How to fix data discrepancy issues in Windsor.ai

When streaming data from Windsor.ai to various destinations, users may notice discrepancies in the reported numbers. 

These mismatches most often arise due to the following reasons: 

• Caching
• Different query levels
• Missing data fields 
• Platform-specific calculations

This guide outlines the most common data discrepancy issues and provides troubleshooting steps to resolve them.

Possible reasons:

1. Windsor.ai caches data for a certain period (6 hours).
2. Data updates may not be immediately reflected in reports.

Troubleshooting steps:

1. Manually clear cache – Use the Windsor.ai onboard settings to refresh data.
2. Adjust refresh intervals – Ensure your data refresh settings match the required frequency.
3. Compare cached vs. live data – Run a direct API query to verify if cached data differs from real-time results.

Possible reasons:

1. Comparing data at different levels of granularity (e.g., campaign-level vs. ad-level in Google Ads).
2. Different filters, time zones, or aggregation methods are applied.

Troubleshooting steps:

1. Ensure consistent querying – Use the same filters, time periods, and aggregation levels in Windsor.ai and the original data source when comparing data.
2. Align data processing logic – Check that transformations and calculations match the expected logic.
3. Use direct API calls for validation – Fetch raw data from the source to verify if discrepancies exist at the query level.

Possible reasons:

1. Certain fields may not be available in the selected data source’s API.
2. API permission issues may restrict access to specific fields.
3. Filters applied in Windsor.ai may exclude necessary data.

Troubleshooting steps:

1. Verify API field availability – Check the data source API documentation to confirm if the missing field is retrievable.
2. Review API permissions – Ensure that your API key or account has access to all required fields.
3. Adjust query parameters – Modify Windsor.ai settings to include missing fields.
4. Compare raw API data – Fetch data directly from the source API to check if the issue is with Windsor.ai’s processing.

Possible reasons:

1. Shopify’s dashboard calculates sales differently from the API.
2. API reports might exclude canceled/refunded orders, while dashboards include them.
3. Timing differences in how orders are recorded and reported.

Troubleshooting steps:

1. Refer to Shopify’s documentation – Shopify explains its unique reporting logic differences here.
2. Reconcile order statuses – Compare completed, pending, and refunded orders separately.
3. Check API filters – Ensure the API query includes all necessary parameters, such as order status and time range.

By following these troubleshooting steps, users can resolve common data mismatches when streaming data to destinations via Windsor.ai.

Ensuring consistency in queries, understanding caching behavior, and reconciling platform-specific calculations are key to achieving accurate data reporting.

If the issue persists, contact our customer support for further assistance. 

Please, follow this quick guide to effectively report on your data discrepancy: https://windsor.ai/documentation/how-to-report-data-integrity-mismatch-issues. In this format, our dev team will be able to identify the problem much quicker.

Usually, investigating data discrepancies takes time (1-3 days), so it’s better to organize the report and send it to [email protected]’ll take a detailed look and reply to you as soon as we find a resolution.