Test a time-tracker integration before relying on it
Verify field direction, update behavior, failure handling and reconciliation.
Test the exact fields, direction and update behavior of an integration. A connection that authenticates successfully may still transfer the wrong record or fail to explain a partial import.
Define the contract on both sides
List source and destination systems, direction, trigger, identifiers, units, dates, review states and expected frequency. Decide whether records are created, updated or summarized. Map project and contributor identities explicitly. A connector logo or supported-app listing does not prove that the particular export fields and correction workflow you need are supported.
Use a fictional change sequence
Send one known 30-minute record, then correct it to 45 minutes, then change its project. Inspect whether the destination updates the original identity or creates additional rows. If the integration supports deletion or cancellation, understand its documented behavior before testing anything destructive. Do not use real production employee data as the first compatibility experiment.
Test recovery from an uncertain result
Record the source record ID, batch/request identity and destination result. Simulate a permitted test failure or inspect documented retry behavior. If the first request times out, determine whether the destination applied it before trying again. Do not assume a retry is safe or that an error response implies no records changed. Require a way to reconcile partial success.
Choose a manual handoff when it is safer
For a small volume, a reviewed export and explicit recipient acknowledgement may be easier to control than a poorly understood automation. Automation becomes useful when its transformation and failure behavior are observable, not merely when it saves a click. Hold an integration whose required identifiers, correction behavior or permission boundaries cannot be demonstrated, and keep the existing accepted process available.
Continue this work-record check
Use the linked guide for the next decision in this workflow. Keep your original records separate from experiments and record any unresolved requirement before changing a live process.
Sources and boundaries
Primary documentation supports the dated product facts. Proposed checks and fictional examples are our editorial method, not observed product results.
- DeskTime — export · checked September 23, 2026
- Time Doctor — export · checked September 23, 2026