Local CSV tool

Analytics to CRM reconciliation

Explain why event rows and CRM records differ. Match four supplied snapshots by inquiry ID, keep delivery states separate and leave missing evidence open.

Compare four exports

Files are read in this tab’s memory. This tool makes no upload, AI or data API request and does not save files or results. Reset clears its inputs and results. With analytics consent, Compare sends only the fixed tool name through the site’s existing analytics.

inquiry_id, inquiry_date, accepted, environment
No file selected.

inquiry_id, event_name, event_date, environment
No file selected.

inquiry_id, original_inquiry_date, record_status
No file selected.

inquiry_id, status
No file selected.

Up to 2 MiB and 5,000 data rows per file. Use exactly the listed headers. IDs: 1–128 letters, digits, periods, underscores, colons or hyphens; begin with a letter or digit. Dates: YYYY-MM-DD. Environment: production or internal-test. Accepted: true/false. Event: accepted_inquiry. CRM status: created/unknown. Delivery: queued/rejected/created/unknown. Supply one final delivery status per inquiry. Extra columns are rejected so personal fields are not needed.

Prepare a matching snapshot

Use the same original inquiry cohort and extraction rule in all four files. CRM matching uses the original inquiry date. This comparison checks date equality; date-only files cannot validate time zones or UTC offsets. A date mismatch needs investigation, not an automatic correction.

The teaching ZIP defines October 1–7, 2026, America/Los_Angeles, extracted October 10. Its result is 120 event rows − 5 tests − 15 repeated events = 100 production inquiries; 95 created + 3 queued + 2 rejected = 100. Download the original synthetic source ZIP to inspect its README and exact CSV headers.

This is a supplied-snapshot comparison. It has no live GA4 or CRM connection and does not infer qualified leads, sales, revenue or acceptable loss. A created record establishes creation evidence only. Queued, rejected and unresolved records need an investigation owner and follow-up.

For the walkthrough and decision table, read the analytics reconciliation guide.

Put the result to work

Practical guides, examples, and worksheets for this task.