Improving the Visual Design and Information Architecture of the Source Types of ProQuest’s Basic Search Experience of Search.proquest.com
Researched and Written by
Mina C. Hami
M.A. Media & Information,
Human Computer Interaction
User Experience Designer
Revised Report December 8, 2019
The study aims to redefine how we visually communicate results to ProQuest’s diverse users including advanced researchers, grad, undergrad, faculty and Ph.D. candidates. The study’s hypothesis claims that current design does not help users evaluate their searches in the most effective way. This research suggests that the current experience misses out on some of the most important information to users needed to help evaluate the results. (i.e. The way a user evaluates a book maybe different then the evaluation requirements for a dissertation or video). Currently the search results do not give the user the ability to evaluate a book based on the cover, chapters, or relevance to the user, nor does the search experience allow users to accurately categorize their search queries.
The study took place from May 2019 until November 15, 2019. The researcher is both a M.A. graduate student at Michigan State University and an employee of ProQuest LLC.
ProQuest LLC. is a global content and technology organization based in Ann Arbor, Michigan. Since it was founded in 1938, the organization provides libraries with information content including an array of historical archives, scientific breakthrough content for scholars, students, and professionals.
For the purpose of this study, research was focused on the primary ProQuest basic search platform (PQ1): search.proquest.com . Within this platform, users are able to conduct basic searches and advanced searches, browse by publication, and search through the 100+ databases available through their library’s subscription with the company.
GOAL OF STUDY
The goal of this study is to address the UI issues & problems with the current ProQuest One platform source types and identify ways to improve the UI & user experience to customers. To achieve these goals, the researcher has already completed secondary research and a comparative analysis of other social media platforms that use icon/thumbnails to support their information architecture. Information found from secondary research was presented to design team in order to inform this user study.
The researcher, Mina Hami, is a graduate student from Michigan State University under the advisement committee of Dr. Carrie Heeter and Dr. Ruth Shillair. The study took place over a five month duration, from July to November of 2019. Mina is a user experience designer employed by ProQuest LLC and assigned to this project under the direction of User Experience Lead and manager, Austin Bedell. The findings from this research will supplement other on-going research and implementations addressing source types and visual components of the main ProQuest basic search platform.
METHODOLOGY + PROCESS
This section of the report describes the process and tools the researcher used to conduct the study, collect data to gather findings and recommendations provided later in this report. The following information provides a brief explanation of those processes and tools used:
Researcher Heuristic Evaluation of Current User Interface (UI) of Basic Search Experience
User Personas - Assignment Researcher + Domain Learner
IRB, Participant Recruitment & Limitations
Preliminary Participant Survey
Participant Consent Forms (IRB and ProQuest Compliant)
Session 1 - October 29, 2019
Session 2 - November 11, 2019
Session 3 - November 14, 2019
Design Process + Prototypes
1.0 Researcher Search Experience Heuristic Evaluation
Before jumping into the design and interactive prototypes, the researcher conducted a heuristic evaluation of the current search experience. The results of the heuristic evaluation can be found below.
According to Usability.Gov, visual design should add aesthetics to a user interface (UI) of a site through items such as colors, images, and fonts . The visual design should not take away or provide distraction to what is upon the screen .] However, upon evaluation of the current search experience, a number of UI issues arise. To address some of the issues the researcher used Jakob Nielson’s ten principles of interaction design to carry out an evaluation. Of Nielson’s ten heuristics evaluated, three principles had serious issues. They include aesthetic and minimalist design, consistency and standards, and match between the system and the real world.
To read the full heuristic evaluation, click here
To help shape this study, the researcher used pre-established user personas provided by the organization. Out of the many user persons that the UX ProQuest team uses, for the purpose of this study Assignment Researchers were the primary focus for recruitment. Assignment researchers are categorized as their primary task being understanding a topic and satisfying it via research. So since the primary participants that were recruited for this research, were graduate students many of them would be categorized as assignment researchers.
However, there were a few users that were more of the domain learners as well. These participants were those who have conducted research or were in the process of working toward a larger project either for work or school. These personas were more in the second and third round of user testing during the mockup feedback sessions. But for the majority of the project Assignment researchers were the primary participants.
3.0 IRB Approval + Participant Recruitment + Limitations
Before testing could take place, the researcher had to get IRB Approval for testing. Once approved the researcher could move forward in participant recruitment. However, the IRB process took a little longer than what the researcher was used to because the researcher currently works for ProQuest, LLC as a UX designer and a graduate it was made sure that every angle of the protocol for testing was covered. These hiccups in the IRB procedure, pushed back the recruitment timeline.
Due to time constraints, the researcher was able to only able to recruit a limited number of users. A majority of participants were graduate students within the assignment researcher persona. Since the study used one of the courses the researcher was in a professor did participate in the study. The results from the professor’s participation is included in a separate section because of its outlier and more advanced feedback. A total of 14 individuals were able to participate in this study. Based on timeline factors to conduct and analyze the results, the researcher decided to use convenience sampling by recruiting classmates and peers within the Communications graduate school.
A link to the IRB protocol and approval letter can be found in Appendix M
4.0 Preliminary Participant Survey
Before each user session, each participant was invited to fill out a Preliminary Participant Survey. The survey asked participants demographic information that was not personally identifiable. The survey also asked participants about their experience with the ProQuest basic search platform search.proquest.com, their research habits and perceptions/ expectations of the different type of scholarly content.
To review questions of the Survey see (Appendix A).
To see the results from the survey see Appendix I or on the following page.
5.0 Participant Consent Forms
For all moderated testing and activities, the researcher provided approved consent forms outlining participants about the research study, their freedom to decline to participate, the importance of the company information being tested and the use of video and voice recording, see Appendix B.
A consent form was also provided at the beginning of the preliminary participant survey (see Preliminary User Survey in Appendix A.
6.0 Moderator’s Script & Testing Activities
Moderated Testing & Activities
The goal of moderated testing was to gather real-time user feedback on the existing visual components of the basic search experience and working prototypes produced by the researcher. There was no interventions used. Participants were given a consent form telling them that their actions are voluntary and that they can resign at any time during testing.
Participants were also given a Preliminary User Survey at the beginning of the sessions. The purpose of the survey was to gather basic information on the participants, their familiarity with ProQuest, and their research habits.
The researcher conducted Session One, with Casey O’Donnell’s Fall Semester 2019, MI 893 Game and Film Design Studio 1 class. The course met on Tuesdays at 10:20 a.m. on Michigan State University’s campus in the Communication school’s Room 175. Session two and three took place on November 11th and November 14th, via the researcher’s msu Zoom meeting space. The sessions were recorded using Zoom’s online software in order to gather data to analyze after the sessions were completed. Participants were provided with PDF attachment the consent form, and survey links to the survey via Qualtrics and prototype links via Adobe XD.
Session 1 | Class Interactive Activity Appendix C
Session 2 & 3 | Prototype User Session (Appendix D)
Preliminary User Survey via Qualtrics (Appendix A).
7.0 Design Process + Prototypes
The researcher used the design thinking methodology to shape the study and create the prototypes used for testing in later sessions. The design thinking methodology is defined by five stages: Empathize (to understand the problem you’re trying to solve), Define (the problem that you’re trying to solve), Ideate (start producing out-of-the-box ideas), Prototype (creating scaled down ideas to experiment/test) and Test (using the prototypes done in the previous step to gather user feedback: what they think of the design, behave and feel).
For this study, once the hypothesis was defined and a study procedure established, the researcher sketched initial ideas based on the known issues already with the basic search provided by ProQuest UX team manager, Austin Bedell, and the outcome of the heuristic evaluation done by the researcher. Based on that feedback below are some of the initial ideas. These ideas were shown to participants during a follow up session for the satisfaction of the MI 893 course project update. While the researcher did not test these ideas with the first moderated session, they were the foundation for the next set of prototypes. Once there was feedback from the first moderated session with the MI 893 course. the researcher sketched (Appendix G) and prototyped four variations to be tested in the second session with users on Zoom and then again with more ideation for the final session.
(The image to the above is a depiction of the prototyping process done to create interactive prototypes)
FINDINGS + ANALYSIS
This section covers the summary of findings from the testing methods used to evaluate the current basic search experience but using a collaborative focus group method for Session 1 with the MI 893 course at Michigan State University. The prototypes designed for Session 2 A-D 1.0 based from the feedback in the Session 1, and also the improved prototypes A through D 2.0 tested in Session 3.
This section includes design recommendations that may improve the visual communication components of the search.proquest.com experience based on the combined feedback from sessions 1 through 3. The recommendations are organized into the following major categories:
Make keywords more visible for users on the results page.
Create a better quick view/look inside preview experience to allow users to view content to help determine its relevance to their research goals.
Improve suggested search term functionality for users that gives a better end experience.
Revisit the importance of the image functionality across multiple source types.
1. Make keywords more visible for users on the results page.
Based on the feedback the design through the different sessions, there was a clear importance and request for keywords to be readily available and easily visible for users. Currently in the UI key words are shown within the meta-data of the query result and under the details page labeled “details.” In the mockups A through D the keywords were presented in different designs. However, it seems the keywords should be pulled out and more readily visible to users. Below are some iteration ideas concluded from this study:
In the top iteration the keywords are part of the text below the title, author and functions.
In the middle iteration mockup below, the keywords are part of the abstract for the resource.
The bottom iteration the details wording was confusing to participants, because they were not sure if details meant keywords, or subjects within the text. To the researcher, this seems like details actually means keywords. Therefore, the wording was modified to include this on the full information page.
2. Create a better quick view/look inside preview experience to allow users to view content to help determine its relevance to their research goals.
The ability to have the Amazon-like feature to ‘Look Inside’ a book, journal, newspaper or quick preview of a video seemed to fair well with users across the user group sessions. However, how to present that information in the most intuitive way where it is useful to users is a challenge. While the idea initially was initially to do an icon with a magnifying glass in the corner of the thumbnail, it was easily apparent to some users, while others struggled with the concept. When the word “preview” was shown to the users in word form, although it was slightly more helpful the placement wasn’t always the best. Therefore, below are some ideas from the collective feedback on the placement and function of the preview function.
3. Improve suggested search term functionality for users that gives a better end experience.
The need for suggested search terms came up was present in a few of the user session feedback. While the platform currently has auto complete function which could double as a suggested search term feature, it is not really apparent to users. Instead of doubling the functionality in the UI (seen first below), the suggestion is to break it out into its own component.
4. Revisit the importance of the image functionality across multiple source types.
Based on the feedback from the user sessions while images that show the book covers or journal covers were useful in some cases, they were not useful in other cases. Visuals should ultimately add aesthetic and meaning to a layout or design. For addressing the book source type, book covers made sense to participants. Yet for addressing scholarly journals and newspapers covers seem to not be the answer, and did not fair well participants in this study. Currently it is unclear what the next steps to address this issue is, thus the researcher believes that feedback from ux team could help with finding a conclusion.
While the basic search for search.proquest.com does a good job at presenting a clean interface and offering multiple sorting options for all levels of researchers, it struggles with offering users the ability to effectively evaluate their search results across different source types. Based on the study, participants did support the claim and offered suggestions for improvement such as the preview vs abstract functionality, speed of website, the ability to see keywords more clearly and adding a suggested search term result. The only piece of the hypothesis that still seems to be unclear is the true usefulness of the thumbnail visuals (ie, book covers, journal covers, etc.). The researcher hopes to open that piece of feedback to the UX team and potentially do a smaller study that just focuses on solutions.
This study, is just one piece of a larger study that has been ongoing at ProQuest LLC, within the UX design team. The valuable information found in this study, will support the validation and verification of some of the new changes that the team has already developed and launched, as well as offer feedback on product improvement. Following this study, the researcher will be presenting key findings to the UX team and gain more collective feedback
The researcher would like to thank her advisors, Dr. Carrie Heeter and Dr. Ruth Shillair for their knowledgeable guidance and support throughout the project as part of the graduate school program at Michigan State University. The participants who took time out of their busy schedules to participate in this study. As well as the ProQuest User Experience team and managers, Austin Bedell and Stephen Kemsley with their guidance and feedback for the longevity of this project.