At a glance
A step-by-step guide to creating a Google Scholar profile: institutional email verification, fixing misattributed papers, merging duplicates, and reading the metrics with appropriate suspicion.
Dr. M, a materials chemist two years into a lectureship, lost an argument with a search engine. A hiring committee had opened her Google Scholar page and found an h-index lower than a rival candidate's. Her record wasn't the problem. A namesake in another country had been quietly merged into her profile years earlier, and roughly forty of her own papers sat unattributed under a hyphenated name variant she stopped using in 2019. Cleaning it up took an afternoon.
That afternoon is worth booking. In a survey of faculty at ten highly ranked universities, published in Scientific Reports in February 2025, more than 60% of respondents who said they weigh citation counts reported pulling them from Google Scholar — more than every other source combined — and 73% believed their colleagues do the same. Whatever you think of that habit, it's the habit.
The profile itself is free and takes about fifteen minutes to start. Getting it right takes longer, because Google's matching is statistical rather than authoritative, and nobody audits the result on your behalf.
What the profile is, and the two conditions that make it visible
Google Scholar Citations launched in 2011 and opened to all authors that November. The product is a public author page: your papers, a citation count beside each one, a graph of citations over time, and three numbers — total citations, h-index and i10-index.
The h-index, proposed by the physicist Jorge Hirsch in 2005, is the largest number h such that h of your papers have been cited at least h times each. Ten papers with ten or more citations apiece is an h-index of 10; one paper with 400 citations and nine with two leaves you at 2. The i10-index simply counts papers with at least ten citations, and it exists only in Google Scholar. Both are reported twice, all-time and over the last five years, which is the more useful column for anyone still building a record.
Two conditions decide whether your profile appears when a stranger searches your name, and people routinely satisfy one and miss the other. The profile must be public — it's private by default, and the wizard won't flip that switch for you. It must also carry a verified email address at your university or institution. Google's help page is blunt about the second part: gmail.com, hotmail.com, aol.com, yahoo.com and qq.com addresses are not suitable for the purpose. The address is never displayed on your public page, and Google states it won't sell or trade it.
If the verification email doesn't arrive, the cause is usually mundane. Institutional filters sometimes swallow it, a typo in the address kills it silently, and clicking the link while signed into a different Google account can leave the address marked unverified. Check spam, then ask your IT team whether mail from Google is being blocked before assuming anything more exotic.
Building it: six steps, one of which people skip
Sign in, then work through the wizard in one sitting. The order matters less than doing all of it.
- Go to scholar.google.com and sign in with a personal Google account, not your institutional one. Library guides at Monash University and Eastern Michigan University both recommend this, and the reason is practical: an institutional account usually dies when you change jobs, and the profile dies with it.
- Click My profile, in the top-left menu or the side drawer. If you see a profile creation form instead of your own page, you're signed in as the wrong account.
- Enter your name exactly as it appears on your papers. If you publish under two forms — married name, transliteration, hyphen — pick the dominant one and fix the outliers by hand later.
- Set your affiliation to the official name of your institution, in the form your co-authors use. Matching leans on this string, and abbreviations fragment it.
- Enter your institutional address in the email-for-verification field, then click the link in the message that arrives.
- Add four to eight plain keywords as areas of interest, in English. These influence how your profile surfaces in searches for topics, not just for names.
The wizard then shows a list of articles it believes are yours. This is where an otherwise tidy profile acquires its first flaw.
Photo by Declan Sun on Unsplash
The article list is where profiles go wrong
Google's suggestions come from a statistical model built on author names, bibliographic data and article content. It is good and it is not careful. For anyone with a common surname, an initialled first name or a large co-author network, the suggested list will contain other people's work. Deselect it before you save, not after.
Then hunt for what's missing. Click the plus icon and use "Add articles" to search the index by title in quotation marks; that catches papers Scholar has indexed but not linked to you, which is the common case for conference proceedings and anything published under a name variant. For work Scholar has never indexed, "Add article manually" opens a form asking for title, authors, date, journal, volume, issue, pages and publisher. It's tedious and it works immediately.
Why some papers never get crawled is worth understanding, because you can fix the cause at source. Google's inclusion guidelines require that the full text or the complete author-written abstract be freely visible without a login, pop-up or scroll; PDFs must sit under 5MB; the title needs to appear large on the first page with authors listed directly beneath and a section headed "References" at the end. Repositories built on DSpace, EPrints or Digital Commons are usually configured correctly out of the box, so depositing a copy there is often faster than arguing with a personal webpage.
Duplicates are the last chore. When Scholar indexes a preprint, an accepted manuscript and the publisher version, you get three entries and one inflated count. Select the duplicates together, choose Actions, then Merge, and pick the record with the best bibliographic data. A crossed-out "Cited by" number on one of the entries means Scholar already recognises the duplication and is counting it once.
Automatic updates, or a monthly email you actually read
During setup Google asks how it should handle new articles, and labels automatic updating as recommended. Plenty of university libraries recommend the other option: don't update automatically, send an email to review and confirm. The confirmation setting costs you a few minutes a quarter and prevents the Dr. M scenario, where a stranger's papers arrive without anyone asking.
Google's own documentation notes that automated updates won't overwrite edits you've made by hand, and that the setting controls only your article list — citation counts and metrics always refresh on their own. Two other switches are worth flipping while you're in there. The Follow button next to your name sends an email when one of your papers is newly cited. The Homepage field accepts a URL, and a link to your institutional page or ORCID record gives anyone who lands on the profile a route to the rest of your work; the ORCID identifier is a more stable anchor than a name, since names get reused.
What the numbers are worth, and where they break
Start with coverage, because the honest answer is that Google Scholar finds more than its commercial rivals and that the surplus isn't all equal. A systematic comparison across 252 subject categories, published in the Journal of Informetrics by Alberto Martín-Martín and colleagues, found that Scholar captured 95% of the citations Web of Science recorded and 92% of Scopus's. It also found a large amount those databases missed: in the humanities, social sciences, and business and management, citations unique to Scholar exceeded half of all citations in the category. Correlations with Scopus ran from 0.92 to 0.99, and with Web of Science from 0.78 to 0.98.
The composition of that surplus is the caveat. Between 48% and 65% of Scholar-unique citations, depending on field, came from theses, books, book chapters, conference proceedings and preprints rather than journals. That's a fair description of how humanities and engineering research actually circulates — and it's also a weaker quality filter than a curated journal index. A 2021 study by Fayaz Ahmad Loan and colleagues in Library Hi Tech, which verified 21,022 publications claimed by the fifty most-cited library and information science researchers, found discrepancies in publication counts for 58% of them, in citation counts for 58%, and in the h-index for 42%. Some of that is error. Some isn't.
The deliberate version is well documented. Yasir Zaki and colleagues at NYU Abu Dhabi scraped roughly 1.6 million Scholar profiles and identified 1,016 scientists whose citations jumped tenfold in a single year, then flagged 114 who had received more than 18 citations from one paper. In an undercover experiment they built a fictional researcher, populated a profile with chatbot-written preprints and bought 50 citations for $300. The invented author briefly ranked 36th most cited in their field, with an h-index of 19.
That's the exception, not the base rate. The base rate is Dr. M: a real record, badly parsed, costing a real candidate a comparison she should have won.
Photo by Vitaly Gariev on Unsplash
What this means for your lab
Treat the profile as equipment that needs calibration, not a set-and-forget form. A few habits cover most of it.
- Keep exactly one public profile. Two profiles split your citations and confuse anyone trying to assess you; if you find a second one, make it private or delete it.
- Sort your article list by title once a quarter and scan for work that isn't yours. Common surnames and initials are where false attribution concentrates.
- Put the public profile URL — it begins https://scholar.google.com/citations?user= — in your email signature, your CV and your ORCID record, so the version people find is the version you maintain.
- Check the verification status of your institutional address after any job change, since affiliation strings and addresses both go stale.
Two mechanical facts save panic later. Deleted articles go to a trash area and can be restored, and university guidance on profile management suggests a deleted profile may be recoverable within about 30 days of deletion. Neither is a substitute for care, but both mean a bad click isn't fatal.
Here's the next step, and it takes ten minutes. Open your profile, click the "Title" column header to sort alphabetically, and read down the list looking for one paper you didn't write. Delete it. Then set a calendar reminder for the first Monday of each quarter, and do the same thing again.
- academic-publishing
- bibliometrics
- h-index
- citation-metrics
- research-visibility

