Back to Resources
Schema Mapping Guide
Standardize your data for SpendSignal's incrementality engine.
Overview
To accurately model incrementality, SpendSignal requires data to be mapped to a canonical schema. Whether you are using Snowflake, BigQuery, or PostgreSQL, your source tables should expose data in the following formats.
Revenue Events Schema
This table tracks conversions or purchase events. It is the source of truth for your "Y" variable (Output).
| Column Name | Type | Description | Required |
|---|---|---|---|
| transaction_id | STRING | Unique identifier for the conversion | Yes |
| transaction_date | DATE | Date of the conversion (YYYY-MM-DD) | Yes |
| amount | FLOAT | Value of the conversion | Yes |
| user_id | STRING | Hashed user ID for cross-device linking | Recommended |
| currency | STRING | ISO Currency Code (e.g. USD) | Optional (Default USD) |
Marketing Spend Schema
If you are importing custom channel data (e.g. TV, Direct Mail, or unsupported Ad Networks), use this schema.
| Column Name | Type | Description | Required |
|---|---|---|---|
| date | DATE | Date of spend (YYYY-MM-DD) | Yes |
| channel | STRING | Name of the channel (e.g. "TV_Spot_1") | Yes |
| spend | FLOAT | Cost incurred on that day | Yes |
| impressions | INTEGER | Number of impressions (Optional proxy) | Optional |