Free tool · Workflow reliability

Success is not the same as done.

Your automation says it ran. The CRM row, booking or notification is missing. Compare the event log with downstream business receipts before deciding what happened.

No loginRuns in your browserEvidence, not a score
NEW HERE? START WITH THE FICTIONAL CASE
01 Load fictional sample02 Check outcomes03 Investigate, don't auto-approve

No customer data required. The sample is a demonstration, not a reported customer success. Start with the checker ↓ · Read the method ↓

01 / Investigate

Compare outcomes. Optionally check expected events.

Paste sanitized CSV rows only—no client names, emails, tokens, secrets or business-sensitive records. This tool does not upload or store your inputs. Use the fictitious sample to understand the findings first.

Browser-local analysis
Required columns: event_id,status. Blank or unrecognized statuses are flagged for investigation.
Required columns: event_id,receipt_id
One column only: event_id. Use pseudonymous test IDs from the same time window. This can reveal events missing before workflow execution; it cannot verify your source export is complete. Leave blank for the original two-file comparison.

02 / A practical guide

Why a green workflow run can still miss the business result

A workflow engine usually reports whether it executed the steps it was asked to run. A business needs something else: was the intended record actually created, was the correct message delivered, or was the booking finalized?

These are different kinds of evidence. For an agency handing over an n8n or API automation, the most useful acceptance check is often a simple reconciliation between the event that triggered work and the system that should show the outcome.

First: did the workflow actually run?

No execution record: check whether the external service called the intended Webhook URL, whether you used the temporary test URL instead of the production URL, and whether the workflow is published. The official n8n Webhook documentation explains the separate test and production endpoints. This reconciler cannot diagnose a missing trigger from two empty exports.

Execution exists, but no CRM/booking receipt: compare stable event IDs with the downstream export using the free tool above. A missing receipt is a reason to investigate, not proof that the write failed.

Receipt exists, but the customer-facing outcome is wrong: this tool can confirm a matching ID was exported, not whether the content, recipient, booking state or external delivery was correct. Check those facts within the authorized downstream system.

Before the workflow: did every source event arrive?

For a focused worked example, read our Shopify order missing from n8n: source-to-receipt guide, then return to compare the three fictional CSV files.

A clean result here does not prove all source events were received. With only execution and receipt CSVs, events that never triggered the workflow are invisible. Add the optional expected-source CSV above to compare a separately exported source event list with execution IDs; unmatched source IDs become investigation findings, not certified missed orders.

For Shopify and similar systems, obtain an authorized, complete source list for the same time window, pseudonymize the matching event IDs, and paste only the single-column event_id list above. The checker can then flag absent executions and compare existing executions to independent downstream receipts. For example, a store can have two orders while its workflow and CRM each contain only one matching record; this checker would show no discrepancy with only the first two CSVs, but can flag the unmatched second source ID when the optional source list is supplied.

Shopify recommends periodic API reconciliation for missed webhooks in its official delivery guidance. Check your event IDs, timing, permissions and provider-specific delivery rules. This site does not connect to your store, authenticate webhook delivery or certify full source coverage. Never paste customer orders or access tokens into these CSV fields.

A small, repeatable evidence check

  1. Pick one outcome, such as “one CRM record for each accepted lead.”
  2. Give every event a stable event_id, and preserve it in the downstream receipt where feasible.
  3. Export only the relevant test run IDs and business receipt IDs. Remove names and sensitive payloads.
  4. Compare the two sets; inspect missing, repeated and unexpected receipts.
  5. Confirm delayed processing windows and system-specific success rules before classifying an exception as a defect.

When you see a missing receipt, the next action is not automatically to retry the whole flow. First inspect idempotency safeguards and partial side effects, so a repair does not create duplicate business records.

Important: A matching receipt is not proof that the downstream transaction was correct. This tool is a narrow, read-only diagnostic aid. It does not replace end-to-end testing or establish payment delivery, regulatory compliance or recovered revenue.

What you can test next

For an actual client release, investigate one agreed missing-input case, a repeated event, a timeout, an application error hidden inside HTTP 200, and a partial-save failure. Record observed outputs separately from proposed tests and unavailable evidence.

03 / Self-service product research

Would recurring outcome checks be worth paying for?

The free checker above is available today for manual, sanitized CSV comparisons. We are studying a planned, not-yet-available standardized product for repeat imports, independent business receipts and exception review across systems. No live integrations, automatic monitoring, pricing plan or checkout is offered here.

If you manage operations or decide what your team buys, tell us how often you reconcile real outcomes, what free or built-in method you already use, what fails and whether there is a realistic software budget. Negative answers are useful too; no private records are needed.

Share an optional buyer-research answer ↗Opens a voluntary email draft with seven questions. Nothing is submitted automatically. This is research only, not a reservation, purchase or paid service engagement.