How duplicates happen
Scopus and Web of Science both build author records by clustering documents algorithmically — matching names, affiliations, co-authors and subject areas. The clustering is good, not perfect, and certain situations reliably break it.
Name variants are the main cause: publishing sometimes with a middle initial and sometimes without, a hyphenated surname rendered inconsistently, or a transliteration that varies between journals. Changing institution splits records too, especially alongside a change in research area.
Common surnames create the opposite problem — one profile holding papers by several different people. That is worth fixing with more urgency, because you may be credited with work that is not yours.
Merging duplicates is one of very few actions that raise your visible citation count without a single new paper.
Why it matters more than it sounds
Every metric derived from a profile — total publications, citation count, h-index — is computed per profile. Split across two records, both look weaker, and the h-index in particular can be substantially understated because it depends on the whole distribution.
This matters wherever someone assesses you by looking you up rather than reading your CV: hiring panels, promotion committees, funders, and evidence-based immigration assessments. They see the profile, not the reality behind it.
It also matters for you. If you have ever reported a citation count from one database and been unable to reproduce it later, a fragmented profile is a likely explanation.
The process
Start by finding every profile that might be yours — search by surname and initial rather than full name, and check name variants and former institutions. Note each identifier.
For Scopus, use the Author Feedback Wizard, which walks you through selecting the profiles to merge and the documents to add or remove. Requests are reviewed by a person, and turnaround is typically weeks rather than days. For Web of Science, corrections go through your Clarivate author record.
Do both, since they are separate systems and fixing one does nothing for the other. Then connect the corrected profiles to ORCID so the merged state propagates and future publications land in the right place.
One caution: check the document list carefully before submitting. Requesting the removal of a paper that is genuinely yours, or the addition of one that is not, creates a different problem and is slower to unpick than to avoid.