by Max Mosterd (Head of Operations & Analytics, Knowledge Unlatched)
Open Access (OA) for books is still in its infancy, but publishers and institutions who wish to accelerate its adoption are becoming increasingly interested in analytics of OA content. Such analytics can help publishers and libraries to make informed decisions around open access books, such as in which discipline to start an OA publishing program or where they should be investing in order to increase the impact of their content. Knowledge Unlatched (KU) is addressing this issue by combining usage data from a variety of different hosting platforms. Quarterly usage reports help publishers to review the OA use of their titles, and libraries to track the usage by their own patrons.
Total Interactions from 2017 – 2019 for KU Select HSS Books
Over the last three years, Knowledge Unlatched has reported on usage data from several platforms for its open access books on a monthly basis. The focus of these standard reports has been on the direct relation of user interactions on a title basis. However institutions and publishers are also interested in how these interactions translate to a specific region or a certain discipline, such as which disciplines show the highest uptake of OA books.
To date, the Knowledge Unlatched KU Select HSS collections have made 1,049 books available open access, of which 968 books include interactions between 2017 and 2019 (see Table 1 below). The remaining 81 books are either still to be published, have no interactions as yet or lack certain metadata tags and therefore have to be excluded from the analysis. It is important to stress that even if the period covered for this analysis is the same for all titles, the publication dates will have an impact on usage (e.g., books from the KU Pilot in 2013 have been online far longer than e.g., KU Select 2017 HSS Frontlist Books). At the same time we can expect to see different usage patterns for backlist and for frontlist books.
Table 1: Usage of KU Select HSS Collection
Books from 2017-01 until 2019-12
The OA books analysed are hosted on a variety of platforms. The platforms to which titles are continuously uploaded, and thus where usage data is available, are JSTOR, OAPEN, Project MUSE and the Open Research Library/Bibliolabs (prior to ORL) — all of which are COUNTER conformant. The OAPEN platform currently hosts the most titles, although a few titles are still missing — which might indicate that no interactions were found for the timeframe 2017-01 to 2019-12. There are many differences across these platforms making it something of a challenge when attempting to compare the usage – including varying publisher admission policies (MUSE focuses on non-profit presses, JSTOR only accepts content from its member presses etc.), different times of ingesting KU books (ORL is only in beta since mid-2019) and of course the different usage metrics and methods applied. These are important reasons which make any attempt to compare usage on a platform-level problematic.
That said, KU does wish to provide analytics on a more general level, so took the decision to aggregate the different usage metrics into to one single figure which we call “Interactions,” meaning the total sum of all Investigations, Chapter Downloads, Chapter Views, Book Downloads, etc. It is clear that this is by no means an ideal way of analysing usage, but it does allow us to observe trends across several dimensions.
Table 2: Usage of KU Select HSS Collection Books from 2017-01 until 2019-12, by Platform
Total Interactions over Time
At KU, the monitoring of and reporting on usage is becoming increasingly important. Institutions funding these OA collections require impact reporting to provide to their administrations and thus to justify the investments which have been made. Hence one of the key metrics tracked by KU is the average number of interactions per title, the aim of which is to increase the total year-on-year as more books are made available. Figure 1 shows how this has developed over the last three years, showing that the total interactions per title has almost doubled each year. We assume that this effect is in part driven by the inclusion of more new titles and platforms, but is also due to better indexing and thus increased visibility of the content to researchers.
Figure 1: Average Interaction per Title
To get a better idea of the long-term trends regarding interactions on a title level, the top performing title is displayed in Figure 2. As could be expected the highest usage takes place within the first few months after publication. Over time interactions decrease slightly, in line with research and teaching semesters. Similar to findings from other studies this demonstrates that a significant portion of lifetime usage takes place within the first few months, followed by a long tale of fewer interactions over time. In line with its unlatching process KU encourages the publication of OA titles in the first two quarters of each calendar year, as well as the fourth quarter, as these tend to be the most intensive periods for research and teaching and where the most usage takes place.
Figure 2: Usage Lifetime for KU Select Title
with most Interactions
Regional Developments of Interactions
Looking at the geographical dimension, most of the interactions originate from North America. This is not surprising seeing as KU Select’s customer base includes many institutions from this region (see Figure 3). Over the course of the three years North America has also increased most in terms of total usage worldwide with an increase of 621% in 2019 compared with interactions in 2017. The continent ranking second in terms of increase in interactions is Africa, starting with 17,000 Interactions in 2017 and increasing by 440% to 70,000 interactions in 2019. Europe (340%), Asia (355%) and Oceania (316%) all follow the same trend, while South America has “only” increased the actual use of KU select content over time by 110%. A possible reason for this might be that all KU Select books are published in English, while the dominant languages in South America are Portuguese and Spanish hence there is less uptake of KU Select books over time. Obviously, the inclusion of more platforms also helps to boost visibility to researchers and thus increase the total number of interactions.
Figure 3: Total Interactions over Time by Region
Interactions by Discipline
The share of the OA book market by discipline is subject to much speculation. Our analysis across publishers and disciplines reveals that “Modern Languages and Linguistics” clearly stands out as the key discipline in terms of average interactions per title.
Figure 4: Total Interactions by Discipline per Title
The Ideological Battle: OA by Business
Model and Size of Publisher
KU Select encompasses books from 78 publishers from the period 2017-2019, including big and small ones, not-for-profits (largely University Presses) and for-profits. A particular area of interest was how title interactions performed according to each publisher category. And we were pretty surprised at the results: titles submitted by non-commercial publishers performed 13% better than the average across all 968 titles. The group of monographs from commercial publishers however underperformed the average by -23%.
Our study cannot yet provide evidence to explain the reasons behind these differences, and we will have to leave it to future research to investigate these further.
Figure 5: Interactions by Publisher’s
Looking at usage based on publisher size the results were also surprising, at least at first sight (Figure 6). According to this analysis the group of monographs from small publishers shows the best performance compared with the interactions of the complete sample, with a standard deviation of +18%. While the books from medium-sized contributors performed in line with the average (+1%), those of the larger publishers significantly underperformed against the average. While we have not yet done any further research on this, we attribute the trend primarily to a technical effect. All large publishers have their own powerful online distribution platforms (not included in the interaction count), and this is not the case for many small publishers. At the same time the results demonstrate that open access publishing can help smaller publishers to overcome the structural disadvantages they often face compared to the larger competitors.
Figure 6: Interactions by Publisher Size
The study of 968 English language research monographs across 78 publishers from all over the world is the first of its kind, and it helps to highlight key issues regarding OA books. Despite its obvious limitations (e.g., lack of reporting standards across platforms; the natural effect that interactions with OA content will also take place outside of the platforms examined), we have identified certain trends:
1. OA books are a rapidly expanding category, not only in terms of publishing output (as others have shown), but also regarding the usage of content.
2. Growth is strongest in the highly-developed and wealthier regions of the world, particularly in North America.
3. Researchers’ Interactions with the content still show significant variations by discipline.
4. Open access monographs from non-commercial and smaller publishers benefit disproportionally high from the access model.
The study also shows that we are still very much in the early days of research into OA book usage as a whole. More analysis needs to be done to contextualize the results, particularly comparing it with paywalled content. That step will be the litmus test as to whether open access does actually bring the benefits which funders are hoping for.
For Knowledge Unlatched the findings have also very practical implications regarding the tailoring of content in future pledging rounds. KU Select will be subject to a stronger topical focus on those disciplines where it has seen strong usage. While one could argue that this decision will be detrimental to those disciplines with lower usage, our focus is on a strong return on library funds invested. For other disciplines KU will offer dedicated partner packages which will help to better reflect the value of these “collections” to specialized audiences in the research world.
In addition, more analytical research will be carried out in the near future on the actual use of OA books by countries in the so-called Global South. We will be interested to examine how not only those disciplines with less usage, but also those regions of the world with currently less interactions, relate to open access. There may be special requirements from such regions which we are not currently fully aware of but which could help to increase the positive outcomes of the access model.
Implications & Conclusion
Moving forward, KU Select plans to apply a stronger topical focus on those disciplines where it has seen a strong engagement. This will already be reflected in the upcoming cycle of KU Select 2020. Usage data thereby helps to inform the OA programs KU promotes in order to encourage a high impact on the OA investments participating institutions and consortia make by supporting KU collections.
In addition, more analytical research will be carried out in the near future on the actual use of the OA books by the global south. As a result of such research, Knowledge Unlatched plans to not only create collections that have a strong impact potential, but also strives to reach those countries and regions that need such Open Access the most. Leveraging these insights therefore holds great potential to accelerate Open Access and reaching relevant audiences.