The current state of app accessibility

With all the tools at our disposal, we are doing a remarkably bad job!

in 2021, when people had nothing to do while sitting at their homes, the data consumption on phones or handheld devices officially surpassed that on desktops! Even though the world has returned to something that resembles normality, the trend is here to stay. One article from Allconnect states that

Almost three-quarters (72.6 percent) of internet users will access the web solely via their smartphones by 2025, equivalent to nearly 3.7 billion people.

Now if we combine that with 2.2 Billion people (source: WHO) suffering from some sort of visual impairment, which is roughly 25% of the global population, we are looking at about 19% of the global population (roughly 1.5 billion people) that will need mobile apps to be accessible or at least usable with assistive technologies by 2025!

The numbers look about right. so?

A study conducted at the 美国华盛顿大学 by Anne Spencer Ross, Jacob O. Wobbrock, and James Fogarty found the horrors of the current state of accessibility today. The study analyzed a little less than nine thousand mainstream apps and tried to figure out how many of them have all the accessibility levels present, especially for images and icons. The reason they chose images and icons is that the screen readers are efficient in reading the text. At the end of the study, they concluded that about 23% of these apps lack about 90% of labels. This means, that a person, relying on a screen reader is essentially dealing with no data or worse, machine-generated unhelpful labels.

This triggered my curiosity and I decided to try a few apps on my own, with TalkBack (yes, I am an Android user!) Incidentally, I landed on the Trip.com app.

Note: I have nothing personal against the app, they are great in providing the intended services. I just happened to be planning my trips while reading the study. If anyone from the org is reading this, please don't take away my coins!

Testing with the automated tools

The obvious first step in testing, any app for accessibility was to run it through the Accessibility Scanner app. With this app in effect, you have to just open the app that you want to test and you will get a full report of what is missing in terms of accessibility.

Here are the report outcomes in short

• Detected 16 issues related to accessibility with 3 screens
• Detected insufficient target sizes
• Detected insufficient color contrasts
• Labelwise, DID NOT report anything noteworthy

In short, the app is one of the top 23% of applications that has more than 90% labels, that must be great! Not really!

The Human testing

After the great report by the accessibility scanner, I decided to try the app with TalkBalk on my own. There I realized that;

Despite having more than 90% labels present on the app, not a single one of them was useful!

on the home screen, in the top row, where the icons show flights, hotels, and trains in the first row, the accessibility labels read: "Primary entry zero", "primary entry one" and so on. In short, there was no way for a person relying on TalkBack to understand what page they were trying to get to. Similarly, for the second row, where the app shows icons for car rentals and attractions, the labels read: "secondary entry zero" and "secondary entry one"

The story continues on the actual flight booking page. the labels on the second page are not helpful either. for the tabs, instead of using the actual tab name in the label, the talkback returns values like "tab underscore group underscore zero" which is not only unhelpful but also frustrating.

In conclusion, despite having all the labels, the app is not usable with assistive technologies!

This poses a new question.

Of all the 23% of apps that have more than 90% labels assigned, how many are actually useful?

So why is this happening?

In the tech ecosystem, we have a lot of applications that can test the ready app for accessibility flows. Companies like Google and Apple have developed tools that set right into their product development workflow and ensure that whatever is being produced is accessible from the get-go. Android ATF and Accessibility Inspector on XCode are examples of such tools that integrate with the Development environment and find out the problems before the app becomes public.

However, these machine-based tools can check the presence of labels and not the validity of the label. By validity, I mean if the label is useful to the person who has no access to any other means of perceiving the element on the screen, other than the label. The creators of the tools have published guidelines about how to write useful labels and those are available on the sites.

Hence, until the AI catches up, the apps need to be checked manually for accessibility labels, if they make sense. It is the collective responsibility of the designers, product managers, product owners, and testers to ensure that the app is shipped with the appropriate labels. But this does not seem to be happening!

when I started enquiring about the same, I found the top five reasons to be the following

1. Not aware of the necessity of accessibility considerations
2. No technical knowledge (especially for the designers. How to do this, caused by complexity aversion)
3. High cost to fix
4. No ROI, caused by framing and survivorship bias
5. Tight deadlines

So what's the solution?

It is found that fixing the accessibility of a developed app costs a lot of money. So it is a good idea to make the designers and product managers, aware of the accessibility considerations right from the scoping phase. Making product managers aware of the necessity, enabling designers to write correct labels, and coming up with accessible flows that make sense to people with disabilities will be of prime importance. This will start taking accessibility considerations into account while coming up with deadlines, budgets, and resource plans. Designers need to be efficient in generating labels that make sense to the people, relying on screen readers. They also need to be extremely efficient in handing over all this data to the development team, so that we ensure that the right labels get assigned to the right elements during the development phase.

I have collated all my experience working with accessibility in a small book, called Mobile Accessibility Rituals. This book aims to explain the accessibility guidelines in a more prescriptive manner so that designers can simply follow the guidelines instead of learning the complex WCAG guidelines. The accessibility guidelines hosted on the W3C Site are aimed at perfection. They deal with all the disabilities across all accessibility standards like A, AA, and AAA standards. This makes the site content heavy and presumably complex. In the book, I have tried to create a version where the designers can fix the issues for AA standards for Visual impairments.

In conclusion,

When it comes to the percentage of apps that are usable with assistive technologies, the 23% benchmark created by the research mentioned earlier in this article is slightly misleading. Efforts need to be made to ensure that we create a just ecosystem for all abilities as I have observed during my work, it is not that difficult to have the right labels for the right elements with a little addition to the efforts. As a representative of the design community, I cannot insist more on the fact that we, the designers, are responsible for ensuring the products that we create are usable by everyone, irrespective of their abilities. With the proper mobile accessibility rituals in place, we can at least start getting into the habit of considering accessibility right from the get-go. The book is available worldwide via Amazon.

We are ending 2023 with this reality check. Let's start the new year, 2024, with the efforts to level out the playing field.

Originally published on: https://www.aakux.xyz/blogs/state-of-a11y