Re-keying is slow, error-prone and usually invisible until something goes wrong. Connecting two systems removes it, but only if you know exactly which fields move, which system is the source of truth, and what happens when data does not match. These four steps answer those questions before anything is connected.
Step 1 — Map what moves
Goal: Know exactly which information travels between the systems.
- List the records that are copied today, such as customers, orders, invoices or tickets.
- For each record, list the fields that are copied and who copies them.
- Note how often it happens and how long it takes.
- Collect a sample of recent errors caused by re-keying.
You finish with: A field-by-field map of what moves and a baseline.
Step 2 — Decide the source of truth
Goal: Avoid two systems disagreeing about the same fact.
- Choose which system owns each field when both can edit it.
- Decide what triggers a sync: a new record, a change, or a schedule.
- Define how mismatches and duplicates are handled.
- Confirm who may approve changes to the mapping.
You finish with: Ownership rules for every field and a mismatch policy.
Step 3 — Connect with narrow access
Goal: Move the data with the least access that works.
- Use each system’s supported connector or API before custom code.
- Give the connection access only to the records and fields it needs.
- Log every record it creates or changes.
- Run it on test or recent records alongside the manual process first.
You finish with: A working connection with logging, run in parallel.
Step 4 — Monitor and hand off
Goal: Make sure errors are caught and someone owns the connection.
- Send failed or mismatched records to a named person to resolve.
- Compare error rates and time spent against the baseline.
- Document the mapping, the owner and how to pause the sync.
- Review the mapping when either system changes.
You finish with: A monitored connection with an owner and a runbook.
Common pitfalls
- Syncing both ways without deciding which system wins.
- Giving the connection access to everything.
- No alert when a record fails to sync.
- Retiring the manual check before the connection has a record.