might this be useful? Here are some potential use cases:
- Analyzing how many distinct accounts were active, or fired a certain event (group by account ID)
- Tracking how many charts are copied, saved, and modified in your organization, and then identifying which ones are interacted with most often (group by chart ID)
- Discovering how conversion rates change between different types of orders (group by order ID)
- Determining how many accounts have converted from free trials to paid accounts (group by either account ID or project ID)
- Tracking how many posts are drafted but never published in your social media platform (group by post ID)
These last two use cases are examples of funnels, so be sure you understand how those work before using group-level reporting in situations like these. Check out our Help Center articles on building and using funnels in Amplitude for more information.
Important Note:
- You can instrument up to five different group types per project.
One advantage of using group-level reporting in funnel analyses is that different members of the group can complete different steps in the funnel, and Amplitude will still interpret that as a conversion. In a standard funnel analysis, the same person must complete all steps of the funnel to count as a conversion.
This is useful for multi-sided marketplaces or B2B2C companies whose conversion processes involve multiple people. An example of this is a product that allows medical practices to bill patients for expenses. Their conversion funnel might include steps like “send invoice” and “send payment.” The former is completed by an admin, while the latter is completed by the patient. In situations like these, group-level reporting is the only way to accurately measure total invoice conversion.
You can use group-level reporting in most Amplitude chart types.
Event-level vs user-level group definitions
Groups in Amplitude can be defined at either the event level or the user level. What's the difference?
- An event-level group is one that only incorporates specific events in the users' overall journey. Users are assigned to a group at the time the event is sent, and users do not remain in that group unless you explicitly assign them to it when the event is sent. A good example use-case might be a multi-product company that only wants users' events associated with a particular group when those events are related to a specific product.
- By contrast, membership in a user-level group is independent of the events being triggered. This is useful when you want to attribute all events triggered by a user to a particular group. Users are assigned to the group once, and they remain in it for all future events. A good example use-case might be one where all users working for a specific company are associated with a user-level group.
NOTE: A user's group type cannot be un-set, and must instead be overwritten.
Add groups to charts
All chart and report types in Amplitude support group-level reporting, with the exception of the Personas and Compass reports. To use a group you have instrumented in your chart, select the group you want to analyze your data by in the Users dropdown, located in the Segmentation Module.
For example, if you wanted to track the number of daily active organizations and group them by region, you can set up an Event Segmentation chart like this:
Another example could be creating a User Composition chart, but this time, focusing on the composition of a user property for groups. For example, the following chart would show you the composition by country of all the organizations that use your product:
Create group-level behavioral cohorts
To create a group-level behavioral cohort, use the dropdown on the left to specify if you want to create a cohort out of one of your groups.
For example, we can create a behavioral cohort of companies who triggered the event Create customized report
, and apply that cohort to a Retention Analysis chart to see the differences in retention between companies that fired that event versus companies that did not.
You also have the option to import a cohort of groups from a file. The file must contain exactly one group name per line.
The following chart shows that organizations using the Slack integration have higher retention than organizations that do not, suggesting this might be important to focus on for organization retention.
View and download groups with Microscope
You can also use the Microscope feature with account-level reporting. This is useful if you are performing an account-level analysis and want to dig deeper into a single data point or bucket.
For example, imagine you have an instant messaging application, and you want to increase the number of new user invites. To do so, create a funnel analysis with steps from 'Activate Account' to 'Invite New Contact':
Then use Microscope to view the groups in the last step's drop-off, or download the groups to understand why those groups drop off before inviting additional co-workers. You can also create a group-level cohort of those accounts, apply that cohort to other charts, or open Investigate Conversion Drivers to perform a detailed analysis.
But before you can do any of this, you'll need to set up account-level reporting.
Explore the behavior of a specific account
Accounts allow you to drill into a single group’s behavior, similar to the User Activity section.
In this example, the group type is org id
, and the group value is 12345678
(listed under Group Name).
To access the Accounts tab:
- Click User Look-Up.
- Click the Account toggle.
- Click one or all of the groups you have instrumented.
- Optionally, search for a specific group or group property.
Group property searches must follow the syntax of name = value
. They are restricted to groups that have been active in the last six months. Amplitude will search across all historical values held by the property, and not just most recent group property value. Use quotes for multi-word strings. Delimiters like commas or semi-colons should not be used. Spaces are optional.
Clicking on an account will take you to that account’s page, where you can view the account’s properties and activities.
You will find the account’s most recent properties in the top panel. The account’s properties can be set three different ways: the Group Identify API, the Salesforce integration, and by using Event Segmentation to create dynamic properties. These properties can be used to describe the account as whole (e.g. '30 day active users', 'account manager', 'plan type', 'renewal date', etc.).
NOTE: If you're using the User Look-up to verify events and the properties set with them, they will not be visible in the Info view. You can find them by clicking Raw instead.
The charts above are automatically generated and depict the account’s activity so you can see how many users in that account were active this week, how well they are retained, and more.
Set properties at the group level
Important Note:
- The limit per project is 1000 account-level properties.
Group properties are simply properties at the account level. These properties will apply to all users who fall into that account.
Dynamic group properties
You can turn your KPIs into dynamically-updating group properties. Add group properties such as “Last 7 Day Active Users” or “Monthly Active Users” to each account in your product. Dynamic group properties are created by Admin-level users via an Event Segmentation chart with the following steps:
- Save a single time series Event Segmentation chart: Save a user-level chart with a single time series metric that you want to track.
NOTE: Dynamic properties are not supported on frequency and property distribution metrics, or on custom formulas.
- Determine time interval: Set the time interval you would like to update the group property on. For example, a rolling window of “last 7 day active users” will update every day, while “weekly active users” will update once every calendar week.
NOTE: Dynamic properties are not supported on static time ranges.
- Name group property: Choose your group type and name for the group property.
For example, to keep track of daily active users in your accounts, you can use the dynamic group properties feature to instrument a group property directly from the UI without writing any additional code. The chart shown below tracks daily active users; then, using the dynamic group property feature, you can create a group property that tracks the daily active users within each account.
NOTE: Editing the chart used to create the dynamic group property will not affect the property.
Create group properties via the Salesforce integration
To learn how to create group properties using Amplitude's Salesforce integration, see this article in the Amplitude Developer Center.