Recommendation Systems: The Role of AI in Helping Consumers Make Choices

Written by Deborah Luyando Haamanjanji

Image courtesy of Pexels

Life abounds with many choices, easy and difficult.

Whether it be deciding what AP or IB Diploma classes to take to what time you will begin an assignment, one factor that will constituent the final decision is information.

Information enables individuals to consider possible benefits and consequences to their choice to ensure that the best decision is made. At the end of the day, the decision is up to our judgment. Economic theory implies that AI will influence human judgment on choice by either providing more information or eliminating some factors that will not be of benefit. This blog will discuss how relevant AI- specifically recommendation systems- ensure that consumers and producers make informed decisions in a market.

What is a Recommendation System?

Despite having limited knowledge or experience to consider how a vast array of choices will result in a desirable outcome, recommender systems use information based on a user’s interests or past decisions to recommend an item. There are three types of recommender systems namely; content-based, collaborative filtering- based and Knowledge-based but this blog will only focus on content-based.

How Content-Based Systems Work and Influence Users/Consumers

A Content-Based system recommends items based on a user’s/consumer’s past preferences. By considering item properties such as genre or artist, the system develops a profile that explores the specific attributes of an item that the user previously liked or disliked. The implementation of classification techniques such as Naive Bayes, Nearest Neighbor Algorithms, and decisions trees will then assign each item into the desired category.

The recommender system will then compare an item’s attributes to the user’s profile to derive relevant items that will form a recommendation list that will suit the user’s preference. Finally, the systems will then remove those items the user tends to dislike so the evaluation of the recommendation is dependent on how accurate the item’s representation and the user’s profile are (Zhang et al.).

An example of a commonly used application that applies this is Netflix. On average how long does it take you to choose a movie without looking at the recommendation bar? I will spend 15 minutes before giving up.

With a plethora of series and movies to choose from, users tend to spend so much time trying to find a movie that they believe will be worth their time. With the content-based system excluding items that may not be of interest, a user’s time and utility can be maximized by enabling them to analyze their recommendation to make a better decision without the stress of scrolling through many options.

Content-based recommendations are as beneficial to producers as they are to consumers. With the incentive to maximize profit in a complex customer-driven market, understanding customer needs and wants and how products can align to those desires helps firms decide how they can achieve allocative efficiency.

By implementing a content-based recommender system, firms can analyze the most preferred attributes of a product, enabling them to assess what good will bring them profit. Although, ethical concerns limit the number of information firms can attain. Many users are not willing to put out or display information that they view to be personal, therefore, firms must have the user’s consent or alert them before using their information.

In consideration of ethical boundaries, firms may use the system to analyze consumers’ search results and their( the consumer’s) previous purchases. This approach would enable firms to offer a specific audience what they desire, to increase sales and a potential increase in profit. Examples of firms that take advantage of this are Amazon, Netflix, and Disney+ by releasing or selling products that suit the desired audience’s preference.

Image 1.1: How Netflix uses the recommendation system

Furthermore, to enhance user experience, firms may choose to exploit the buyer’s intent/journey models or purchasing models to ensure that consumers consistently rely on the firm to receive adequate recommendations and information (Payne).

Image 1.2: Buyers Intent/Journey Model.

Image 1.3: Purchasing Model

AI has an immense role in the choices made in a market or society. By considering the importance of making an informed decision, recommendation systems ensure that consumers will experience utility as the choices given to them will suit their preferences. Likewise, profit recommendation systems will enable firms to analyze the market and make decisions that will bring many benefits to consumers.

Citations

Blog Editorial Team. “Artificial Intelligence Getting Prepared for Your Journey,” Software One, 7 Apr. 2021, www.softwareone.com/lb-lu/blog/articles/2021/04/05/implementing-artificial-intelligence-part-1.

Fingent. “6 Ways Artificial Intelligence Is Driving Decision Making | Fingent Blog.” Fingent Technology, 8 Sept. 2019, www.fingent.com/blog/6-ways-artificial-intelligence-is-driving-decision-making/.

Houstonconsumers’ Deng. “Recommender Systems in Practice.” Medium, Towards Data Science, 13 Feb. 2019, towardsdatascience.com/recommender-systems-in-practice-cef9033bb23a.

https://www.facebook.com/krishnaseo. “What Is the Buyer’s Journey and Why It Matters for You.” DigitalNuisance, 15 Oct. 2019, www.digitalnuisance.com/buyers-journey/. Accessed 22 Oct. 2021.

Mishra, Utsav. “What Is a Content-Based Recommendation System in Machine Learning?| Analytics Steps.” Www.analyticssteps.com, 29 May 2021, www.analyticssteps.com/blogs/what-content-based-recommendation-system-machine-learning.

Payne, Matt. “Recommender Systems for Business - a Gentle Introduction | Scalr.ai.” Www.scalr.ai, 24 Aug. 2021, www.scalr.ai/post/recommender-systems-recommendation-systems. Accessed 22 Oct. 2021.

Zhang, Qian, et al. “Artificial Intelligence in Recommender Systems.” Complex & Intelligent Systems, vol. 15, 1 Nov. 2020, 10.1007/s40747-020-00212-w.v