Within the Order Attribution data model, second order attribution fields (such as second order date, second order revenue, and related attribution dimensions) were incorrectly displaying data from the customer's most recent order instead of their actual second purchase. This affected repurchase rate calculations and customer segmentation analyses that relied on second order behavior. The issue was limited to the Order Attribution model — customer profile data was not affected.

We corrected the underlying logic to accurately pull each customer's second chronological order for attribution. The fix has been deployed and validated across both Snowflake and BigQuery environments. Merchants using the Order Attribution data model should review the following areas for expected shifts in data:

Transactional Sales Report, Product Affinity, Order & Order Line Revenue, LTV Time Series, Lifetime Value First Order SKU, Dynamic LTV, Customer Flags & Info

Customer profile data, first order fields, last order fields, and overall customer metrics. Merchants without the Order Attribution data model are not impacted.

January 20, 2026

Similar to Google Ads, merchants required product-level insights for Facebook campaigns that the standard "Ad Spends" report did not provide.

We introduced the "Product Performance Report" for Facebook Ads, allowing for product-id breakdowns within campaign mode.

January 20, 2026

Merchants needed more granular data to understand how specific products were performing within their Google Ads campaigns.

We added a new "Product Performance Report" to the Google Ads extractor, providing SKU-level breakdowns for better marketing analysis.

January 20, 2026

Users had to navigate through menus to reach their most important data after every login.

We added an option for users to set any Dashboard or Look as their custom homepage. This report will now appear as the first tab on the home screen automatically.

January 14, 2026

Downloads from dashboards were failing to include data that came from merged queries.

We updated the download engine to support merged queries and improved overall performance, resulting in significantly faster file generation.

January 7, 2026

Exceptionally large Sephora EDI 852 files were being read as empty strings, causing the parser to fail and data to be missed.

We updated the file handling logic to properly manage high-volume EDI files, ensuring they are read and parsed completely regardless of size.

January 7, 2026

The download component was failing to show real-time updates and was experiencing errors when generating specific formats like PDF and PNG.

We optimized the data stream to the UI for instant progress updates and resolved the formatting bugs for PDF, PNG, and CSV exports.

January 6, 2026

A validation error was preventing users from editing an existing TikTok Audience if they selected the "Create New Custom Audience" option during the edit process.

We corrected the validation logic to allow users to seamlessly switch to a new custom audience while editing their integration.

January 6, 2026

Data duplication and timing issues were occurring in Northbeam integrations due to how sync keys (like campaign names and status) were being handled.

We refined the sync keys to use specific IDs instead of names and moved Northbeam integrations to dedicated workflows. We also added an editable "wait timeout" to ensure the integration waits long enough for Northbeam's data to settle before extraction.

January 5, 2026

Users could not apply dashboard filters to their scheduled export jobs, limiting the usefulness of automated reports.

Scheduled jobs now support filters. The platform can read existing dashboard filters, allowing users to apply them when creating or updating their automated exports.