Find and merge duplicates

The Duplicates page scans one object at a time and groups records that look like the same thing. Nothing is ever merged automatically — an incorrect merge is far more expensive to unpick than a duplicate is to combine — so every group is a suggestion you confirm.

What counts as a match

Three signals form a group: the same name, the same email address, or the same web address. Web addresses are compared by domain, so acme.com, www.acme.com, https://acme.com/about and ACME.com/ are all understood to be the same site rather than four different ones.

Why a shared LinkedIn or Twitter is not a match

Comparing by domain only makes sense for a domain the company actually owns. Everyone's LinkedIn page is on linkedin.com, everyone's listing is on a directory site, and plenty of small businesses share a website builder's domain — so treating those as "the same domain" would fold every unrelated company into one enormous group.

Solstral therefore recognises social networks, business directories, review sites, site builders and link shorteners, and compares the part of the address that identifies you: linkedin.com/company/acme and linkedin.com/company/globex are correctly seen as different, while the same profile written two different ways still matches. A bare twitter.com with no profile on the end carries no information at all, so it is ignored rather than matched.

Why colleagues are not duplicates

The same field can mean two different things depending on the object it sits on. A website on a company identifies that company. A company domain on a contact identifies their employer — and forty colleagues sharing it are forty different people.

Solstral works this out from your data rather than asking you to configure it. When a web address field's values are mostly unique, a repeat is treated as a duplicate. When the values repeat by design, sharing one is not enough on its own: the names have to agree too. So two records for Jane Doe at acme.com are still offered as a duplicate, while her forty colleagues are left alone.

If a single web address pulls in an unusually large crowd of records, it is treated as a shared platform we haven't catalogued rather than a duplicate set. The page tells you how many records were skipped for these reasons, so an empty result doesn't look like something is broken.

Merging

Pick the record to keep, then merge. Everything from the others moves onto the survivor — activities, emails, notes, files, tags and relationships — and the duplicates are removed. Where both records hold a value for the same field the survivor's wins; where the survivor's is empty, a value from one of the duplicates fills it, so merging never loses data you had.

Merges are written to the audit log with the records involved, so the change is traceable afterwards.