Tracking Plan Management
Best Practices
Groups

Intro to Group analytics

💡
Group analytics is a terminology used in product analytics tools, that refers to grouping by unique product concepts other than unique users.

The concept of grouping product analytics by unique users is well established. E.g. counting unique users who perform specific actions in your product, or individual user retention. In addition to unique users, most products have other core concepts to group by, such as unique workspaces, unique chat channels, unique games, etc.

Group analytics are supported by product analytics tools and CDPs like Segment, RudderStack, Amplitude and Mixpanel. You can also do group analytics in a fairly straight forward way in your data warehouse, as long as you provide the unique id of the respective concept in the relevant context.

In this article we dive into Managing Group Analytics in practice (functionality, consequences, use cases, how to manage), for each of the two main types of group analytics:

Grouping users

You can associate users with groups. Then that user profile, as well as every subsequent event associated with that user, will be attributed to the respective groups.

This is useful for example for B2B products, when each individual user of the product belongs to a company, and every action of that user should be attributed to that one company.

See further use cases and which analytics platforms support grouping users below in Associate user with group.

Grouping events

You can associate events with groups without associating the user with a group. This is useful if you have product concepts where each user can perform actions in different groups.

For example if you have a chat app and want to analyze the performance of different channels.

See further use cases and which analytics platforms support grouping events below in Associate event with group.

Managing Group Analytics

Associate user with group

Functionality: This action will associate this user with a specific group ID in your downstream data. This is supported by Avo, Segment, RudderStack, Mixpanel and Amplitude, and additionally you can use this method for your data warehouse via custom destinations in Avo.

Consequences: All events triggered with this user ID will be available for group analyses. See Amplitude (opens in a new tab), Mixpanel (opens in a new tab), Segment (opens in a new tab) and RudderStack (opens in a new tab) docs.

Use case: For example, if you have a "Workspace" group, you can associate a user with a specific workspace, according to your internal workspace IDs. This allows you to explore the workspace onboarding conversion rate, where different onboarding steps may be done by different users in the workspace.

How to manage in Avo: Add the "Update Group" action type to an Avo event. Then select which Groups to associate the user with.

Associate event with group

Functionality: Associating an event with a group will associate that specific event with that specific group ID. It will not permanently update the group information or permanently associate this user with this group ID. This is supported by Avo, Amplitude, Mixpanel, and additionally you can use this method for your data warehouse via custom destinations in Avo. This is not supported by Segment and RudderStack.

Consequences:

For Mixpanel and Amplitude: All events logged with a group will be available for group analyses such as calculating metrics based on unique groups. See docs for Mixpanel and Amplitude.

For your data warehouse: The events will have the Group ID associated with them. If you maintain a table with each group’s metadata (see Update Group Properties section), you can additionally enrich your events with the current state of the group metadata. This will allow you to query your events by the historical state of the group metadata, as it was at the time of the event.

This is not supported by Segment and RudderStack.

Use case: You should log events with a group when you have product concepts where users can be associated with many groups through their lifetime.

Example: If you have "Branches" in your product, of which users can create many through their lifetime, then you can create a "Branches" group to understand the lifetime of a branch, for example time from branch open to branch closed, etc. Then, instead of permanently associating the user profile with a specific branch, you associate only the branch related events to specific branch IDs.

Example: If you have chat channels, you can log the “Message Sent” event with the “Channel” group and the specific channel ID, so you can do analysis of channels such as retention by channel (as opposed to retention by individual users).

How to manage in Avo: Add the "Log Event" action type to an Avo event. Then add the Group you want to associate this event with.

Update Group Properties

Conceptually the group properties are quite similar to user properties. You can maintain group metadata, aka group properties, for the group IDs that you associate users or events with. The first step is defining your groups (group types and group ids). Then you can attach and update specific properties about the groups.

Functionality: This action will update metadata associated with a specific group (e.g. member count of a team). This is supported by Avo, Segment, RudderStack, Mixpanel and Amplitude, and additionally you can use this method for your data warehouse via custom destinations in Avo.

Consequences:

For Mixpanel and Amplitude: Future events associated with this specific group will have the updated metadata about the group available for filtering, segmentation, etc.

For Segment and RudderStack: Note that you can not update group properties without associating the current user with that group. This means that when you update group properties, it will automatically associate the user permanently with this group.

For your data warehouse: If you’re using a custom destination to send data into your data warehouse, you can use Update Group Properties to update a table that maintains the current state of group metadata.

Use cases:

Example: If you have a "Workspace" group, and that workspace just added a new member, or upgraded their subscription plan. You can then update the group metadata (aka group properties) with the new member count or subscription plan, to view workspace retention segmented by e.g. their current subscription plan.

Example: If you have a “Company” group type, then group id could be “Microsoft”, with the group properties ”Location”: “Redmond”, “CEO”: “Satya Nadella” and “Employee Count”: “150 000 - 200 000”. When something changes, you can update the corresponding group property. For example in 2014 the Microsoft’s “CEO” property was updated from “Steve Ballmer” to “Satya Nadella”.

How to manage in Avo: Add the "Update Group" action type to an Avo event. Then select which Groups to associate the user with.

Simplified groups terminology guide

This guide simplifies the understanding of group analytics across various platforms, focusing on Avo's "Group Types" and "Group Ids" and their equivalents in other tools.

Key concepts

  • Group Types: Categories for organizing groups, such as companies, branches, or chats.
  • Group Ids: Unique identifiers for individual groups within a type, like specific company names or unique codes.

In Avo, groups are sorted by their type and each one is uniquely identified by an ID. For example in the “Company” group type you’d have group ids as identifiers of the companies, such as “Alibaba”, “Moderna”, “Company A” or "s5WLukvxA8". This organization method is prevalent in many analytics platforms. However, platforms like Segment and RudderStack identify groups solely by their IDs, without categorizing them into types.

Leveraging group properties

Utilizing group properties is a powerful method for adding metadata to groups. This extra layer of information can enhance analytics by providing more context about each group, facilitating more detailed and targeted analysis.

Equivalents across platforms

PlatformEquivalent to Avo’s Group TypeEquivalent to Avo’s Group Id
AvoGroup TypeGroup Id
AmplitudeGroup TypeGroup Name
MixpanelGroup KeyGroup Id
Segment-Group Id
RudderStack-Group Id
💡
The approach to group analytics varies across platforms. Some focus on unique identifiers without categorizing them into types, while the addition of group properties allows for enriched metadata, enhancing the understanding and analysis of group data.