Importing a file where two people share a phone number — a household landline,
a shared office number — created only one contact. The second row matched the
first person on the number and updated them instead: the second person was
never created, their email address was added to the first person’s record as a
second address, and their phone numbers and note went onto the first person’s
timeline. The completion screen reported the row as updated, with nothing to
say a merge had happened. Re-importing the missing person did not bring them
back, because the import then matched the address that had just been grafted
on.
What you may notice
- Two rows with different names and different email addresses now import as two
contacts even when they share a phone number. - A contact is still matched on a phone number when the row names the same
person, and when the row names nobody at all. - An email address that is a contact’s second address now has to agree with
the name on the row before it matches. Their primary address still matches on
its own, so re-importing someone under a new surname still updates the
contact you expect. - A contact with no name yet — one created from an inbound call, say — still
takes the name from a row that matches it on a phone number, rather than
becoming a second contact.
What this doesn’t change
Contacts already merged by a previous import are not separated by this change;
they still have to be corrected by hand. The review screen does not yet preview
which rows will match an existing contact.