Focusing on the users

Figuring out what users want and how they behave are some of the biggest challenges for publishers and for the companies that they outsource their technology to. This is made more complicated by changes in the information that users are accessing, the technology and in the users themselves.

Firstly, there have been dramatic changes in the amount of information available. In addition to the inevitable yearly increases in published works, there is the current push towards digitising historical articles, some going as far back as the 19th century. And the bundling of content into packages and consortial licensing arrangements is allowing libraries to purchase more publications by lowering their per-article cost. Meanwhile, the nature of the information is also changing. In addition to traditional journals in electronic form, publishers are also offering new publication types such as electronic books, reference works and media collections.

These changes mean that users may now encounter publisher websites that offer thousands of publications and millions of articles and other materials. Such extensive collections require sophisticated tools to allow the user to navigate and use the content efficiently. Users will modify their content-finding behaviour and navigation strategies from the way they used smaller collections. They must also master additional types of tasks that are now enabled by this expansion, such as access to reference material.

Interconnected content

Another factor that has evolved is the way that electronic content is interconnected. Thanks to greater support of linking standards such as OpenURL and Digital Object Identifiers, along with wider distribution of metadata, users have never had more choice in where they begin their online research. New indexing services for scholarly material, such as Google Scholar, provide new avenues for users to find content. The full text of articles is now linked not only from the publisher’s website but also from metadata found through external indexing services, in aggregated databases, federated search engines and other external sites. These myriad choices are having significant effects on how users approach research tasks.

Underneath all these issues, the underlying technology of the internet is also moving on. With the advent of sophisticated ‘Web 2.0’ applications that are available to a wide internet audience, such as Google Maps, users are beginning to expect a richer, more interactive user experience from scholarly content on the web. New programming techniques such as ‘AJAX’ (Asynchronous JavaScript and XML) allow websites to offer a user-experience that approaches the speed and interactivity of desktop applications, providing a much more efficient and enjoyable research experience.

Understanding how the above trends affect the actual behaviour and needs of e-content users is vital in developing the user experience of services such as our new MetaPress 2.0 e-content hosting platform. For this reason, our research and development team carries out an ongoing series of studies of user behaviour and attitudes. We conduct usability studies with actual users because there is often a discrepancy between what we think users will do and what they actually do. We also mine large volumes of website analytics data.

The e-content user experience

Many factors affect the e-content experience for end-users. These include how they navigate to content, how they select content for viewing, and which content they actually view. Some factors are directly related to attributes of the content itself, while the environment in which the content is presented also plays a major role.

The most important factor related to the content is the quality and desirability of the material itself. No matter how easily and rapidly content can be located; if it is not of interest to the user, then usage will be low. The content quality factor should be the publisher’s primary concern.

After content quality, findability (the ease with which a content object can be located) and usability (the ease with which an object can be utilised by the user) are the main factors that affect usage and the quality of the user’s experience. But findability precedes usability – if the user cannot find the content, he or she will never have a chance to use it. Optimising findability and usability requires understanding of what users actually do when using e-content websites.

Guiding principles

There are three basic principles of user-behaviour on all websites that provide a basis for understanding more subtle aspects of user-behaviour on e-content sites, as Steve Krug described in his 2004 book Don’t Make Me Think (New Riders):

• ‘Users don’t read web pages. They scan them. Users are often under time pressure to finish a task on the web quickly, which contributes to this phenomenon. Users also know that they don’t need to read everything on a page to gain value;
• Users don’t make optimal choices. They know that there is not much penalty for making a bad choice on the web, as the ‘back’ button is just a click away; and
• Users don’t figure out how things work. They muddle through. Users often navigate sites inefficiently, don’t use the ‘right’ features at the right time, and often take circuitous paths to content. When faced with many options, users will not systematically explore to find the best choices.’

Understanding these principles helps us interpret web analytics data, such as from web server logs, as well as directly observed user-behaviour. And our studies have uncovered some additional principles that particularly relate to scholarly e-content:

• E-content access is a secondary activity. Users generally access e-content not as a primary goal but as a means to a different end, such as writing a paper or collecting evidence for research projects.
• Users just want the content, and quickly. Our analysis of web analytics data shows that users do not linger at e-content websites. Sessions are generally very short (less than four minutes on average on MetaPress publisher websites) – once users find a relevant content object (for example, an article of interest) they generally take that object away and end their session.
• Few habitual users. The vast majority of users of a given e-content website will be first-time visitors or novices. There are very few habitual users and expert users of a given e-content site.
• The Google effect. Thanks to the ubiquity of Google-style searching, users have become habituated to being able to find large volumes of relevant information quickly and do not expect to be required to spend large amounts of time attempting to locate the ‘perfect’ content object. Users generally spend very little time examining search results and often view only the first few content objects in a list.

Part of the design

Understanding these issues has helped us in designing MetaPress 2.0 to take advantage of actual user-behaviours. For example, this latest release of our platform has a flattened navigational structure. Text is simplified and available choices are kept simple to increase the likelihood that users will make useful decisions without being overwhelmed. It offers Google-style searching by default to allow users to utilise their existing knowledge to quickly find content with practically no learning curve.

The other trend that has been addressed is the way that most users will arrive at a publisher’s site via a direct link to a content object from another site, such as a federated search engine or an aggregated database. For this reason it is important to support durable URLs, OpenURL linking, and DOIs for all content objects. Wide distribution of metadata also enables linking from as many source sites as possible, such as Google Scholar, databases, and e-resource access tools used by libraries. RSS feeds and email alerts further enable direct linking and shorten the user’s path to content.

Facilitating discovery of related content is also important for users. Our platform offers links to material that is related to the content or search results that the user is currently viewing. Users can interactively filter search results, such as by subject, without having to resubmit an entirely new search. This allows the user to filter out unwanted content and be exposed to more content that is related to their goal.

Finally, it is important to note that every user is different. To really give a good user experience it is important to adapt the presentation of content to the end-user’s computing environment and research style. For example, in MetaPress 2.0, users with high-capability computers can use advanced interaction capabilities when they view extremely high-resolution graphics but users on lower-end devices will still have a basic interface that still allows them to see the images.

Other ways to adapt to different users and research styles include online article collections, the ability to save search criteria and be notified of new results, support for multiple content formats and RSS feeds. And providing the option to access articles as HTML or PDF files is necessary because some users like to read articles online while others, despite the recent emphasis on technology, prefer to print articles out and read them away from the computer screen.

Gary Coker is director of research and development at MetaPress