Why your citation count differs between databases

Three databases, three numbers, one paper. The disagreement is not an error — they are counting genuinely different things.

They index different literature

Google Scholar crawls the open web for anything resembling scholarly output: journal articles, conference papers, preprints, theses, technical reports, working papers and book chapters. Its coverage is the broadest of the three and the least curated.

Scopus and Web of Science are curated. Each has selection criteria, an editorial board deciding what enters, and a review process that can discontinue titles. A citation only counts if it appears in a source the database covers.

So a paper cited in a preprint, a thesis and a technical report has three citations in Scholar and possibly zero in Web of Science. Neither number is wrong. They answer different questions — "who has referenced this anywhere" versus "who has referenced it in curated literature".

Google Scholar usually returns the highest count because it indexes the most, including preprints and theses. Curated databases count less on purpose.

The other sources of divergence

Coverage windows. Web of Science coverage of a given journal may begin in a particular year, and citations in earlier volumes simply are not counted. This affects older work disproportionately.

Document matching. Databases sometimes fail to link a citation to the cited paper — a mangled reference, a missing DOI, a variant title. Those citations exist but attach to nothing.

Profile fragmentation. If your own author record is split across two identifiers, an author-level total will undercount even when each individual paper is counted correctly. This is the most fixable cause and worth ruling out first.

Which to quote

Use whichever your field treats as authoritative — commonly Scopus or Web of Science in the sciences and engineering, with Scholar as broader supporting context. Some disciplines, particularly parts of the humanities and computer science, are better represented in Scholar.

Whatever you choose, name it and date it: "42 citations (Scopus, retrieved 12 August 2026)". A bare number invites the question of where it came from, and an unattributed figure that cannot be reproduced does more harm than a smaller attributed one.

Do not mix sources to assemble a flattering picture — quoting a Scholar citation count next to a Scopus h-index is not comparing like with like, and a reader who checks will notice.

Frequently asked questions

Usually Google Scholar, because it indexes the widest range of sources including preprints, theses and reports. That breadth is also why it is noisier — it may count duplicates or non-peer-reviewed citations.

Not less credible, less curated. For some fields it gives the most complete picture. For formal assessment, curated databases are usually preferred because their coverage criteria are explicit and stable.

Either the citing source is outside that database's coverage, or the citation failed to match your paper — commonly a mangled reference or a missing DOI. The first is expected; the second can sometimes be reported and corrected.

Report the one appropriate to your field, and name the source and date. Quoting the highest available figure without attribution reads as cherry-picking to anyone who checks, and costs more credibility than the extra citations gain.

How we can help

Author positioning

Building a coherent, findable academic record across the databases that count.

Citation growth

Growing genuine reach and citation of work you have already published.

Research visibility

Getting your work in front of the readers most likely to build on it.

Indexing support

Publishing where your discipline actually looks, so the record is counted.

Is your publication record showing up as one person?

Book a free 30-minute consultation. We will look at how your work appears across Scopus, Web of Science and ORCID — fragmented profiles quietly understate a record that is already there.

Book a meeting →Contact us