The Impact of Customer Perception on Purchasing Decisions in E-commerce
The growth of e-commerce has changed the way customers make purchasing decisions. With an abundance of products available, customers rely on perceptions to make quick judgments between options (Du & MacDonald Reference Du and MacDonald2016).
These perceptions are derived from prior experiences and available information, acting as mental shortcuts for customers to simplify decision making (MacDonald & She Reference MacDonald and She2015). For example, customers tend to judge how absorbent paper towels are based on the presence of quilted lines (MacDonald, Gonzalez & Papalambros Reference MacDonald, Gonzalez and Papalambros2009). While relying on perceptions can help customers simplify decisions, it can also mislead customers to make uninformed decisions (MacDonald & She Reference MacDonald and She2015). This is often seen with sustainable products where features perceived as sustainable may not contribute to engineered sustainability.
In this article, a feature is defined as either a visual aspect of a product such as material or as a descriptive aspect like ‘a nice gift’. Moreover, engineered sustainability is defined as real sustainability according to well-studied methodologies such as a life cycle analysis (LCA). For example, customers may perceive a stainless-steel coffee maker as more sustainable than a plastic one, but according to an LCA, it is the energy efficiency that has the largest environmental impact.
Despite this, designers tend to focus on engineered sustainability requirements while neglecting perceived sustainability (MacDonald & She Reference MacDonald and She2015). This is validated by a lack of market success for sustainable products despite market research indicating customers are willing to pay more for them (The Sustainability Imperative: New Insights on Consumer Expectations 2015). A robust literature exists on customer perceptions of features when purchasing products.
Previous literature supported designing-in features based on perceptions, but a method was lacking for identifying features as perceived by the customer. Literature has shown online reviews to be a treasure trove of customer perceptions of design features. For example, Ren, Burnap & Papalambros (Reference Ren, Burnap and Papalambros2013) used Amazon Mechanical Turk (MTurk) respondents to assess the perceived safety of car designs using online reviews and machine learning. Building on this, we previously developed a method to identify features perceived as sustainable from online reviews using crowdsourced annotations and natural language processing (El Dehaibi, Goodman & MacDonald Reference El Dehaibi, Goodman and MacDonald2019) (refer to section ‘Combining machine learning with collage approaches’ for a deeper overview). We extracted features perceived as sustainable using French presses as a case study and demonstrated they are not fully aligned with engineered sustainability.
In a subsequent study, we confirmed that participants identified the extracted French press features as sustainable using a novel collage activity (El Dehaibi, Liao & MacDonald Reference El Dehaibi, Liao and MacDonald2021). Participants placed products on a set of axes and selected features from a list. We found that they more often selected features perceived as sustainable when evaluating product sustainability on the collage.
The benefit of our approach is that we rely on customer reviews to determine customer perceptions. Our goal is to provide designers a robust method to identify product feature perceptions from online reviews so that they may differentiate their products and drive purchase decisions.
As more purchases occur online, several papers have explored the changing context in which customers form perceptions, using tools like machine learning and collage activities to extract perceptions from online reviews. Moreover, an active body of research has leveraged machine learning to uncover customer perceptions from online reviews. The gap in this literature lies in understanding how actual features, such as ‘handle shape’, would influence customer perceptions. This research aims to fill this gap by identifying generalizable methods that link specific visual or descriptive product features to customer perceptions.
In this section, a literature review is presented on how customer perceptions shape online decision making. Wang et al. (Reference Wang, Li, Ye and Law2016) investigated the impact of online reviews embedded in product descriptions on purchasing decisions. The scholars investigated how the participants interacted with pages that had and did not have online reviews in the descriptions. The results showed that product pages with online reviews in descriptions had longer fixation time on average, suggesting these descriptions aligned closer with customer perceptions.
To determine the influence of purchase decisions, the scholars then collected historical data from Taobao for two products, a shaving gel, and an electric shaver. The data included sales, reputation, price, and whether the products had online reviews embedded in the descriptions. Using a hierarchical multiple regression model, the scholars found that descriptions embedded in online reviews positively influenced purchase decisions.
Maslowska, Malthouse & Viswanathan (Reference Maslowska, Malthouse and Viswanathan2017) studied the influence of product price and customer perceptions of reviews on online purchase decisions. The scholars used shopping data provided by two online retailers, one that sells unique and high-priced items while the other sells health and beauty products. There were 2.5-3 million observations from each retailer. For each observation the scholars had access to the number of reviews for a product, the average number of stars, whether the customer clicked on the ‘review tab’, product price, and purchase decision.
The scholars used a logistic regression model with the purchase decision as a dependent variable and found that the product price plays an important role in how ratings and reviews influence the purchase decision. For lower-priced products, average ratings can have a large influence with fewer reviews while for higher-priced products, more reviews are needed for the average rating to have an influence.
von Helversen et al. (Reference von Helversen, Abramczuk, Kopeć and Nielek2018) investigated the relationship between customer age and the influence of perceptions of product attributes and reviews on purchasing decisions. The scholars designed three between-participant conjoint analysis surveys where they presented pairs of positively rated household products to participants. A mixture of highly positive and negative reviews was shown with a mixture of low and high ratings. The scholars found that younger customers relied more on average ratings when product attributes were similar between paired choices, while older customers were quickly influenced by negative reviews.
Li et al. (Reference Li, Zhuang, Lu and Chen2019) investigated how return policies influence customer perceptions of products and decision making depending on the market stage of a business. The scholars propose a multistage hidden Markov model which models randomly changing systems. They test it on 50,000 purchase records spanning three years from Taobao including returns, discounts and total sales. The results showed that promotions and return policies had a varying influence on repurchase behaviour across different stages of market growth. For example, a company in the growth stage could benefit from flexible return policies and frequent promotion while a company in the introduction stage would not.
A gap in previous literature is understanding how actual features, such as ‘handle shape’, would influence customer perceptions. This presents an opportunity for designers to determine how specific visual and descriptive product features align with customer perceptions to drive purchasing decisions. The development of e-commerce and social media provides a wealth of information that designers can tap into online.
Zhang, Yan & Zhang (Reference Zhang, Yan and Zhang2020) use Airbnb, an online marketplace for accommodations, to study the influence of accommodation host self-descriptions on customer trust and booking behaviours. The scholars annotated 4179 host descriptions from Airbnb listings based on the perceived trustworthiness of the hosts. The scholars then used a deep learning model to predict perceived host trustworthiness for 75,000 host descriptions.
Using this data, they extracted textual features including readability, sentiment intensity, and semantic content. Semantic content included personal information about the host such as family and work. From regression analyses, the scholars showed that readability of the self-description had a positive influence on perceived trust, while semantic intensity had a U-shaped relationship with trust. Moreover, semantic content had a positive influence on trust if the content was related to sociability. When looking at Airbnb booking decisions, the results showed that higher perceived trust of hosts led to more booking decisions.
Liu, Jiang & Zhao (Reference Liu, Jiang and Zhao2019) use natural language processing to identify product competitive advantages from social media content. An example review could include: ‘the sound system in the Passat sounds better than the one in my old Camry’. The scholars first preprocessed the reviews by removing stop words and performing named-entity recognition. They then performed a sentiment analysis using logistic regression and a domain-specific lexicon to assess customer sentiments towards features of the Volkswagen Passat compared to its competitors.
Zhou et al. (Reference Zhou, Ayoub, Xu and Yang2020) developed a method to extract relevant product features from online reviews. The scholars collected 91,738 review sentences across several products created by Amazon (Fire tablet, Echo, etc.). The scholars manually labelled a sample of 10,000 reviews as either relevant or not to the product and trained a fastText algorithm to filter out irrelevant reviews; about 20% of the reviews were filtered out to remove noise. The scholars then used a topic modelling approach called Latent Dirichlet Allocation (LDA) to identify ‘topics’ that in this case are product features from the remaining review sentences.
Park & Kim (Reference Park and Kim2020) proposed a method to improve the accuracy and diversity of extracted features from online reviews using keyword embedding and two clustering methods. The scholars tested their method with 27,201 laptop reviews, 20,823 wearable device reviews and 19,159 smartphone reviews, and also collected 27,201 reviews for 61 laptops. The reviews were preprocessed and embedded into vectors, and then clustered using X-clustering on noun phrases for noise reduction. Special clustering was then applied to extract features.
Kim, Park & Kim (Reference Kim, Park and Kim2021) build on the above method by applying it to investigate the role of COVID-19 pandemic in changing customer preferences. The scholars collected 8548 reviews for smartphones dating before the pandemic and 7263 reviews for smartphones dating after the pandemic began. Smartphones included both new and refurbished products. The reviews were preprocessed, embedded into vectors, and clustered to identify key product features: screen, memory, camera, battery, security and price.
The scholars then used a sentiment library to calculate the sentiment for each review sentence according to the clustered product features. The results showed that for new phones, sentiments decreased across many of the product feature clusters, while for refurbished phones that difference was less significant. Similarly, Bag, Tiwari & Chan (Reference Bag, Tiwari and Chan2019) develop a method to predict a customer’s purchase intention based on review polarity and sentiment scores. The scholars used natural language processing to extract features from 29,069 online reviews.
Then linear and nonlinear regression analysis was performed including both review polarity and brand social perception scores to extract salient product features. Previous literature identified methods to extract customer perceptions from online content but has not yet identified methods that can link specific visual or descriptive product features to customer perceptions. Moreover, previous literature has not developed a method to validate perceptions of product features by testing those perceptions on users in terms of liking and evaluating products.
This gap is particularly crucial for sustainable products where designers often focus on engineered requirements while neglecting perceived requirements. Motivated by this gap, we previously conducted two studies where we first developed a natural language processing approach to extract product features perceived as sustainable from online revi... Selling products and gaining profit is the main mission of a private Sales forces with their specialized relationships can determine the fate of transactions and make companies successful in achieving their ideal goals.
Buyer-seller relationships can significantly reduce the marketing costs of the organization. Therefore, the impact of sales forces in an organization is relatively high in the past. A better understanding of the sales professional dimension will lead to the formation of long-term relationships and increase a company's profitability. However, although much research has been done on the issue of company-customer relationships, there have been extensive studies on the perception of consumer behavior by vendors.
Therefore, the aim of this study was to investigate the effect of trust, reputation, and reputation of the seller on the commitment to the seller with the mediating roles of good interaction and perceived risk. The statistical population of this study was 250 customers of a retail company.
In order to collect the required data, an electronic questionnaire was designed and distributed among the statistical reporters using the available sampling method.
Purpose: People enjoy digital shopping because it allows them to make purchasing decisions anytime and wherever they choose. The goal of this study is to discover the social media variables that impact customer purchasing decisions as well as to examine consumer perceptions of the items, and determine how social media impacting factors directly affect customer purchase decisions of fashion products.
The study relies on secondary and collected sources. This is based on a review of the existing literature, with data obtained from various exploratory articles or papers, journals, books, and magazines. Science Direct, Research Gate, and Google Scholar were used to gather the data required for the study.
This result demonstrates that the concept of consumer perception - usually it is linked to a specific behavior such as emotional and cognitive because of these commitments the virtual market is becoming increasingly competitive.
MacDonald et al. (Reference MacDonald, Gonzalez and Papalambros2009) identified a relationship between perceived and engineered requirements using a discrete choice analysis survey with paper towels. The results showed that customers constructed their perceptions on an as-needed basis and are not inherently found in people. For example, in a paper towel survey participants claimed they would not purchase nonrecycled paper towels for any price, but later reported purchasing from brands with 0% recycled paper towels the last time they went shopping. In a subsequent study, She & MacDonald (Reference She and MacDonald2017) demonstrated how perceived sustainable features led participants to prioritise engineered sustainability concerns in a decision scenario with toasters.
For example, an embossed leaf pattern on a toaster led participants to prioritise energy and shipping concerns of the product. While the embossed leaf pattern does not contribute to sustainability, it communicates information to customers that helps bridge the gap between perceived and engineered sustainability. In doing so, customers are better informed to align their intent with their purchase decisions. The previous literature highlights why it is important for designers to design-in perceptions, and for designers to meet customers where they are. Specifically for sustainability, it is important for designers to create a product that is both engineered to be sustainable and also perceived as sustainable by the customer.