Bi‑Directional Sync Rules Between HubSpot and Salesforce: RevOps Best Practices

Two‑way sync sounds simple until conflicting edits, timing differences, and competing automations collide. This playbook defines pragmatic, low‑drama sync policies that protect your core business processes while giving marketers and sellers the context they need where they work. The key is to make decisions field‑by‑field instead of declaring the entire object one‑way or two‑way.

When to Choose One‑Way vs Two‑Way

Make one‑way the default unless you have a clear case for two‑way. Two‑way is appropriate when both systems legitimately create or improve a value that drives execution—for example, a phone number the prospect updates in a HubSpot form and a rep corrects in Salesforce. Treat stage and status fields as one‑way from Salesforce to HubSpot.

Conflict Resolution Policies

Choose a conflict policy per field and publish it: system‑of‑record‑wins, last‑write‑wins, or timestamp‑wins with a tie‑breaker. Avoid silent merges. Log conflicts so you can spot patterns that hint at unclear ownership or missing validation.

Sync Frequency and Triggers

Near‑real‑time sync is ideal for speed‑to‑lead and rep notifications; batch is acceptable for enrichment or low‑stakes attributes. Document trigger sources—form submits, imports, API writes, or workflow updates—and confirm they behave identically in your connector.

Deletions, Merges, and Recoverability

Deletions must be explicit and rarely automatic. Favor soft deletes or archival flags so you can reverse errors. For merges, pick a primary key and a deterministic rule set—then propagate merged IDs to the other system to avoid re‑introducing duplicates.

Testing Strategy for Sync Rules

Write scenario tests that exercise both human and automated changes. Validate not only field values but also process outcomes: Did the right owner receive the record? Did the status change subscribe the person to the correct sequence? Keep a suite of synthetic tests you can run after a metadata change.

Monitoring and Alerting

Dashboards should show success rates, error codes, retry counts, median delay, and backlog age. Alerts should target SLOs users care about (e.g., “median lead create > 2 minutes” or “status mismatch > 1% of synced records”) rather than raw connector errors that often self‑resolve.

Operational Playbook

Run a weekly triage for new errors, a monthly review of conflict logs, and a quarterly audit of sync directions versus real usage. Rotate ownership so tribal knowledge spreads across RevOps.

FAQ

What is the safest default sync direction?

One‑way from Salesforce to HubSpot for process fields (stage, status) and two‑way for low‑risk profile data with validation. Err on the side of protecting sales process integrity.

How do we handle rapid edits from two systems?

Prefer timestamp‑wins with a deterministic tie‑breaker. Add small jitter to automated updates to reduce collision windows, and avoid redundant automations in both systems.

Should we sync deletes?

Usually not. Prefer archival flags and periodic cleanups. If you must sync deletes, require approvals and implement a recycle bin strategy for recovery.

How do we test two‑way without disrupting sellers?

Use sandboxes and seed realistic data. Run synthetic tests after each change, then perform a limited production canary with close monitoring.

What causes most data thrash?

Overlapping automations, repurposed fields, and missing validation. Consolidate ownership and enforce contracts to reduce churn.