Segment
Segment is SignalSmith’s module for building computed customer attributes, defining audience segments, and activating those audiences to downstream destinations. It brings together traits, audiences, audience syncs, templates, splits, and priority lists into a unified workspace for marketers and data teams.
Why Segment?
Traditional CDPs force you to move data out of your warehouse into a proprietary system before you can segment and activate it. Segment takes a different approach: all computation happens warehouse-native. Your data never leaves your warehouse — SignalSmith generates and executes SQL against it directly.
This means:
- No data copies — Traits and audiences are computed as SQL queries against your warehouse tables
- Warehouse-scale performance — Leverage the full power of Snowflake, BigQuery, or Databricks for audience computation
- Freshness you control — Schedule trait and audience evaluation on your terms, from minutes to daily
- Full SQL flexibility — When the visual builder isn’t enough, drop into SQL for any computation you need
How It Works
- Define your schema — Entity types (User, Account, etc.) with their attributes and relationships provide the foundation
- Compute traits — Build computed attributes like lifetime value, purchase frequency, or engagement scores using SQL, aggregations, or formulas
- Build audiences — Combine trait conditions and attribute filters with AND/OR logic to define customer segments
- Activate — Sync audiences to destinations like ad platforms, CRMs, email tools, and more
Key Features
Traits
Traits are computed attributes attached to entity types. SignalSmith supports three types:
| Trait Type | Description | Best For |
|---|---|---|
| SQL Traits | Custom SQL queries that compute any metric | Complex calculations, multi-table joins, warehouse-specific functions |
| Aggregation Traits | Visual builder for count, sum, average, min, max, count distinct | Common metrics without writing SQL |
| Formula Traits | Arithmetic and logical expressions combining existing traits | Derived metrics like ratios, scores, and flags |
Audiences
Audiences are segments of customers defined by conditions on traits, attributes, and entity relationships. The visual filter builder lets you construct conditions with AND/OR grouping, nested logic, and a full set of comparison operators.
Before activating, you can estimate audience size with sampling queries and preview sample members to validate your segment definition.
Audience Syncs
Audience syncs activate audiences to destinations. Unlike model-based syncs that map arbitrary columns, audience syncs manage membership lists — tracking which customers enter and exit the audience over time. Three sync modes (mirror, additive, subtractive) give you control over how membership changes propagate.
Templates, Splits, and Priority Lists
- Audience Templates — Pre-built segment definitions for common use cases like high-value customers, churn risk, and new users
- Splits — Divide an audience into random percentage-based groups for A/B testing and holdout experiments
- Priority Lists — Resolve audience overlap by defining which audiences take precedence when a customer qualifies for multiple segments
Getting Started
If you’re new to Segment, the recommended path is:
- Ensure your schema has at least one entity type with mapped attributes
- Create a few traits to compute key customer metrics
- Build your first audience using those traits
- Sync the audience to a destination
For a guided walkthrough, see the Audiences Quickstart.