Free Feedback for #accessiBe

On Friday, 12 February, at the request of Chancey Fleet I joined a call accessiBe set up with her. I have made it a point to only engage accessiBe publicly and with publicly available information, declining invitations from accessiBe in the past.

Michael Hingson, formerly of Aira, had recently joined accessiBe as its Chief Vision Officer and was engaging in outreach in the blind community to people who had been critical of accessiBe. He asked Chancey for a meeting to discuss her “misconceptions”. After she assented, he said he was inviting one of the founders (though two attended the call), and so Chancey asked me to attend.

About 40 minutes into the hour-long call, one of the founders asked Chancey and me to propose solutions. While the request felt like it came from frustration with our criticisms, it is a valid ask. Qualifying my unwillingness to give away free consulting, I started to present suggestions. Unfortunately, as I laid the groundwork for my first suggestion, it was perceived as an attack and I only got through parts of two suggestions.

This post is the result of that conversation.

• accessiBe Must Address its Credibility Problem
• accessiBe Must Engage the Community
  • But Maybe Not Like This (added 3 March 2021)
• accessiBe Needs to Stop Weaponizing Disabled People
  • C− for Token Effort (added 11 August 2021)
• accessiBe Should Join the W3C
  • accessiBe Has Joined W3C (added 20 June 2021)
• accessiBe Needs to Give Back to the Industry
• accessiBe Should Offer an Offline Service
• Wrap-up

accessiBe Must Address its Credibility Problem

As it stands now, accessiBe has a credibility problem. I have documented this: its product cannot protect from lawsuits, at least some of its praise is paid for, its understanding of the ADA is off, as is its comprehension of WCAG, it has never addressed its demonstrated black-hat marketing tactics, and (most importantly) its marketing promises are in direct conflict with its Terms of Service.

In TechCrunch’s article covering accessiBe’s latest round of funding ($28 million), it raises the point (however anemically) that accessiBe cannot catch or fix all issues. The author writes, AccessiBe CEO Shir Ekerling said that this concern has been ameliorated by recent improvements to the technology. Since accessiBe’s marketing has for years promised that it can remediate all accessibility issues, it raises the question — was accessiBe lying then or is it lying now?

Any one of these would rightly erode trust in a vendor. Whether or not accessiBe feels these claims are fair, they are absolutely real concerns for the people most affected by the product — users. Many of these users have influence over purchasing within organizations, and nearly all users control where they spend their own money.

As long as accessiBe fails to acknowledge and address these credibility problems, accessiBe’s lack of credibility will follow the company.

accessiBe Must Engage the Community

Mr. Hingson is starting some of this engagement, which appears to be part of his role. So far it appears to be laser-focused and may rely on his existing professional relationships. This needs to grow; accessiBe cannot expect to have a proper dialogue with the community when singling out critics for two-on-one or three-on-one calls.

On the call, one of the founders referred to a silent majority of users and customers who are happy with the service. That belies a form of selection bias. The vocal minority is vocal precisely because harm is being done, while non-disabled users will not notice them. Chancey shared her experiences with problematic content on sites using accessiBe (broken tables, insufficient alternative text, un-addressed charts, failure to add captions or audio description, and more), which track closely with the issues I have raised as well.

Instead of engaging in selection bias, accessiBe can set up relationships with existing advocacy organizations. It can have webinars moderated by trusted disability community partners and advocates to gather honest feedback where end users are given a voice. No talking over by founders, no assertions that the product is a net good. Listen for, and hear, the concerns about unintended consequences.

This activity can be a good first step in helping address accessiBe’s credibility problem. It could maybe even prevent embarrassing ableist accessiBe social media posts that demonstrate accessiBe does not understand the assistive tech on which end users rely. It might also lead to an understanding that disability tourism is actually harmful.

But Maybe Not Like This (added 3 March 2021)

Selling out? Really? How much have you really analyzed #accessiBe, what it says and what it does? Do that and then tell me specific problems based on what #accessibe says it can do that you find it can't. Tweet and email me at michaelhi@accessibe.com. No generalizations, #facts.

Especially since making any effort to understand the person’s criticism before responding would have yielded the answer:

I literally wrote a blog post where I detail my user experience testing AccessiBe on a fairly simple webpage. No advanced images, just headings and tables.

Once again, Chancey is on point.

Broken headings and tables. Misleading alt text. False advertising. Trapped keyboard focus. Zero support for non-HTML content. Sock-puppet reviews.
Here's a vision: read our stories in the #AccessiBe thread before aggressively singling out critics for homework. #BelieveUsers twitter.com/mhingson/status/136612…

An organization that truly wants to engage with a community does it in the open, not by asking you to contact them privately one by one without even acknowleding your experience.

accessiBe Needs to Stop Weaponizing Disabled People

I am not the only one to raise concerns with overlays. Lainey Feingold gathered feedback from disabled users who tried accessiBe and its competitors, and all found them lacking in fundamental affordances. Yet, instead of accessiBe honestly stating these gaps and challenges, its marketing presents a message that companies should use this tool to target people with disabilities, that they are someone else’s problem, that they can be relegated to a ghetto of quick-fix add-ons.

While accessiBe touts its goal of making the entire web accessible by 2025, it sees that only through the lens of its tool — promoting a future where disabled people are dependent on the good will of accessiBe to improve their experience. Meanwhile, accessiBe actively markets its product with this message on its home page: accessiBe is trusted to protect over 100,000 websites worldwide!

Its message is clear — accessiBe frames its value proposition as primarily a protection from the dangers of litigation from disabled people. Until accessiBe stops weaponizing and dismissing disabled people, it will never win the broad support of the community it claims to want to help.

C− for Token Effort (added 11 August 2021)

Clearly spending some coin, accessiBe paid for a slick ad to promote itself by centering disabled people. Well, for most of the ad. Two quotes stand out:

1. The Internet is inaccessible. Well, no. It and the web were and are quite accessible before we broke it. As I have noted to accessiBe, the famous TBL quote even asserts that.
2. …don’t just do it to be compliant… This is amusing because their entire ad campaign says to do it to avoid being sued.

Clearly accessiBe’s outreach is not working because the National Federation of the Blind dropped accessiBe as a sponsor for its National Convention, likely at a significant revenue loss. I am not sure what is such a clear and stark rebuke from the community accessiBe keeps insisting it exists to serve.

accessiBe Should Join the W3C

The W3C is the standards body behind the Web Content Accessibility Guidelines (WCAG). It defines the very guidelines that are law across the world. The same guidelines that accessiBe promises its service satisfies. But accessiBe does not participate in the definition (or debate) of those guidelines in any way, nor does it contribute to the research.

W3C also defines the Accessibility Conformance Testing (ACT) Rules, which generally describes a method to programmatically apply WCAG. Knowing how to programmatically identify WCAG failures is the first, and necessary, step in providing the most basic programmatic remediation. This is exactly where I would expect accessiBe would want to participate, if only to try to sway the rules to their methods.

I was able to suggest this on the call, and while the initial response was positive, one founder made it clear he was unwilling to join because I, Adrian Roselli, have demonized accessiBe for so long. While I am wary that accessiBe’s failure to join the W3C so far is because of me (though I have been blamed by wealthy orgs before), here is a great opportunity for accessiBe to rise above my mosquito-scale annoyance and contribute.

Not only could accessiBe’s membership improve its comprehension of WCAG, it could also have a hand in defining WCAG 3 by bringing its self-described extensive research to bear.

accessiBe Has Joined W3C (added 20 June 2021)

It has been a few weeks now, but accessiBe finally joined the W3C. So far accessiBe does not seem to be active in any way, which is good for the W3C as a revenue source. Even if accessiBe’s team does start to participate, the W3C process is designed to prevent any one organization bullying its agenda through.

There was rightly some consternation, but in the end anyone can become a member and contribute. The quality of accessiBe’s contributions will likely be minimal, but at least accessiBe can no longer honestly claim it is unfamiliar with the process. Likely accessiBe’s failure to understand WCAG will be obvious to anyone who bothers to look.

There is no denying accessiBe will continue to lie (I use that word intentionally, given its history of documented lies) and now claim it has a hand in accessibility standards development. In fact, expect accessiBe to use it as a marketing tool. But now there will be public evidence of accessiBe’s failure to engage, or if accessiBe does, failure to comprehend.

accessiBe Needs to Give Back to the Industry

The digital accessibility industry has a long history of giving back, both to the community it ultimately serves, and to others in the industry. My blog is an example of that on an individual level. It exists solely because of what others before me have documented. Conferences are another example, particularly the free ones, where companies and individuals share techniques, lessons, research findings, and more. W3C, open source contributions, and more, are built on sharing information that ultimately benefits end users, people.

On the call with accessiBe, the founders were clearly passionate in their belief that they have improved the web through their product. For some people, accessiBe may have genuinely taken something utterly opaque and given it a peep-hole of accessibility. But accessiBe’s efforts seem to stop there. There is no contribution to open source, no pattern library for the taking, no research reports on its claimed AI techniques, no lessons learned about failures in popular frameworks. In this industry, accessiBe is one of the rare companies that takes more than it gives. Looking at its financing, this is certainly not for lack of resources.

With little effort, accessiBe can turn that around. Present at CSUN, start a development blog, finance accessibility research through trusted partners, contribute accessibility fixes to popular frameworks and libraries, get a corporate membership and actively participate with the W3C. Give back to the industry that has provided the foundation on which you have built your business.

accessiBe Should Offer an Offline Service

The burden for making a web site accessible in accessiBe’s model falls to web site visitors. Because accessiBe’s customers only add client-side script to their site, accessiBe’s service never fixes the underlying page. All accessiBe does is attempt to remediate content after the fact. Over and over. For every user. On every visit. And then, only if the script is not blocked and the user has a stable connection and capable computer.

Given that the community accessiBe claims to serve is historically under-employed and over-burdened with expenses, it is unfair to expect users on notably under-powered machines to do all the remediation for the sites they visit. In some cases, the script never runs, leaving users with an arguably worse experience.

Instead, accessiBe can offer a service that performs its remediation on the customer’s server. Just as most publishing platforms offer hooks to third party services to publish elsewhere, accessiBe can write a service to accept, remediate, and return content when it changes. The same can be done for templates. As more organizations rely on an automated deployment pipeline, accessiBe can provide a service to live in that pipeline. And accessiBe can still charge the same (or higher) rates for that service, without bodies in seats 24-7, while the founders are sleeping. Essentially hitting the points that investors like to see.

No longer could accessiBe be so easily (and rightly) accused of providing a “separate but equal” service, but instead it could be treating disabled people as first-class users by removing that extra burden.

An additional bonus is that accessiBe could also expand its mobile support beyond solely Android 8 and iOS10, extending its service to those who have more modern and capable devices.

Wrap-up

If accessiBe wants to build loyalty and reduce its critics, it needs to act sooner rather than later. It is flush with a new round of cash, so the timing of when accessiBe acts will tell us how seriously accessiBe takes its stated goal of making the entire web accessible. So far, accessiBe’s actions demonstrate otherwise.

If you came to this piece because you are evaluating accessiBe and therefore care little for my advice, take a few minutes to read my much shorter post Sub-$1,000 Web Accessibility Solution. It provides a high-level overview of how you can avoid relying on someone else to handle your accessibility needs when on a budget.