Editor's note: Reading a social media feed with a mobile screen reader breaks down in three places, and each one is worth checking on your own site.
- Post readouts: profile pictures repeat the account name, collapsed posts may not expand, and dates and like counts are often not spoken.
- Image descriptions: few posts have meaningful alt text, AI descriptions are inconsistent, and emojis or graphics outside the standard set are not recognized.
- Audio and video: autoplay talks over the screen reader, player controls often cannot be reached, and video posts can trap focus in the feed.
Michael Taylor, a UsableNet contributor, is blind and relies on a screen reader to use the web. He shares his first-hand experiences to help others understand what it is like to navigate the web with a disability. This is the second post in a four-part series on social media accessibility.
Continuing the discussion about social media accessibility, I will be breaking down the process of reading posts in this blog. Perhaps the activity that I spend the most time on when using social media is browsing my feed and digesting content. This is not always an easy process with a screen reader due to common deficiencies and usability concerns, which I will elaborate on below. Just to reiterate, mobile touch-based screen readers will be the focus here.
On most social media apps, the screen reader will focus on an entire post as one block or web element. This, in and of itself, is good, because having to hear each component of a post separately, swipe by swipe, would be incredibly frustrating. However, there are still some potential issues here that can make hearing a post's contents clunky.
This first one pertains to usability. The screen reader will often read the profile picture as the poster's account name, in addition to the visible account title at the very top of the post. As an example, this can sound like "John Smith Profile Picture, John Smith." While this may not technically be incorrect under accessibility standards, it adds clunky redundancy that slows down my navigation.
If I want to move through my feed quickly and digest as many posts as possible in a short amount of time, the more concise the post's identifying information, the better. If the profile picture is labeled with a name identical to the poster's account name, which is spoken anyway, the picture announcement is extraneous.
Some social media platforms show longer posts in a collapsed view to conserve space on smaller screens, and users need to tap on the post to expand it and reveal the rest of the content. This process is often problematic for screen reader dependents. For one, the expand/collapse element, most commonly an icon button, is sometimes not labeled for screen reader recognition. Secondly, a post will sometimes refuse to expand when I activate it with my screen reader, making it impossible to examine the full contents.
What works best for me is when the code allows the screen reader to speak the entire text even if it is visually collapsed. Some social media platforms do this, and it is incredibly helpful.
When I listen to a post's details, some important items, such as the date and time of the post and the number of shares or likes, are often not spoken by my screen reader even though they are clearly visible on screen.
While I have mentioned the importance of alternative text image descriptions on other types of websites, their impact on social media is tenfold. Some social media posts, in fact, have no text at all and are entirely visual. It is shocking how few of the social media posts I come across have meaningful descriptions for the visually impaired. While AI-based image analysis on social media can sometimes generate image descriptions, in my experience, their consistency and clarity leave something to be desired.
In my opinion, any and all images uploaded to social media must have some sort of description for the visually impaired. I fully understand that the photos on these sites are added by end users and are not the direct responsibility of the platform developer. However, I believe there needs to be an enforced framework that requires picture descriptions that meet certain standards of quality and clarity.
Without these descriptions, so much of what is on social media is lost on the blind community because we cannot take in visual content or appreciate how it relates to and supports the written commentary provided with a post. The same needs to apply to any symbols, icons, emojis, or small graphics included with a post that are outside the commonly recognized set of Unicode characters. My screen reader will not recognize these unless they include proper labels and descriptions.
Many social media platforms these days encourage posts that include, or entirely consist of, short audio or video tracks. Engaging with these is a frustrating process for the visually impaired. Firstly, in most cases, platforms play video or audio by default once a post becomes visible on screen, which can be detrimental to accessibility and usability.
Screen reader users cannot appreciate video content if the screen reader is speaking over the audio, and they cannot effectively hear navigation announcements if an audio track plays incessantly in the background. Furthermore, there is often no screen-reader-accessible way to use the video player controls to pause or rewind playback in real time.
Some social media platforms have an option buried in the settings to disable video autoplay. However, it is not universal, and for some reason it sometimes fails to work correctly if a screen reader is enabled.
When browsing my feed, I often find that posts with audio or video act as focus traps when I swipe through. Focus gets stuck on the currently displayed post, and the only way to free it is to scroll down and focus on the next post using direct touch navigation.
As things currently stand in terms of accessibility and usability, visually impaired individuals like myself cannot always get the same level of enjoyment from social media as our sighted counterparts. This must change for us to feel fully included on social media.
Why does a screen reader announce the account name twice on a social media post?
When a profile picture's label matches the account name shown next to it, the screen reader speaks the name twice, as in "John Smith Profile Picture, John Smith." It may not be a standards failure, but it slows down every post in the feed.
Can screen readers read collapsed or truncated posts?
Only when the expand control is labeled and responds to a screen reader, and in the feeds described here, neither was reliable. Feeds that let the screen reader speak the full text while the post stays visually collapsed avoid the problem.
What happens when a social media image has no alt text?
An image with no description gives a screen reader user nothing to read, and on a post that is entirely visual, the whole post is lost. AI-generated descriptions sometimes fill the gap, but not consistently or always clearly.
Does autoplay video cause problems for screen reader users?
Yes. Autoplaying audio and video talk over the screen reader, so neither the content nor the navigation announcements can be heard clearly. In the feeds described here, posts with video also trapped focus, and the autoplay setting did not always work with a screen reader on.
News feeds, product galleries, review lists, and blog indexes work a lot like a social feed, with repeated cards, truncated text, images, and embedded video. Move through one with a screen reader on and check whether dates and counts are spoken, whether a "read more" control works, whether video plays before anyone asks it to, and whether focus can move past it.
An automated accessibility testing tool can help find missing image descriptions and unlabeled buttons at scale, while focus traps and autoplay behavior are worth testing by hand. If your team posts on social media, your images need descriptions too, and gaps like these can lead to accessibility complaints.
Images and video are both on the agenda for our upcoming webinar with OuterBox, Web Accessibility, SEO and AI Search: Building Clearer Website Content, on November 18 at 12 pm ET. Register to hear Jeff Adams of UsableNet and Colin Gacek of OuterBox share best practices for images, video, and the everyday content decisions behind them.