Beyond Accessibility Compliance: Exploring the Role of Information on Apparel Shopping Websites for the Blind and Visually Impaired

Abstract

In response to increasing numbers of people switching from offline to online shopping after the COVID-19 pandemic, this study sought to obtain an in-depth understanding of apparel website design and digital accessibility for all people, including people with visual impairments (PVI). Given the Convention on the Rights of Persons with Disabilities Article 9, which mandates that all 191 international parties take measures to ensure people with disabilities have equal access to information and communication technology, this study explored the challenges PVI face while accessing informational content about apparel products online. To achieve this goal, Culnan’s dimensions of perceived accessibility to information, a lens for understanding how consumers experience and evaluate the accessibility of information systems, were used as the theoretical framework. We applied phenomenological methods to explore the daily “lived experience” in depth through observations and semi-structured interviews with eight female participants in their 20 s to 60 s, each lasting more than 45 min. Based on thematic analysis, the findings highlighted the unmet website meta descriptions for product information and navigation functionality for assistive technology, which, as a result, negatively impacts digital accessibility for PVI to shop online for apparel. The study concludes with contributions that extend the theoretical framework to the digital landscape, addresses the gap of inclusive digital apparel retailing practices, and emphasizes the opportunities for apparel educators to incorporate an inclusive design curriculum.

1. Introduction

Increasing numbers of people switching from offline to online shopping after the COVID-19 pandemic highlights the need for a critical review of apparel shopping website design to ensure that digital accessibility exists for all people, including people with visual impairments (PVI). In 2023, the global fashion e-commerce market amounted to 781.5 billion US dollars, with China and the US accounting for 309 billion US dollars in revenue—emphasizing the prevalence of consumers’ preferred apparel shopping channel [1]. People with “visual impairments” refer to those with low vision or blindness, who are categorized by the level of visual acuity and visual field that affect their functional abilities [2]. PVI require an alternative strategy to access information on apparel shopping websites through assistive technology (AT), which provides a human interface to the website’s alternative text in various modalities, including screen readers and zoom text. In 1988, the US Congress enacted the Technology-Related Assistance Act for Individuals with Disabilities [3] and defined AT as a device, equipment, or system used to increase, maintain, or improve the functional capabilities of a person with a disability. Meanwhile, the United Nations [4] mandates that all 191 international members (e.g., states, countries, and the European Union) take measures to ensure people with disabilities have equal access to information and communication technology. Therefore, the study aimed to explore the ability of PVI to equally access information about apparel products online—allowing them to fully participate in society—and highlight the lack of inclusive design within apparel retailing practices.

The World Health Organization’s [5] International Classification of Diseases (ICD-10) is utilized as the global standard to define visual impairment into two groups: (a) low vision; (b) blindness. Individuals with low vision have a visual acuity range between 6/18 and 3/60. Visual acuity of 3/60 would mean an individual would need to stand three feet away from an object to see it with the same clarity as a normally sighted person standing 60 feet from the same object [2]. Individuals with blindness have a visual acuity below 3/60, and those with a visual field (peripheral vision) of less than 10 degrees are classified as having total blindness. It means that, of the 285 million PVI globally, 39 million are legally blind, and 246 million people have low vision. Of those who are legally blind, about 15% are classified by ICD-10 [5] as having total blindness or no light perception (NLP), and approximately 34 million people have some light perception and can see varying color hues, shades, and outlines of objects along with 246 million people with low vision. PVI will be the term used to identify the population with either “low vision” or “blindness” throughout the study.

Accessibility refers to the design and implementation of products, services, environments, and content in a way that ensures they can be easily and effectively used by people with various abilities and disabilities [6]. It is not limited to physical space but also extended to digital spaces; it is the fundamental agreement of over 191 international members of the United Nations, which spearheads international social and legal frameworks to provide effective and nondiscriminatory communication through digital accessibility centering on access to products, resources, and services across hardware and software in digital spaces [7]. The United Nations [4] encourages digital accessibility through the use of the World Wide Web Consortium’s (W3C) Web Content Accessibility Guidelines (WCAG) 2.1 [8], which defines how to make web content more accessible to people with disabilities, including the visually impaired. For example, inaccessible web content means that people with disabilities are denied equal access to information. The US Department of Justice [9] stated that ensuring web accessibility for people with disabilities is a priority. The United Nations [10] also declared access to published digital content as a basic human right.

However, many apparel shopping websites are not compliant with WCAG 2.1 [8] and limit PVI’s ability to access apparel. We have little understanding of how current apparel shopping websites communicate their products to all people, especially to PVI, and if their information is digitally accessible. If not, what are the issues that PVI can experience? The finding of this study was expected to close the gap in our understanding of digital accessibility within the apparel shopping context and provide insights into how the industry can enhance accessibility in the digital landscape through communication between AT and consistent website design. To achieve this goal, Culnan’s [11] dimensions of perceived accessibility to information are used as the theoretical framework, which provides a nuanced lens for understanding how consumers experience and evaluate the accessibility to information systems through physical, interface, and informational dimensions. To achieve this, a literature review highlights blindness as a spectrum and its challenges toward independence, the emergence of digital accessibility and AT functionality from the PVI perspective, and Culnan’s [11] framework adapted for the digital landscape. Given the lack of existing literature, a qualitative approach to data collection and analysis was applied through observation and in-depth interview methods with visually impaired consumers using their own combination of AT in an online apparel shopping setting. The study concludes with implications and future research opportunities.

1.1. Blindness Spectrum and Challenges for Independence

Total blindness, low vision, and visual impairment encompass a diverse range of conditions, each with its own unique characteristics and implications for daily life [11]. A common misconception about “blindness” is the person is experiencing a world of total blackness [12]. However, this is far from accurate; a variety of eye diseases, genetic disorders, and birth defects, as well as aging or suffering an injury, can interfere with healthy vision [13]. These visual impairments do not all “look” the same. People with total blindness are those who have a complete lack of light perception and are legally described as having no light perception (NLP). In fact, only about 15% of the PVI have total blindness; the remaining 85% have some level of vision where they may be able to see full-color hues, shades, or intensities along with object outlines [13]. They may also have a blind or blurry spot in the middle of their field of vision or peripheral vision, yet can still decipher images and colors. However, they may have one visual impairment or a unique combination of visual impairments. Therefore, blindness is a spectrum of visual impairments that affects 285 million people worldwide [13].

Over the last two decades, society has become more aware of the barriers facing PVI. Some researchers elucidated the various challenges in offline and online environments to obtain informational resources or navigation information, such as visual impairment advocacy programs or access to government resources [14,15]. However, only a few studies are available on PVI’s challenges for obtaining information while shopping. Lee et al. [16] explored the offline shopping environment, where PVI reported that the purchasing process is an inconvenience and that it is difficult to obtain information about products. Zor and Vuruskan [17] argued that, for individuals with low vision to total blindness, daily activities such as mobility, communication, and shopping may be challenging. They further explored the offline shopping environment with a smart glove concept and found the apparel shopping experience is a rather long and difficult process for PVI, specifically for locating a store, choosing a product (identifying color, silhouette, design style, etc.), combining garments from closet memory, purchasing, and caring for the product according to the instructions. In their study, they used physical shopping environments and focused on the inability to access product information through tactile senses, not the online shopping environment.

Although their study is one of a few dedicated to apparel shopping, Teixeira et al.’s [18] research involved both online and offline shopping, and the study context was very general without any specific bounds. They interviewed PVI about past apparel shopping experiences (both online and offline) and found that PVI were confronted with descriptions that were not detailed and did not accurately represent the actual products. Hence, the understanding of specific challenges by PVI who navigate apparel online shopping websites is very limited. In sum, the research on PVI’s apparel online shopping challenges is extremely limited, with little or no understanding of their daily challenges despite the importance of apparel in all our lives.

These challenges and lack of adequate, accessible apparel shopping for PVI result in barriers to social participation, contributing to the disablement process [19]. Moreover, when choosing apparel, the consumer develops certain expectations about the ability of the product to fulfill functional needs as well as symbolic needs [20]. Previous studies focus on fulfilling the functional needs of PVI with little attention to the symbolic needs of independence, expressionism, and dress as communication [21]. This limits PVI to participate in all aspects of life and supports the apparel retailers’ role in the disablement process. A guiding theme among the blind and visually impaired community is to promote independent and healthy living skills through the development of vision-specific skills along with the use and application of technology [22]. In fact, Lee et al. [23] argued that the notion of “independence” is frequently used to motivate technology design in the human–computer interaction sub-field of accessible and assistive technology for people with disabilities. For example, Yang [24] studied mobile application design for better clothing habits of PVI and found that enabling them to choose what to purchase without assistance is important to the quality of their apparel shopping experience. Other studies found that the ability of PVI to obtain product information is heavily reliant on store associates, family members, and friends, which denies them a chance to independently search for apparel, and they often feel like a burden to those around them [21,25]. Despite the challenges, PVI reported that online shopping can give them a sense of control and is generally their preferred mode of shopping for clothes compared to the offline equivalent when it works well [26].

1.2. Digital Accessibility

Although the web remains a key digital environment for people with disabilities, opportunities and barriers evolve through technological generations—resulting in periodic cycles of gain and loss of accessibility. Digital accessibility is measured by a chain of dependencies where training, hardware, software, content, and standards must work together harmoniously, and each of these elements must be understood as a dynamic process [27]. Sánchez-Gordón and Moreno [28] state there are many operating systems, applications, and websites that have addressed their inaccessible problems, only to become inaccessible again when an update, new version, or campaign is launched. The common pattern is that minor accessibility corrections are made after software production rather than during the software design itself, where key criteria like usability and security are considered, yet accessibility is excluded [28]. Because of that, Grieves and Kaneko [29] argue that accessibility must be considered qualitatively for every aspect of the design and production process; then, it can improve the user experience for everyone.

In an effort toward independence, many researchers addressed accessibility through the creation of intricate and costly AT devices to shop offline. These include a wearable, portable, and cloud computing device for recognizing the images of products for PVI [16]. Other initiatives like mobile applications such as Be My AI™, Aira©, Seeing AI, and Google Lookout were created with the offline experience in mind. The field has witnessed fewer research projects aimed at bridging the accessibility gap for online shopping. The proposed technology has many limitations such as annual subscriptions, additional manufactured devices, and dependence on unaccompanied personal shopping assistants. Programs like Beyes^® and BrowseWithMe™ incorporate a computer screen reader, text-to-speech output, and artificial intelligence to improve the online apparel shopping experience for PVI. Specifically, BrowseWithMe™ is the first automated technology that enables PVI to actively ask for and receive only the information they are seeking on a webpage [21]. However, their study observed that inconsistent structure and incomplete product information among websites, uncompliant to WCAG 2.1 [8], is a limitation to its proposed technology.

AT is a technology designed for individuals with some form of impairment or the elderly population and enables them to overcome various physical, social, infrastructural, and accessibility barriers to independence [30]. To obtain information online, PVI utilize ATs that have the following functions: screen readers, text-to-speech software, text magnification, zoom tools, braille displays, refreshable braille interfaces, accessible web design features, voice control, screen curtain software (blocks out visual information on the screen), etc. The research surrounding accessibility and inclusion for PVI centers around developing ATs with human-machine interfaces [30] (Bhowmick and Hazarika, 2017), like Jobs Access with Speech^® (JAWS^®) and Fusion™, which were developed to prevent those with vision loss from seeing screen content or navigating with a mouse [31]. Although both require an annual license, JAWS^® is the most widely used screen reader, and Fusion™ is the ultimate accessibility tool for PVI to access information online [31]. PVI may use one specific AT or a combination to fulfill their needs; however, Fusion™ was designed to supplement JAWS^® for those with low vision. JAWS^® utilizes tab command to navigate through webpages and synthesize content. However, AT navigates through webpages in a linear approach (e.g., a website structure where users progress in a single, straight-line path). This is much different than the approach for the sighted, who operate in a non-linear navigation path using a cursor or a touch screen. Linear navigation, when executed with accessibility for PVI in mind, can help users predict what element of the webpage will come next. Yet, most webpages are designed with the sighted in mind with various visual cues that motivate non-linear navigation, creating inefficiency and limited exploration for PVI. AT performs best when website design is balanced between providing a clear, consistent path and offering flexibility for PVI to access information in a way that suits their needs.

1.3. Theoretical Framework

To gain a holistic understanding of the needs of PVI and the challenges they face while shopping online for apparel, we explored the communication between website design and widely used AT through the lens of Culnan’s Dimension of Perceived Accessibility to Information [11]. The theory explains perceived accessibility as a multidimensional concept encompassing physical (the extent to which an individual has physical access to the terminal and the system), interface (the extent of translating a request into a non-natural language), and informational (the ability to retrieve the information independent of any subsequent judgment as to the item’s relevance) access to the source which helps design and evaluate a wide variety of information systems and services (see

Table 1. Culnan’s framework and digital accessibility adaptation.

Dimensions	Perceived Accessibility to Information 1	Digital Accessibility Adaptation 2
Physical	Gaining physical access to a terminal and system.	Gaining physical access to a personal computer and an apparel website through assistive technology.
Interface	The extent of translating a request into a non-natural language.	The use of assistive technology features to command the language to formulate a website query and navigate between webpages.
Informational	The ability to retrieve the information independent of any subsequent judgment as to the item’s relevance.	The use of assistive technology to retrieve desired information based upon command language of website query and subsequent judgment to the output’s relevance.

¹ Reference to Culnan’s dimensions of perceived accessibility framework. ² Reference to author’s digital accessibility adaptation for study context.

).

The framework has been widely used in communication technology research in the offline environment, including organizational communication, library science, and management information systems. Culnan’s [11] study explored three sources of information (computer-based, libraries, and individuals) and found the three dimensions significantly influenced peoples’ perception of accessibility of a system or source. Within the online shopping context, the informational dimension is a focus. Indeed, Culnan argued that a computer-based information source requires a more complex structure for command languages than information sources, which are derived from in-person exchanges. In the field of information technology (IT) services, the challenge for designers and researchers is to develop systems that not only meet users’ needs for information but are also perceived as accessible. In this light, Djamasbi et al. [32] explored web accessibility for visually impaired users through the perceived accessibility dimensions. This study found the informational dimension to be a key determinant of ease of use and usefulness by both users with and without visual impairments. The finding’s applicability to users with and without visual impairments is important to note as web accessibility continues to evolve and influence all users’ access to information.

In sum, given the apparel online shopping context in this study, we focused on the informational dimension of digital accessibility faced by PVI. That is, we focused on digital accessibility within the context of information exchange between AT and apparel shopping website design. Following Culnan’s [11] informational dimension with the concept of digital accessibility, we aimed to assess how PVI retrieve the proper product information they need to purchase apparel and how website content must be communicated through the limited functions of AT to add to the multi-disciplinary academic literature (e.g., apparel retailing, human–computer interaction, and digital accessibility). Specifically, the study raised two research questions: (a) how does online apparel shopping website content exchange the product information with PVI with the use of AT? (b) how does such website content exchange navigation information with PVI while using AT? To achieve the study aims, Section 2 highlights the qualitative approach to data collection and analysis applied through observation and in-depth interview methods. Section 3 presents the results of overarching themes and supporting subthemes. Section 4 concludes the paper with a discussion about implications and future research opportunities.

2. Materials and Methods

2.1. Qualitative Inquiry

We used a phenomenological research design to explore the participants’ lived experiences and depict the essence or basic structure of the phenomenon. A phenomenological approach is well suited to studying emotional, affective, and intense human experiences [33]. Phenomenologists are interested in the “lived experience”, and such a focus requires researchers to go directly toward the phenomenon itself [33]. To obtain the essence of PVI’s daily experience of apparel shopping online, task observations, a demographic questionnaire, and in-depth interviews were employed. In qualitative research, interviews are the primary method for data collection. However, Merriam and Tisdell [33] argued that when searching for the essence of the phenomenon, observations are another primary tool because they take place in the natural setting where the phenomenon occurs and offers a firsthand encounter. In the phenomenological process, prior to observing and interviewing those who have had a direct experience with the phenomenon, the researchers examine and set their own personal viewpoints and assumptions aside [34].

To address the tension between a phenomenological approach and the use of a pre-defined framework, we employed Culnan’s [11] framework as a set of sensitizing concepts [33] to guide data collection and analysis. While developing semi-structured interview questions, we were mindful of Culnan’s [11] two relevant dimensions (informational and interface) by incorporating open-ended questions that allowed participants to discuss their experiences in their own words. For example, instead of directly asking about the complexity of communication between website design and assistive technology to obtain apparel product information or navigation, we asked participants to talk aloud through the observation, which allowed for further probing. During data analysis, we initially coded and identified emergent themes within each participant’s narrative. Subsequently, we revisited these themes within and across all participants’ narratives through the lens of Culnan’s [11] framework, allowing us to connect our findings to relevant academic literature, therefore directly addressing the gap that remains for PVI to equally access apparel retailing websites. This approach allowed for a rich, nuanced understanding of the phenomenon while also providing a structure for identifying specific accessibility barriers with actionable suggestions. The Section 3 illustrates the overarching themes and their connection to Culnan’s [11] informational and interface dimensions.

2.2. Data Collection and Analysis

Upon approval of IRB (#2099807), two selection methods were used. First, five participants were recruited via purposive sampling from the researcher’s personal connection to the blind and visually impaired community at a non-profit rehabilitation center for PVI. Second, three participants were approached through snowball sampling by previously selected participants while at the rehabilitation center. In total, eight individuals who had experience shopping online for apparel in the past year and were legally blind or visually impaired participated in the study (See

Table 2. Demographic profiles of participants.

Participant 1	Age	Vision 2 and Perception Level 3	Online Shopping Frequency	Assistive Technology	Assistive Technology Experience Level
Dani	53	Blind, SLP	Often	Magnifier, colored keyboard, reading lamp	Knowledgeable
Eva	42	Blind, NLP	Often	JAWS®, braille keyboard	Expert
Suzy	50	Low vision, SLP	Often	Zoom text, magnifier	Learning
Nora	51	Blind, SLP	Often	JAWS®, Zoom text, magnifier	Knowledgeable
Piper	69	Low vision, SLP	New	Zoom text, magnifier	Learning
Rita	39	Blind, SLP	Often	JAWS®	Knowledgeable
Mia	32	Blind, SLP	Often	Fusion	Knowledgeable
Kate	28	Blind, SLP	Often	JAWS®	Expert

¹ Reference to each participant is indicated by a pseudonym. ² Reference to each participant is indicated through ICD-10’s categories of low vision or blindness. ³ Reference to each participant is indicated by some light perception (SLP) or no light perception (NLP).

for demographic information of the study participants). Although eight participants may seem a relatively small number, a review of the transcribed interview transcripts, observation notes, and researchers’ analytical memos showed recycling of the emergent ideas and a lack of new information from participants, suggesting theme saturation [33]. The research location site was chosen to offer a natural setting for participants who visit the rehabilitation center to learn independent living skills through the development of vision-specific skills through technology. Therefore, it was believed that participants were knowledgeable about shopping online for apparel and experienced AT users.

We conducted one-on-one observations and interviews with eight female participants in their 20 s to 60 s, each lasting more than 45 min. Each observation and interview were audio- and video-recorded via Zoom with participant consent and then transcribed for data analysis. The task observations were conducted by one of the researchers at the location site and took place on three consecutive days in February of 2024. Participants were tasked to browse Amazon Fashion with their chosen AT for an apparel item and compatible accessory, share their screens, and select their items. Amazon was chosen due to its prevalence in society, notable accessibility standards, and user-driven technological advancements [35]. The researcher and participant were isolated in a room together; however, participants used Zoom’s screen sharing function so the researcher could follow their online apparel shopping pathways from a computer that was not using AT. The video-recorded screen sharing captured visual images for data analysis purposes. During task observations, the researcher bracketed personal experience as a sighted, online apparel shopper and utilized silenced probing.

Interviews followed each observation and lasted up to 25 min. To build rapport with participants with a disability, the researcher spent about ten minutes sharing their personal and research background prior to beginning the data collection. Second, a questionnaire was used to collect personal and demographic information such as age, gender, vision and perception level, frequency of online shopping, AT, and user experience. Next, semi-structured interviews included questions guided by Culnan’s informational and interface dimensions about communication between AT and website content, which allowed for participants to freely talk about their experience in their own words. Semi-structured interview questions included the following examples: “How would you describe the process of gathering detailed information about the apparel product?”, “What features of your chosen AT are most useful?”, “What could be changed or added to improve your online shopping experience?”. Participants’ responses were further probed to obtain a deeper and clearer understanding of the meaning of the specific experience and yield substantive information about the topic in the participants’ own narrative [33].

The 235 pages of interview transcripts, seven hours of video-recorded participant screen sharing, and 10 pages of field notes were interpreted in four cycles of analysis to uncover the essence of the phenomenon, thematically and holistically [36]. First, the finest details were coded within each interview transcript and video recording, then moved upward to general observations. Second, the emergence of similarities and differences among all within and across the study informants were further categorized by related codes [37]. Third, through the horizontalization of the newly developed understandings across all informants, codes were relabeled or reclassified to consolidate the essence of the phenomenon. Lastly, two researchers reached a unanimous agreement regarding data codes for two overarching theme categories aligned with Culnan’s [1] informational dimension and supported by six evolved sub-themes.

3. Results

The final holistic interpretation led to the creation of a conceptual model that illustrates the relationships among the participants’ experiences, individual themes, and the overarching theme categories that emerged from the study data (See Figure 1). Interpretation of the interview data, demographic profiles, and observations indicated a complex range of challenges for the participants to successfully shop online. The thematic data analysis generated six sub-themes under the two overarching theme categories of product and navigation information. The first overarching theme highlights barriers faced by PVI to equally access apparel product information online—extending Culnan’s informational dimension to the digital landscape. The second overarching theme illuminates the inability of PVI to equitably navigate within webpages—emphasizing the need to incorporate Culnan’s interface dimension across apparel and human–computer interaction disciplines.

3.1. Difficulties in Decoding Product Information (Culnan’s [11] Informational Dimension)

Theme one, difficulties in decoding product information, revealed three sub-themes: (a) color names that “make no sense”; (b) the power of words in product titles; (c) time drain, size maze.

3.1.1. Color Names That “Make No Sense.”

During task observation and interviews, seven out of the eight participants shared difficulty in obtaining information about color options when selecting a product. Specifically, two key barriers of color emerged: (a) improper or vague lexical names (color names make no sense); (b) numerical names (e.g., black-150). First, although naming colors is the process of assigning lexical terms to reference objects in the world through linguistic representation [38], the meta descriptions for product color names read by AT seemed to cause serious confusion. Eva, who is legally blind and sees total black, considers herself an avid online shopper and an expert user of AT, stated that, in general, “finding information about the product is difficult because they have these weird names for colors”. Dani, who is legally blind with some light perception and can see shades of colors in black and gray, echoed Eva and expressed her extreme frustrations by stating, “colors are hard for me…because when they say certain colors, what does it mean?”. Toward the end of Eva’s shopping experience, after about 30 min, she expressed her frustration and stated, “some colors don’t even make sense”.

From the perspective of the participants, color-hue identification is important when selecting and assessing apparel products. Kate, who is just like Eva in terms of total blindness and has extensive online shopping experiences, shared that when searching for a product, “first, I like to look at the colors”. In agreeance Eva, she shared, “color is what I’m thinking of first”. Here, the first author, as a sighted person who is interviewing the PVI, was surprised how colors are important to those who have never actually seen the colors the way the sighted people do. Piper, who has low vision and can see some colors, stated, “if the color selection was good, I might buy a few pairs”. However, more often than not, the participants faced various lexical color names that did not help in their understanding. Some of the adjectives used in color names, in fact, provided little to no context for them, creating confusion. For example, Piper viewed a pair of pants and expressed, “It’s in 8 colors. Nice!”. However, as she reviewed the color options, she began to ask questions, “ash heather, that tells me it might be gray-ish?”. Piper also read aloud, “livia?”, “rodeo?” and “Niagara?”, then proceeded to leave the product landing page, suggesting she thought these color names did not mean anything to her. Similarly, Eva shared, “please tell me what the difference between black and black rinse is” and “I have no idea what bungalow khaki is”. Ultimately, the participants did not select the items due to the inability to determine color hue identification. Although color identification should reference specific objects in the world, terms like “carbon”, “ash”, “bungalow”, or “rinse” provided weak context for the participants.

Another barrier of lexical color names was when the color reference point was not consistent with the traditional meaning of the reference object. We all use our own perceptions and experiences when interpreting color names. However, when the color name disagrees with the traditional meaning assigned to the reference object, the participants experienced confusion. Dani, who has some light perception and can differentiate between varying color values (i.e., shades of colors), shared that she faced uncertainty when JAWS^® read, “camo multi-3”. Here, the print refers to camo designs with various shades of black and gray, yet in her mind, she believes camo shades should be green and tan.

Dani: “Camo multi-3 [which is exactly ready by JAWS^®]. So it should have 3 colors [in my mind], but I can only see black and maybe a gray? I was assuming by the name, that it was shades of green and tan…Isn’t that what you think when you think camo? (…) It’s hard for me to differentiate but usually those [green and tan] are lighter shades [than black and gray] with my vision and these look almost black”.

Similarly, with the same product, Dani continued, “What is leopard multi-2? So is it black and gray or is that a navy blue?” By using the term leopard to label this fabric pattern, we both believed it would be a combination of different shades of brown in relation to the animal. However, the print was indeed in navy blue and gray. In the case of “camo multi-3” and “leopard multi-2”, Dani was able to differentiate the color intensities in hues of black, gray, and white; however, the color names made no sense to her, even to the researchers, who are sighted people. Similarly, Rita, who is legally blind with some light perception, shared, “It says [JAWS^®] it’s army green, but it looks like it might have other colors on it as well”. As a sighted observer, the primary researcher shared with Rita that that particular product was primarily a shade of green but had bright pink accents on it. Rita replied, “it would be nice if it [the products’ color options] had listed it as ‘dark army green with pink accents’ so I understood”.

The final instance where color made no sense was numbers assigned to the color options. This is not surprising since the field of computer–human interactions utilizes numerical color naming conventions for image processing and computer graphics [38]. However, numbers attached to colors furthered participants’ confusion about color information. Dani reviewed a workout outfit and shared, “it’s [JAWS^®] telling me A1-black and A2-black. But, what is the difference? It looks like a shade that is not black”. Similarly, Eva searched for shoes and shared, “what is the difference between black and black-bf2?” and the same item also listed “gold1”. As a sighted observer, even the primary researcher was unable to differentiate between “black” and “black-bf2”. The only difference between the two was the fact that the black option had a black interior of the shoes, and the black-bf2 option had a nude interior. Kate experienced a similar issue, JAWS^® read the color options as “150 dash black” and “166 dash silver”. The silver option to me appeared to be silver glitter with silver rhinestones, and the black option appeared to be black leather with no surface design elements. Again, for half of the participants, gathering product information about color when the color naming included numeric identifications further confused them as consumers.

3.1.2. Power of Words in Product Titles: Keywords vs. Clarity

All participants used the keyword search box to search for and obtain product information from the product title (i.e., header content), and they were categorized by keywords and clarity. Participants who searched common styles or phrases, like “workout set for women” or “pajamas plus size set for women”, were able to generate the intended items but shared that they were confused on certain terms in the product titles. In almost all cases, the seller’s brand name was listed first in the title, and participants shared that adding these random brand names makes it more difficult to find information about product attributes.

When reviewing an exercise apparel item, Dani asked about the term “naked feeling” in the product title, then selected a similar item, which included “seamless” in the title and stated, “oh seamless, so it will be so light and soft. Maybe seamless should be paired with naked feeling”. Similarly, as Eva searched for a pair of shoes, she shared, “this is going to be terrible trying to search for shoes, but that’s why I’m going to do it”. Eva utilized “women’s black dress shoes for work” as her initial keyword search, which generated product titles with confusing phrases like, “monte juliet loafer”, “adley boot”, and “parigi pump”. As the AT read product titles, Eva shouted “no” exhaustively and asked, “what does that even mean?”. She then narrowed her keyword search to “black flats for women” and was able to choose a product that met her criteria through a product title with keyword clarity (e.g., flat shoes, women, comfortable, round toe, memory foam, and women’s flats).

Most titles were vague or lacked detail in design elements like silhouette, construction, details, etc. To overcome this when searching for a product, participants shared that being as descriptive and detailed as possible with keywords is how they narrow their search. For example, Mia, who is legally blind with some light perception and considers herself a frequent online shopper and a knowledgeable user of AT, was successful in generating multiple detailed keyword searches for the same item. Mia shared, “I’ve noticed the long string of tags that you need in order to make an item more findable. But, those key phrases make it difficult to find what you’re looking for”. She generated desired items by searching “long high waisted skirt” and “wide soft belt dark brown”, which included silhouette, line, and color elements (i.e., length of skirt, where skirt hits at waist, width of item, texture, etc.). Similarly, Kate noted product attributes and tiny details help form her keyword search and “if the title can be as descriptive as possible, like focusing on the smallest details it helps to make the most informed choice possible”.

3.1.3. Time Drain, Size Maze

The final, but not the least, barrier to product information for all participants was obtaining size options due to inaccessible and inconsistent charts, text descriptions, and size drop-down menus. Specifically, the time spent to obtain inconsistent size options from the three different locations on the product landing page created frustration among all participants. For example, Dani and Mia shared, “where’s the size chart? Do you see it?” and “it’s very difficult to find the right size because how do you know where the size charts are?” From a sighted perspective, the “size chart” function on Amazon’s product landing pages is located under the size drop-down box in small, blue hyperlink text. Participants used tab navigation and found the size drop-down box, but after their AT read the size options, they often asked where they could find specific size information. Through linear navigation, the “size chart” hyperlink would be next in their navigation pathway, but almost all participants stopped looking for sizing information after the drop-down box options.

When participants found the hyperlink, it led to an infographic, which furthered their confusion. The screen reader function read a string of random numbers and size values, which needed more context to select a size. Often, column and row labels were not read, creating more questions for the participants. Inaccessible size charts are in the form of images or PDFs that lack color contrast and alternative text, which creates issues for AT functions like screen readers and zoom magnifiers. In cases where the screen reader recognizes the chart, it rarely communicates the column-based format and interprets the size chart with correct pronunciations and context. For example, Nora was unable to complete the task of adding an item to her cart due to sizing frustrations. The size chart for the desired product was read by JAWS^® as “seven-to-eight, nine-zero-one-chat, and one thousand one hundred and twelve” instead of size seven to eight, nine to ten, and eleven to twelve. Even with this incorrect information, Nora navigated to the drop-down size box of the same item to double check if her size (which is 9–10) was available. When she clicked the drop-down size box, JAWS^® read “seventy-eight, nine hundred and ten, and one thousand one hundred and twelve”, which was incorrect and inconsistent with the size chart. JAWS^® should have read “nine to ten” which was listed. Instead, it read “nine hundred and ten” or “nine zero one chat”, both of which made no sense to her. So, Nora thought that they did not have the correct size for her and left with disappointment that they did not have her size.

To further the issue, Mia was unable to add the desired product even though her detailed keyword searches generated a product that met all her criteria. Fusion, an AT program, read the infographic size chart by each row without any pauses, creating difficulties for people to differentiate between size options. She stated, “dang, they don’t have an XXL. This frustrates me”. The size chart did not list the size option for XXL. When she returned back to the product landing page, the XXL size was listed as text in the product description—another example of inconsistency between the size chart and the product descriptions. Despite the confusion, she navigated to the size drop-down box and was surprised to hear the XXL option. However, it did not provide her with the waist measurement in inches (which should have been in the size chart), so she did not add the item to her cart. Overall, Mia spent 9 min just on the size information for one product even with the help of AT. Ultimately, half of the participants failed to add desired items to their cart because AT (i.e., JAWS^® and Fusion) did not correctly read the infographic size chart even if the retailers had the right size for the participants.

3.2. Difficulties in Navigating Webpages (Culnan’s [11] Interface Dimension)

Theme two, difficulties in navigating webpages, revealed three sub-themes: (a) online labyrinth; (b) forest of products, no reviews in sight; (c) voices in the void.

3.2.1. Online Labyrinth

Navigation is another key dimension discussed in the digital accessibility literature. Indeed, for the participants, it was the most complex challenge while completing the task. Although experienced online shoppers who are expertly or knowledgeably skilled AT users, participants felt overall frustration selecting apparel online—five out of eight participants shared they would rather have a family member purchase the item once they have browsed, in fear that they have chosen an incorrect item. All participants struggled with linear navigation; in fact, the apparel selection process was disrupted as they tried to move around the website pages due to visually stimulated navigation features such as mouse hover zoom function, complex page formatting, and overwhelming advertisements. Participants shared general dislike, “this is why I don’t shop on Amazon. It starts going all over the place and drives me crazy” and “how did I get here? I’ll have to start again, that bothers me”. Yet, participants expressed they believed the issue stemmed from how the seller listed the item and Amazon’s lack of consideration for linear navigation in website design rather than their chosen AT. Here, Rita expresses a general dislike toward navigating the online apparel shopping labyrinth:

“I find shopping online to be rather frustrating, either in the case of websites not being accessible with JAWS® or having so many links, advertisements, and stuff like that [visual stimulated shopping pathways]. Sometimes it’s hard to navigate around and just ‘quote, unquote’ see what’s on the screen”.

3.2.2. Forest of Products, No Reviews in Sight

Although all participants shared that they check product reviews with each purchase, they share general frustration navigating past sections of the webpage like advertisements, brand stories, and similar functions for items you may like. Eva seemed disheartened about the time spent navigating to reviews and shared, “sometimes I don’t think sighted people realize that AT uses the Tab key and special commands so when you develop the webpage with information that is visually laid out to attract the reader, it overwhelms us”. All participants shared that reviews help inform their selection process by gathering information about the size, quality, color, and accuracy of product descriptions. Four out of eight participants shared that they check the star rating listed in the product description first because it is more accessible than the review section. In fact, they shared that if the rating is above 4 stars, they will spend the time navigating to the review section. Only two participants spent time navigating to the review section during the task; Kate spent up to 5 min and Eva up to 6 min.

3.2.3. Voices in the Void

Navigating to customer service was another navigation issue for all participants. When asked about customer service, five out of the eight participants either laughed or sighed. Only three participants verbalized their frustrations and shared, “they want you to access your online account by yourself which doesn’t help” and “they will tell me to monitor my account online but I keep saying I just want to talk to someone”. This might be a common solution for many online retailers; however, for the participants, monitoring online accounts is another challenge. Eva expressed extreme frustration about the excessive amount of prompts and shared, “I don’t want to have to say I’m legally blind just to talk to a customer service person, but sometimes I’m like whatever! And do it, so I can get what I need”. When asked if they would like to navigate to customer service during the task observation, no participant was willing to spend the time, and navigating to customer service was frustrating. Overall, participants did not want to draw attention to their visual impairment and wanted to independently solve their customer service issues. They expressed that, in the past, they experienced positive experiences once they were able to speak to a live person, but the multiple prompts and the amount of time spent to find a live person are not worth it when they are unsure if they are buying a product with accurate descriptions.

4. Discussion

In response to increasing numbers of people switching from offline to online shopping after the COVID-19 pandemic, we aimed to obtain an in-depth understanding of apparel shopping website design and digital accessibility for PVI. By employing a phenomenological research method grounded in the information dimension of Culnan’s [11] perceived accessibility theory, the study result of eight participants’ experiences as PVI revealed two major themes: (1) difficulties in decoding product information; (2) difficulties in navigating webpages. Theme one was supported by three sub-themes: (a) color names that “make no sense”; (b) the power of words in product titles; (c) time drain, size maze. Moreover, theme two was supported by three sub-themes: (a) online labyrinth; (b) forest of products, no reviews in sight; (c) voices in the void. Overall, findings highlighted the unmet website meta descriptions for product information and navigation functionality for AT, and therefore, negatively impact digital accessibility for PVI to shop online for apparel—resulting in apparel websites’ possible lack of compliance with Article 9 [4].

The study has several contributions and implications. First, this study illustrated how Culnan’s [11] dimensions of accessibility may be useful for accessing informational content in the digital landscape for all consumers, including PVI. The dimensions of perceived accessibility to information have been widely used in communication technology research in the offline landscape, but the present study applied it within the online landscape and extended the context for the information exchange between AT and apparel website design through digital accessibility. In addition, the study also pointed to the need for further integration of AT functions within Culnan’s [11] interface dimension and for expanding it in such a way that it could be used to reveal a standard format for product landing pages. The findings from theme one that PVI struggled to obtain clear and accurate product information for color and size supports Culnan’s finding—in that the delivery strategies for computer-based information need to include an understanding of the types of users who will be using the system. For example, participants found the apparel shopping website not accessible and required another person to obtain product information content. Participants shared that they would rather ask a family member or a friend to purchase the apparel products they have browsed to confirm whether they chose the right product. This finding further supports Culnan’s implication that users are not willing to justify the time spent using the system when it is perceived as less accessible. This study helped us to highlight the informational dimension of accessibility (or lack of) for apparel shopping websites, further supporting Djambsi et al. [32], who found the dimension to be a key determinant for ease of use and usefulness for PVI to navigate online. We show specific examples of the complexity that Culnan mentioned for accessing computer-based information and also highlight the difference between usability and accessibility from the PVI perspective.

The second contribution of this study is showcasing apparel shopping websites’ possible violation of Article 9 [4]. Findings from theme one, difficulties in decoding product information, show the legal implications for apparel shopping websites. In fact, all participants could not access size information through the three different and inconsistent approaches on Amazon’s product landing page. This finding highlights the need for text-based sizing charts alongside graphic images and the inclusion of written text relating to size in the item descriptions. Accessing information about color showcased the issue of numerical names and the lack of standardized formatting among sellers on Amazon. The United Nations mandates Article 9 [4], yet the question of how to enforce apparel shopping websites to be compliant with nondiscriminatory and accessible practices remains. In fact, this finding highlights the need for the international legal system to develop specific standards and guidelines for digital accessibility with a particular focus on online shopping.

The third contribution extends to the business sector, where the findings raised the issue of whether apparel websites are offering an inclusive approach to website design for digital retailing with effective communication of product and navigation information. The finding that PVI are interested in color information, even if they have never seen color the way a sighted person does, provides an opportunity for apparel website designers to include color options inclusive to PVI. The findings of improper or vague lexical names from theme one clearly illustrated the lack of inclusive color naming by apparel retailers. In fact, choosing color names that have a societal-wide common understanding (e.g., camouflage or shades of green and tan) and then producing the apparel item in colors not associated with the origin of the name creates confusion for PVI, who form perceptions about colors through societal influences. The findings from theme two, difficulties in navigating webpages, showcase apparel retailers’ lack of accessibility efforts for navigation information. This barrier in the digital landscape negatively impacted the apparel shopping experience for all of the participants. Even though Amazon offers the most options for apparel selection, participants shared that they do not shop often on Amazon when retailers like Kohls demonstrate higher accessibility standards. Digital merchandising and retailers should consider website design from the perspective of PVI users to improve customer experience and brand image by demonstrating a commitment to inclusivity.

The final contribution offers opportunities for educators in digital retailing to develop and adapt curriculum from sighted-biased practices and offer a more inclusive and accessible perspective to online apparel shopping. The findings from theme one, specifically the power of words in product titles, emphasize the opportunity for digital retailing educators to incorporate PVI-inclusive keywords in product titles when developing product landing pages through search engine optimization and digital marketing education programs. Color names that “make no sense” from theme one reveals a market for educators and students to raise awareness about the complexity of blindness from the perspective of PVI consumers. Specifically, how choosing color names with accurate context that is inclusive to blindness as a spectrum has market potential to reach seven million PVI consumers in the US.

As a qualitative study and broad research design, the limitations of this study surrounded sampling and study bounds. The sample comprised all female participants with varying levels of visual impairments and age. The online apparel shopping setting was limited to a personal computer rather than a mobile phone, which PVI may prefer as their shopping channel. Future research implications include additional user testing of apparel online retailing websites to build on the product information content barriers identified here and to further explore the market potential of an inclusive formula for website landing pages to increase communication efficiency between retailers and PVI consumers, which will benefit all consumers.