7. Rating Systems for Restaurants : Likes vs Stars
Posted on Sunday, 26 March 2017
#Post #LiteratureReview #RatingSystems #UserExperience #UserInterface6
Introduction
Rating systems are an important aspect of designing an experience for users to discover restaurants, and displaying ratings to users in an effective and efficient manner is of great importance. While it is not possible to identify the ideal rating system that caters to every situation (Winkelmann, Herwig, Poeppelbuss, Tiebe, & Becker, 2009), identifying the rating representation which caters to the requirements and preferences of the largest number of users is critical in designing an effective and efficient user experience.
An example of a 5-Stars Rating System in the Dining Application
The Like vs The Star
Multiple ways of displaying ratings to a prospective user are possible, and two important ones are “Percentage or Number of Likes (and Dislikes)” system used by Facebook and YouTube, as well as the “Five Stars Rating System” used in many E-Commerce Websites.
A conference paper written by researchers from the University of Newcastle, showcasing a study made on the usability and effectiveness of web-based trust rating systems (Pranata, Skinner, & Athauda, 2013), was reviewed. This paper surveyed the usability of three different types of rating systems, gathered qualitative user opinions on them, and identified the most and least preferred type of rating system.
Type 1 : Binary Rating Systems
The Binary Rating System refers to a rating system in which users rate an entity by choosing “1” ,“0” or “-1”. Users can review the rating based on a number or percentage of likes (equivalent to a 1) and dislikes (equivalent to a -1).
While the Binary Rating System scored the highest in the fact that it was easy to use (Pranata et al., 2013), there was a significant number of respondents who felt that it was “insufficient in reflecting the real rating” (Pranata et al., 2013), and that there was a shortage of rating options used (Pranata et al., 2013).
An example of a Binary Rating System
Image credits : https://www.researchgate.net
Type 2 : Notation-Based Rating Systems
The Notation-Based Rating System refers to a rating system in which English words are used to demarcate rating choices, such as “Very Good”, “Good”, “Bad”, “Very Bad”.
A significant number of respondents indicated that it is less ambiguous than the binary rating system (Pranata et al., 2013), while only a minority indicate that it provides too many options (Pranata et al., 2013).
An example of a Notation-Based Rating System
Image credits : http://garciniacambogiarevealed.com
Type 3 : Five-Stars Rating System
The Five-Star Rating System allows for users to rate based on choosing a number of stars to award to the particular entity, with 5 stars signifying the highest quality and best rating score. Users who view the rating are usually provided with an average of the five-star score, averaged over all the users who provide ratings.
The Five-Star Rating System scored the highest, with the majority of respondents indicating that “it is able to reflect the real rating” of the entity (Pranata et al., 2013). Furthermore, this system emerged as the “most preferred rating system” for both users who are asked to give their ratings, and for users who are viewing the ratings (Pranata et al., 2013). In contrast, the Binary Rating System emerged as the least preferred for both classes of users (Pranata et al., 2013).
An example of a Five-Stars Rating System
Image credits : http://www.reactiontools.com
Evaluation and Conclusion
This paper has provided evidence that, from the user’s perspective, a Five-Stars Rating System is generally the most preferred, because of a good balance between an intuitive representation, and the ability to provide a wide enough range of responses.
Two groups of users are targeted in this study — users who submit ratings, and the users who view the ratings. Of these two groups, only one group is representative of the target users of the social dining application which is being designed in this Final Year Project.
Pranata et al.’s study has one minor limitation — in the first part of the study, where user opinions of each individual rating system are surveyed, the results do not distinguish between these two user groups, thereby possibly obscuring minute differences between their opinions.
However, in the second part of the study, the two user groups are clearly distinguished, when the most and least preferred rating systems are surveyed, thereby effectively demonstrating that the user group of interest in this project, users who view ratings, do indeed prefer the five-star rating system.
Informal discussions with peers also showed a general preference for the Five-Stars Rating System. Taking into considering the limitations, as well as the sufficient strengths of this study, a design decision was made : Five-Stars win.