Case Study: Book Recommendation App
Over the Spring 2022 semester as part of my UX Design & Strategy class at Claremont Graduate University, I conducted research to evaluate a new product concept through interviews, competitive analysis, and a survey. My final deliverable was a data-driven persona. View the summary deck here and keep reading to learn about my process!
Have you ever wanted to read a book, but just can’t find something you’re in the mood to read?
This is a problem I’ve encountered and wanted to solve. Today, when readers want to find new books to read, they have to search the internet, look at Goodreads lists, or ask their friends. This can be a time-consuming and futile process because there is no centralized place to find book recommendations, recommendations given from these current sources may not match the reader’s interests, and readers still have to go to another website or store to access the books. My goal was to evaluate the concept of a one-stop shop for readers to discover and access the books they love.
Interviews, competitive analysis, and a survey were used to understand readers' book discovery process and interest in a new book app.
I developed a research plan to understand readers’ processes for finding books, determine the prevalence and severity of pain points, and evaluate if they would benefit from another book app. Specific hypotheses to test were as follows:
Readers are not satisfied with their current methods for finding books to read.
Readers use several sources to find book recommendations including, but not limited to, Goodreads, social media, and word of mouth.
Readers would like to receive book recommendations that take into account their interests in genres, tropes, movies, TV shows, mood, and more.
Readers want to have options for where to purchase a book.
Readers are willing to use and download another app to receive book recommendations.
There is room in the market for a new book app that provides readers with curated book recommendations.
Exploratory interviews were used to evaluate hypotheses 1, 2, and 3, competitive analysis was used to evaluate hypothesis 6, and a survey was chosen to further evaluate hypotheses 1, 2, 3, 4, and 5. See the research plan for full details on methods, recruitment, timelines, and research materials.
Methods
I interviewed 3 readers to uncover insights about their reading habits and book discovery methods, resulting in a provisional persona.
I first conducted interviews with three people who consider reading a hobby and search for new books at least once a month. These interviews were used to get to know prospective users and understand their current reading habits and book discovery process. I leveraged thematic analysis to identify themes among participant responses and build the provisional persona shown in Figure 1. View the interview script.
Exploratory Interviews
Figure 1. Provisional persona
Likewise and Goodreads are formidable competitors in the book recommendation market, with Likewise utilizing machine learning to enhance user-specific suggestions, while Goodreads serves as a hub for readers to track, rate, and connect over books.
After the interviews were completed, I performed a competitive analysis to evaluate the current market. I identified two direct and three indirect competitors. Likewise and Goodreads emerged as the most threatening competitors because they offer similar book recommendation experiences and are widely used. View the full competitive brief.
Likewise is a direct competitor that offers users book (and other media) recommendations based on the user’s interests. The recommendation engine uses machine learning to improve recommendations over time. See Figure 2 for more details.
Competitive Analysis
Figure 2. Likewise competitive analysis summary
Goodreads is an indirect competitor that provides readers a place to track books they’ve read, make lists of books, rate books, connect with other readers, and discover new books. Goodreads is very popular among readers and is owned by Amazon. See Figure 3 for more details.
Figure 3. Goodreads competitive analysis summary
In the final phase of this research project, I surveyed 28 readers to learn more about their book discovery methods, their satisfaction with these methods, and the challenges they face.
As this last phase in this research project, I ran a survey to obtain more data about readers’ book discovery process. The survey employed multiple-choice questions, free-response open-ended questions, and Likert scales to measure how readers currently find book recommendations, their satisfaction with each method they use to find books, the prevalence and severity of pain points identified in the interviews, and their interest in the concept’s value proposition. The full questionnaire can be viewed in the concept evaluation plan.
Survey participants were recruited via convenience and snowball sampling to reach readers while constrained by time and resources. Ideally, 100-400 responses would be collected; 400 responses would reach statistical significance, but 100 responses would be sufficient to identify themes. Because this was a class project and was limited by time and resources, 28 responses were collected.
I analyzed survey data using descriptive statistics to summarize participant behaviors and preferences uncovered from the multiple-choice questions, as well as thematic analysis of free-response answers to identify motivations, challenges, and needs.
Survey
All insights informed a data-driven persona.
The survey and exploratory interview insights were used to create the data-driven persona shown in Figure 4. This persona is an updated version of the provisional persona, based on real survey data from 28 readers.
Data-Driven Persona
Figure 4. Data-driven persona
I’d recommend integrating this concept with Goodreads rather than making it a standalone app, but more research is needed to make it a success.
While survey respondents claimed they would try out this app and responded positively to the value proposition, the solution should be implemented as an addition to Goodreads for the best results. The solution should not be a standalone app because:
The problem of finding good book recommendations is present, but not severe.
Readers would like to receive curated book recommendations and are already using Goodreads, so would welcome the addition to Goodreads.
Goodreads is already widely used, however, other book-related apps have low adoption. It would be difficult to build a successful book app that is only offering one main feature (recommendations).
To bring this solution to market with Goodreads, further research should be conducted to prioritize features and test a prototype. The research questions include:
How do readers want to receive book recommendations? Do they want them delivered or do they want to search manually? Both?
How often do readers want to receive new recommendations?
How often do readers’ book interests change?
Is the proposed design usable and delightful?
Recommendations & Next Steps