Calculated filters

Calculated filters allow you to build very precise segments over an extended period of a user’s history. For example, use calculated filters to target users who purchased a particular product a certain number of times in the last 16 months, or who completed an eCommerce event that matches specific property values. Refine this audience by using event properties to make targeting even more granular.

How it works

Braze segments give you powerful targeting tools to create dynamic groups of users. For most use cases, this is enough to reach your audience effectively. Calculated filters are designed for advanced use cases where you need to analyze behaviors from up to two years ago or apply complex logic—without compromising data retention or system performance. Use User activity filters for purchase and eCommerce event criteria, or Data Object filters for account and custom object targeting.

For example, Braze default segmentation finds users that fit specific criteria you define, such as identifying a user who recently purchased one of your products. Calculated filters let you go deeper—like identifying users who bought a particular color of a specific product at least twice between 18 to 24 months ago. Calculated filters are an enhancement, not a requirement. If you need more advanced filters or a longer historical window, they’re a great tool to help while keeping your data usage optimized.

Calculated filters and SQL Segment Extensions

SQL Segment Extensions and calculated filters both help you build audiences from purchase behavior, but they use different tools and data sources. SQL Segment Extensions use SQL you write against your connected Snowflake data.

Behavior	Calculated filters	SQL Segment Extensions
How you define the audience	Choose purchases or eCommerce recommended events, and counts, time windows, and optional property filters	Write SQL against your Snowflake connection; use templates, incremental refresh, or full refresh
Where the logic runs	Criteria and refresh are managed in Braze as calculated filters	Query runs in your warehouse context according to your extension configuration
Filter list page	A shared list for user activity and data object filters, the Segments column shows how many segments use each filter, and processing statuses reflect generation state	Includes a Type column and filters that vary by extension type
Typical use cases	Purchase frequency and property-based rules over your selected window	Warehouse-backed logic, joins across tables, spend aggregations, and historical windows beyond the calculated filter form

When to use calculated filters

Use calculated filters when dashboard-guided user activity or data object rules are enough and you do not need arbitrary SQL across warehouse tables.

When to use other segment extension types

Use SQL Segment Extensions when you need full SQL, Snowflake-backed data, templates, or refresh modes designed for large or complex warehouse queries. Use CDI Segment Extensions when you need SQL that directly queries your data warehouse using data from Cloud Data Ingestion connections. Use Segment Extensions (simple extension) for message channel interaction criteria when you need to look back beyond recent interaction data.

Use calculated filters and Segment Extensions together

A segment can reference a calculated filter alongside a SQL or CDI Segment Extension—for example, a warehouse-defined cohort from an extension plus purchase rules you maintain in the calculated filter builder.

Create a calculated filter

To create a calculated filter, choose a filter type if prompted, define your criteria, then save and activate the filter before using it in a segment.

Step 1: Set up details

1. Go to Audience > Calculated Filters.
2. Select Create filter.
3. If your workspace has Accounts enabled, select a filter type:
   • User activity filters: Actions and behaviors for users.
   • Data Object filters: Attributes and relationships for data objects.
4. Enter a name that describes the audience you intend to target. A descriptive name makes the filter easier to find when you add it to a segment.
5. (Optional) Add tags to organize calculated filters in your workspace.

For User activity filters, select Enable recurring audience update to refresh the filter on a recurring schedule. If you don’t turn on this setting, the filter doesn’t refresh unless you update it or select Update audience. Data Object filters update hourly.

Step 2: Choose your criteria

• data object filters
• user activity filters

Step 3: Save and activate your filter

Select Save as draft to save a new calculated filter without activating it. For an activated filter, select Save changes to save your updates. You must select Activate filter before the filter is available in the segment builder.

After you activate a calculated filter, Braze starts calculating its audience. When processing is complete, you can select the filter when building an audience.

Use a calculated filter in a segment

After you create and activate a calculated filter, add it when building a segment or defining an audience for a campaign or Canvas.

1. In the segment builder, open the filter list.
2. Under Other Filters, select Existing calculated filter.
3. Select the calculated filter to include in the segment definition.

After you add the filter, select the icon next to the filter dropdown to view the filter’s details and confirm the criteria applied to your audience.

For more information on building segments, see Create a segment.

Manage calculated filters

Go to Audience > Calculated Filters to view, edit, and manage calculated filters in your workspace.

The Calculated Filters page lists user activity and data object filters together. You can narrow the list with the available controls, but the page does not include a filter-by-type control or a Type column. Use the Segments column to see how many segments use each calculated filter.

Status labels

Each calculated filter displays one of the following statuses. Processing and Processing failed show when membership generation is in progress or did not complete successfully.

Status	Description
Active	The filter is activated and available to use in segments.
Draft	The filter is saved but not activated.
Archived	The filter is archived.
Refresh disabled	Recurring audience updates are disabled. Braze may set this status automatically when a filter with scheduled refresh goes unused.
Processing	Braze is processing an update to the filter.
Processing failed	The most recent processing attempt did not complete successfully.

Edit and manage individual filters

Open a calculated filter’s row menu to take an action. The actions you see depend on the filter’s status.

For filters that aren’t archived, the row menu includes Edit, Messaging use, Archive, and Update audience. Update audience is available for active filters that aren’t processing. You can edit a calculated filter while it’s processing, but you can’t save your changes until processing is complete.

Archive

Before you archive a calculated filter, check whether it’s still in use. Archiving can silently break live campaigns, Canvases, or segments that reference it.

• Data Object filters: If an active campaign, Canvas, or segment still references the filter, archiving it means that campaign, Canvas, or segment will match no users the next time it launches or sends. Remove the reference first to avoid this.
• User activity filters: If a draft campaign or Canvas still references the filter, archiving it freezes that draft’s audience at the filter’s last refreshed results. It won’t update going forward unless you remove the reference first.

To find out where a filter is used, select Messaging use from its row menu.

You can’t archive a calculated filter that’s referenced by an active (not draft) campaign, Canvas, or segment. Remove those references before archiving. If you’re archiving multiple filters of different types at once, you see a separate confirmation for each type.

Unarchive

You can unarchive a filter in any of the following ways:

• Select Unarchive in the filter’s row menu.
• Select one or more archived filters, then select Unarchive.
• Open an archived calculated filter and select Unarchive on its page.

When you unarchive a filter, its status returns to what it was before you archived it:

• A draft returns to Draft.
• An activated filter returns to Active, counts toward the active filter limit, and Braze starts an audience refresh.

Wait until processing finishes before you unarchive a filter that shows Processing. If you’ve reached the active filter limit, archive an active filter before you unarchive another active filter.

Save versus activate

You can save a calculated filter without activating it. Inactive filters remain in your workspace but cannot be added to segments until you activate them. Select Activate filter to use the filter in segmentation.

Considerations

Calculated filters evaluate how many times a user completed an event within your selected window. They don’t sum values across events, such as total revenue or lifetime spend. For spend thresholds or other aggregations across purchases, use SQL Segment Extensions.

Property-based purchase and eCommerce rules

When purchase or eCommerce events include the properties you need—such as product category, store ID, or promo code—you can target users with event property filters. Examples include users who used a specific promo code during a date range, purchased a product category in the last 12 months, or shopped at a particular store. Lapsed-buyer patterns (for example, purchased in one window but not another) work by combining calculated filters in a segment.

Profile attributes and location-based targeting

• Profile attributes: Target loyalty points, tiers, or other values stored on the user profile with default segment filters, not calculated filters.
• Distance or proximity: Braze doesn’t calculate geographic distance in calculated filters. Precompute distances in your own service, store the result on the user profile, then segment on that attribute.

Message channel interactions

For message engagement criteria (such as email opens or push interactions), use Segment Extensions when you need to look back beyond recent interaction data. Calculated filters with message channel interaction criteria (when available in your workspace) evaluate recent message interaction data. They don’t use the same two-year data backfill as purchase and eCommerce criteria.

Frequently asked questions

Can I create a calculated filter that uses multiple custom events?

When using calculated filters, you can select one custom event, one purchase event, one eCommerce event, or one channel interaction. However, you can combine multiple calculated filters with an AND or OR when creating the segment.

You can add multiple events or reference multiple Snowflake tables when using SQL Segment Extensions.

Can I use arrays in calculated filters?

Yes. To use arrays, append brackets ([]) to your property name. If your property is location_code, you would enter location_code[].

Braze uses [] to traverse arrays and check if any item in the traversed array matches the event property. For example, you could create a calculated filter of users who match at least one value of an array property.

How does Braze calculate the time period for a relative time period of “last X days”?

When calculated filters calculate the relative time period (“last X days”), the start time is set to midnight UTC. For example, for a calculated filter that refreshes at 2024-09-16 21:00 UTC and specifies 10 days, the start time is set to 2024-09-06 00:00 UTC, not 2024-09-06 21:00 UTC. Calculated filters always use UTC for time windows; your workspace timezone does not apply.

However, you can specify time zones by using SQL Segment Extensions to identify users who performed an event 10 days ago based on midnight in company time, or users who performed the event 10 days ago based on the current time.