How do students and researchers find content? This is a topic that comes up again and again in this column, since it seems to be so frequently misunderstood. Much of the scholarly landscape changes rapidly, but the discovery process remains, I think, fairly consistent. So it is all the more peculiar when academics focus on one small part of the discovery process and analyse it as if it were the only, or the most important, path to discovery.
A recent article (Open Access publications drive few visits from Google Search results to institutional repositories) by three Spanish researchers looks at the discovery of content in institutional repositories. Now, institutional repositories are just one of many locations where researchers will find content, and almost certainly not the first – they were not designed for this purpose. Lots of research has been done to discover how users find content, including the well-known survey carried out every five or so years by Inger and Gardner (and which I wrote about in 2021). They reveal several routes to discovery, including Google search, the library discovery service, aggregators, and others . Their results can be summarised by the graphic on the front of their 2021 article (shown above). Just 10% of article discovery is via Google, and that includes both Google and Google Scholar.
There are plenty of other articles that confirm this assessment, for example, Kibble (2022) comparing UK and US students, which mentions the importance of library discovery services, and points out that Google Scholar is used more than Google itself.
If Google search is such a relatively little-used way of finding content, why devote a whole article to this discovery channel? Moreover, why use a commercial SEO tool, Ubersuggest, billed as a “FREE SEO tool that enables you to do keyword research, analyse competitors, do search audits, and more!”, to judge the content’s discoverability? Not only is this article asking the wrong questions, but it is also using inappropriate tools to measure the results. Every web resource does not have to be judged using SEO tools.
I am a great defender of repositories and the work they do, but I will readily admit they are not the first place to go to when looking for new content. Repositories were set up to foster open science and open research, by providing access to scholarly content that would not otherwise be available – the principle of “green” open access. Using tools such as Unpaywall, which locates a freely-available full text version of a paper, they have an essential role in providing wide access for those researchers who do not have a lavishly funded library that subscribes to many journals. I’m not quite sure how you would measure such usage, but that’s not the point. Providing fine-tuned keywords and SEO tools are not part of the scope of repositories; it is sufficient that an article is available at all. Leave SEO tools to commercial sites, where your position in the search results rankings is all important, and recognize repositories for what they do.
Google Scholar researcher
This study and analysis looks way off. If only “10% of article discovery is via Google, and that includes both Google and Google Scholar” — the repository is very likely not being indexed properly in Google Scholar. For anything properly indexed, about 90% of traffic will be coming from Google Scholar and web search.
Michael Upshall
Thanks for the comment, but I think it reveals exactly what I was trying to explain. Not all sites need to be “properly indexed” for Google Search or for Google Scholar. It may not be appropriate for some sites to have 90% of their traffic coming from search engines. Strange as it may seem for dedicated adherents of SEO, this is not the only measure of success!
Aaron Tay
I have to read the new Spanish paper but for sure for our repositories the vast majority of visits are due to Google than Google Scholar (i was a bit surprised cos I expected it to be the other way around). Even more surprised Spanish repositories are different and Google/Google Scholar drives a minority of visits.
That said I wonder about thecsurvey https://www.consultmu.co.uk/how-academics-discover-new-content/ . Is it self reported? If so I wonder how accurate it is.
Secondly I don’t think there’s a conflict between saying researchers rarely find content via repositories (repositories are a last resort if yoy have institutional access）and saying repositories get most visits due to Google or Google Scholar.
Michael Upshall
Thanks for your comments, Aaron. The study of how academics discover content was by Inger and Gardner, done every few years, and the specific research I referenced was the 2021 version. Unfortunately, the original site where the research was published is no longer running, but you can find several linked studies to this research, for example https://content.iospress.com/download/information-services-and-use/isu800?id=information-services-and-use%2Fisu800.
My point about comparing Google (Search or Scholar) and other paths to discovery was simply to point out that even if users could access repository content via Google, that is not the point of repositories. If repositories really wanted to make their content more discoverable via Google, they would invest in keyword services and such like – but they don’t, and I think quite rightly. Repositories are not set up to be discovered via Google, and I think that’s quite sensible.