As more and more people go online for shopping, it is becoming increasingly important to understand how they rely on e-commerce recommender systems to make purchases. Researchers at Pennsylvania State University suggest that it is not only the recommended ones, but the recommended methods and reasons that help shape consumer opinion.
In one study, researchers investigated how people responded to the two product recommendation systems. The first system generated recommendations based on the user’s previous purchases. This is often referred to as a content-based recommendation system. The second provided recommendations based on what others purchased. This is called the co-recommendation system.
Researchers reporting their findings in Advertising journalDiscovered that people who prefer to think and solve their own problems (personality types that researchers describe as “highly cognitive”) find content-based recommendations to be more compelling. Did. However, those with low cognitive needs are more persuaded by collaborative recommender systems that can serve as a signal that other buyers are already scrutinizing the product.
S. Shamsundal, a professor of media effects at Donald P. Bellisario College, says that the nature of recommender systems and the degree of confidence in suggesting the right products guide people when making online purchases. It states that it can be very important above. Communication and Co-Director of Media Effects Institute.
“In the pre-internet era, before artificial intelligence, we asked another person at a cocktail party.” I heard you went to Italy. Can you give me some advice? There next month I will go. “Gather information to make our decisions.” Sundal, an affiliate of the Institute for Computational Data Science, Pennsylvania State University, said. “Now we are online and have access to information from almost everyone who went to Italy last month, not just the friends we met at the cocktail party. Now we have information about the collective experience of others. You can now get it, as well as how it matches your own background and previous trips. “
According to Mengqi Liao, a PhD student in the mass media and the lead author of the treatise, the subtle “bandwagon effect” may be persuading people.
“From an amateur’s point of view, we may not know that these are actually two different recommender systems,” says Liao. “In some systems, recommenders may only tell customers that they are based on what they have previously purchased, but co-recommendation systems tell them that many others have purchased this product, It further enhances the compelling appeal. “
Researchers have also found that the effectiveness of the recommended system is related to the type of product the system recommends. Consumers with a high cognitive need are more likely to respond to information about how recommended products reflect their personal tastes when deciding on a movie, travel destination, dining experience, etc. .. This is expressed as a product match rate. Recommended for content-based filtering systems.
However, consumers with low cognitive needs preferred collaborative filtering because they were more persuaded by the percentage of other people who bought the recommended item and also promoted their intent to buy the item.
No such difference was seen in the recommendations for “search products” that can be searched online for information. Both personality types preferred a collaborative recommendation system.
“It can be thought of as a kind of cognitive outsourcing,” says Sundar. “For example, a customer might look at a smartwatch ad and see a feature, but” I’m not going to do the hard work of going through all the details and deciding which one is better. ” Just outsource this to others. “If they say it’s a good smartwatch, they’ll buy it. “
According to Liao, most research on recommender systems focuses on optimizing the proposals for these systems. These findings require developers to consider other factors, such as personality type and product type, to improve the user experience of the system, rather than focusing solely on the accuracy of algorithm proposals. It suggests that.
“It can be very dependent on how users receive information about the recommendations that the system provides,” says Liao. “It’s important why these systems provide product and experience recommendations.”
Researchers recruited 469 people from online crowdsourced microtasking sites for research and randomly assigned them to experimental websites that use collaborative or content filtering algorithms.
For collaboration systems, researchers use percentage ranges to indicate the number of similar people (or percentage matches) who have used the recommended product and serve as a clue to the bandwagon effect. For content-based systems, the same percentage numbers were used to show how well the recommended products match the personal characteristics of the consumer based on the user profile. There were three levels of concordance indicators: low, medium, and high.
When testing two different types of products, search and experience, researchers use smartwatch recommendations as examples of search products and tourist destination recommendations to help participants experience the product. I investigated the reaction.
Before browsing the e-commerce site, all participants answered a series of questions to determine if cognitive needs were high or low.
Researchers have tested only two products and two general recommendation systems, so future research may look at the psychological effects of other systems and explore other types of products. Researchers have said this may help validate their findings.
One study found gender bias in music recommendation algorithms
When Mengqi Liao et al, an e-commerce personalization system shows and conveys: A survey of the relative compelling appeal of content-based filtering and collaborative filtering, Advertising journal (2021). DOI: 10.1080 / 00913367.2021.1887013
Courtesy of Pennsylvania State University
Quote: Consumers make decisions based on how and why the product is recommended online (2021, May 4th). Obtained from https: //phys.org/news/2021-05-consumers-decisions-based-products-online.html on May 4, 2021.
This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.
Consumers make decisions based on how and why the product is recommended online
Source link Consumers make decisions based on how and why the product is recommended online