Two systems, one record, and a value that changes depending on which one updated last.
Every mapped field has a sync rule, and HubSpot documents four of them (HubSpot knowledge base):
| Rule | What it does | Use it when |
|---|---|---|
| Prefer Salesforce unless blank | HubSpot writes only when the Salesforce field is empty, and Salesforce values overwrite HubSpot | Salesforce owns the value, but HubSpot can fill a gap once |
| Always use Salesforce | HubSpot never writes to Salesforce, and Salesforce values overwrite HubSpot | Salesforce is the only source, such as owner or stage |
| Two-way | Whichever system updated most recently wins, and deletions sync both ways | Genuinely shared fields, which are rarer than they look |
| Don't sync | Nothing passes between the systems | Fields that should live in one place only |
Two-way is the setting people choose because it feels safe. It is the one that produces the coin flip: two systems take turns overwriting the same value, and the record shows whichever job ran last.
Marketing captures original source on the HubSpot contact. A rep then sets Lead Source in Salesforce from a picklist, and with a two-way mapping the rep's value replaces marketing's. The fix is to write original source once into its own Salesforce field, mapped as Prefer Salesforce unless blank and locked against edits, so HubSpot fills it once and nothing replaces it.
HubSpot deduplicates contacts by email. Salesforce allows duplicate leads unless duplicate rules are switched on, and the same person can exist as both a Lead and a Contact. When the connector cannot find a clean match it creates a record, and the duplicate count climbs on the Salesforce side.
The inclusion list decides which HubSpot contacts sync to Salesforce. Too wide, and every newsletter subscriber becomes a Salesforce lead nobody will work. Too narrow, and engaged contacts never reach a rep. Build it on your written lead definition rather than on whoever set it up two years ago.
HubSpot can create companies automatically from email domains. Letting those flow into Salesforce as new accounts produces accounts with no owner and no territory. Decide which system is allowed to create companies, and match on normalized domain before anything is created.
Records that fail validation rules, required fields or a picklist value that does not exist in Salesforce never sync, and the integration settings list them quietly. A weekly owner for that list catches most of the drift before it reaches a report.
Heavy two-way mappings also consume Salesforce API calls on every update. Cutting mappings nobody reads is the cheapest performance fix in the integration.
Usually because the field is mapped as two-way, so whichever system updated most recently wins. Change the rule to match the system that should own the value.
Write original source once into a dedicated Salesforce field mapped as Prefer Salesforce unless blank, and lock it against edits, so HubSpot fills it once and nothing overwrites it.
The rule that decides which HubSpot contacts sync to Salesforce. It should match your written lead definition, so reps see engaged people and not every subscriber.
Switch on Salesforce duplicate rules, decide when a person is a Lead or a Contact, match on email and domain, and clear the existing duplicates before tightening the sync.
In the Salesforce integration settings inside HubSpot, which list records that failed and why. Give that list an owner and review it weekly.
All of it free.
Leave an email and I will reply myself, usually the same day. No sequence, and no calendar link unless you ask for one.
Would rather just book time? Here is the calendar.
We use browser storage to run the site and to understand how it is used. You can accept all or necessary only. See our privacy notice.