Accessibility Data is Not The Enemy

The following blog post has recently done the rounds and it is yet another article about accessibility that doesn’t quite add-up, and is frankly rather misleading in places.

Please have a read of?the original article?first that sparked off my need to write a response.

Apologies to the author, but until I read their blog I had no intention of writing this article. But sometimes it’s good to react to something to find a way of exploring an issue.

Here’s how it breaks down.

1. ‘How many people with disabilities use our site?’

This is?an interesting question as long as we all have a common understanding of what we mean by ‘disability’ and why we are asking the question.

I’ve previously posted about this so if you’ve already read that reply, skip to point number 2, however I did miss something out. I didn’t mention that the author confuses disability with assistive technology usage.

The question is about “people with disabilities” which is a demographic question. But the answers he gives are concerned with people who experience barriers. The two things are not the same.

Assistive technologies are used by some disabled users as these can enhance both the accessibility and usability of a website if it was designed to support them. But this is not true for all disabled users, and different people with the same conditions will have varying preferences because user experience is more about lived experience than labels.

There are many other people who identify as disabled or neurodivergent who also use assistive technologies or need features so although assistive technology is important, it is only one aspect of the audience in question.

Then they go on to say that?“accessibility benefits everyone”, which is true, but to understand that you have to accept that disability and assistive technology use are not the same thing or even accurate identifiers of each other.

Many people use assistive technologies who do not realise that have a disability or identify as disabled, or use them to overcome transient situational disabilities rather than ones related to conditions that are not designed for. Pinch Zoom, Closed Captions, High Contrast, large touch target areas, font readability, voice control etc. There are many great examples of this.

It’s also simply the wrong question for a designer to ask, it’s loaded and not really answerable because first you have to agree what you mean by ‘disability’.

Is this the medical or social model, and which is the most meaningful to design?

If it is the medical model, do you include conditions that do not in context create user need when using a website/app, like being an amputee with one leg.

Also, do you include hidden disabilities to which there are no reliable statistics?

If it is medical, then is this a demographic question or one where you want to understand the significant barriers commonly experienced by people with particular conditions, and how do you deal with compound need from audience intersectionality such as people who have a vision impairment and dyslexia?

Or is this a social model question?

If so is it better to ask, ‘How many people does the design of our product or service disable?”

This is a far more useful question as you can instead look at the barriers created by the design from not thinking broadly enough about human need and preferences.

Then you can focus on how many of those barriers are permanent and how many are transient.

2. ‘Looking for this data hints at trying to find return on investment.’

Here’s where the contradictions start and BTW why shouldn’t ROI come into this? Ethics and opportunity do not have to be not mutually exclusive.

BTW ROI does not just have to include financial return. It can include engagement, satisfaction and even brand perception.

So if we instead have data that looks at the whole scope of the impact of a programme then we could have the ROI case nailed.

But to do this we require meaningful data and the willingness to accept that we are talking about barriers or needs and not assistive technology when it comes to determining who is disabled in terms of user experience.

Before anyone says it, getting the ethics and law aspect of this right is paramount.

We definitely should not be trying to?detect assistive technologies?because if nothing else it is incredibly difficult to do with any degree of accuracy.

Most techniques detect the likelihood of an AT user but can’t differentiate between a screen reader, switch user, magnification software , eye tracker or voice control. All used by people with different conditions so the data you get is completely useless.

There are more sneaky ways of doing this but I will not publish them because they are totally unethical and potentially illegal.

So please don’t ask me how to do it, I will not tell you.

I believe that if we accept that accessibility is about the needs and preferences of users, and that needs are shared by multiple user groups for a mixture of accessibility and usability reasons, then we can find a less divisive or political data model.

Take a look at the closed caption data coming from organisations like Facebook, Twitter, Netflix or the BBC.

Usage ranges from 40% to 80% of a total viewing audience.

That does not mean that all those people have a hearing impairment. But they do have a common need and preference, which is for the inclusion of closed or option captions. If you then drill down into that group you find that people have different requirements in terms of captioning, but the baseline is a visual only modality or visually weighted modality of AV content.

If you looked at the individual groups by situation and condition you will get smaller numbers, but if you consider the need collectively it adds up to a small investment in time and production to increase the reach and engagement with the majority of your audience.

This is an example of where ROI can be so powerful.

Going back to the initial question, “‘How many people with disabilities use our site?’

Instead switch it to “how many people is our site disabling?”

You can have a much better discussion.

Alternatively you can even ask, “how do we design and build our product to maximise reach and engagement by not disabling our customers?”

Frame it whatever way you think will be helpful.

3. “…our organisation has three good reasons to prioritise accessibility that exist regardless of a number of users with disabilities.”

So if we can prove that #a11y benefits everyone because different aspects of it meet different general user needs, then I believe ROI should be on the table for discussion.

We can’t make the claims that accessibility benefit everyone and then refuse to provide proof. By openly worrying that the audience might be too small we are contradicting this statement. So which is it to be?

By using aggregated metrics of need rather than the demographics of condition, we can remove the issue of privacy and audience size, and also make it more about audience experience rather than identity.

It’s a simple case of matching using the right question and then finding a methodology that matches it.

4. “Privacy trumps metrics.”

This is a misleading statement as it is based on an oversimplification of what accessibility is.

Privacy should be respected, of course it should, especially when it comes to people’s medical information, how they identify themselves, stigma and potential discrimination.

If data compromises any user in any way then this is bad data, but not all data is bad.

Going back to the idea that accessibility is about designing for users’ needs, and then accepting that needs are intersectional across multiple user groups, then this becomes about UX rather than medical data.

The issue isn’t about data vs privacy, the issue is about using the right data in the right way to inform the right choices. Add to that, give every user the option to opt out of being included in any data set, even it their data can ultimately benefit them. As users that should be our choice.

But the author then comes full circle with, “Standards organisations are careful not to add features to the Web Platform that allow such tracking, because it would invade individual user needs too much.”

When they talk about user needs are they presuming that everyone with a need is also an assistive technology user.

This is the overriding problem with this article. The two things are not the same.

5. “Your analytics don’t show market potential”

If you use analytics to determine need and not assistive tech use or medical condition, then yes, your metrics can show market potential.

Why not?

If we go back to the?“benefits everyone”?statement and we combine this with?“need”, then this is a great opportunity to make UX Design more inclusive, and measurably so.

The point made using Tim Cook’s example is very specific about one particular user group and the features Apple develops and provides on their platforms for that group.

These features are not just about people who are blind of vision impaired, even if the statement makes it seem like they are. Apple’s accessibility features support people with cognitive conditions, motor impairments and hearing loss or deafness. Many of the features and settings are shared between user groups so in some way’s Tim’s statement is great from an ethical standpoint because it is an answer to a question about active discrimination and the ethics of Apple.

You would get a similar answer from the BBC if you asked them about the money they spend on bi-media production and audio description, which ensure upwards of 80% of their broadcast content is accessible to people who are vision impaired. As there is no way yet of accurately measuring AV accessibility for people who are blind as AD only relates to some formats o programming, this will always be an estimate.

The problem here is that we mix-up the reasons we make certain decisions. There are decisions driven by the ethics of the brand promise or policy. There are decisions we make because we are legally required to do so. And there are decisions we make because of the breadth of the benefit they have. Do not confuse these, but do think about how you report the impact of your work based on them.

So if we think about the metrics of ‘benefitting everyone’, and measure the impact of the aspects of accessibility and inclusive design that do just that, then we can shift a large amount of what we do from worthy to worthwhile.

Why should we be afraid of working out what these metrics are and including them in any ROI objectives? No-one has given me an answer to that. Please comment if you have some thoughts to share.

6. “Even if we could accurately measure how many people with disabilities used our site, it isn’t a very meaningful number. If our site is inaccessible to people who use voice control, chances are those people are shopping with our competitor instead.”

Here again the author mixes up the concept of disability with assistive technology use. Voice control is mainstream. If you are a user of Siri, Cortana, Alexa, Google Dot… does that mean that you are a disabled person or that you prefer to use your voice as a way of interacting with a product?

User preferences are important and are changing all the time, and assistive technology is increasingly becoming mainstream. People are talking to devices, outputs are becoming increasingly multi-modal and we are becoming more aware as designers that the barriers people face are situational, like the spike in usage for taxi-booking apps at night. People can be temporarily impaired due to tiredness, exhaustion, low charge on a battery (reducing the brightness of the screen), and they can be motor, cognitive or even vision impaired through the use of alcohol or recreational drugs.

These are lived experience barriers which impact the whole of the audience, so apps need to be designed to meet these user needs or they are failing all their users.

Clever designers will then learn from people who experience related barriers as part of their everyday experience, and therefore designing for those users they will effectively be designing in benefit for everyone.

This section of the article references the WHO report on disability.

“For the potential, we could look at the?World Health Organisation’s Report on Disability, published in 2011. In a comprehensive chapter on demographics, they conclude 15–20% of the world’s population has a disability. These numbers aren’t exact, as countries have different methods of counting, but they give a reasonable estimate that we can work with.”

There are some big problems with these numbers especially if you are relating them to user experience, and this is hinted at by the author.

Statistics on inclusion can often not be very inclusive or tell you the answer to the question you are asking. The WHO statistics are about a specific lens on disability that uses many methodologies, some effective and others not, to give an estimation of a demographic.

A big issue with these numbers is that they largely discount cognitive as well as other hidden disabilities. I would say that at best they tell us that there are no less than 15 to 20% of the population with either a sensory or physical disability.

This is all medical and not social model numbers so they are more useful to marketing and diversity data than they are for UX design.

If you want to know how many people there are in your audience who have specific needs that should be designed for, if you want to avoid designing UX barriers for them, then ignore the WHO stats. They are fairly useless in this regard.

7. “Accessibility features on our site won’t benefit everyone all the time, that would be an exaggeration, but they often benefit many more people than just specific groups of people with disabilities.”

I mostly agree with this statement, even if it does mix-up the medical model of disability with user experience. If instead the author focused on the intersectionality of user needs they wouldn’t be so reductive in assuming accessibility features only directly relate to specific disabled user groups.

8. “We probably don’t need to know how many people with disabilities use our sites, as regardless of what that number would be, we should want to build accessible sites, for many ethical, legal and business reasons.”

Here’s the big missed opportunity.

I agree that knowing the demographics of disability for a website, mobile application, video game or enterprise system are largely useless because they tell us very little about user needs, so why bother collecting that data except for diversity and marketing stats… marketing, HR and diversity folk, please discuss.

But instead if we use data to understand the needs of all our users and think about how we are providing designed experiences that avoid disabling people, we find opportunity and turn accessibility from something that is a worthy pursuit to something that is worthwhile, ethical and perfectly valid for ROI.

So think about the drivers of your accessibility programme. There are 3 of them.

1. Ethics. Of course we all want to be inclusive but society isn’t inclusive so we focus on cultural change within an organisation.
2. The law. If you don’t want to be sued then this is all about compliance. There is lots of sensible stuff in there but being compliant does not mean that you are being holistically inclusive.
3. Opportunity. I always put inclusion as part of opportunity because I include anyone who experiences barriers if their need are not considered within this group. But to be fully inclusive you need to understand both the scale and nature of the barriers being designed, and the impact of the decisions you make to design past the barriers, and for that you need evidence.

My Conclusion

If we continually avoid the need for reliable and meaningful UX data then we will continue to fail at being inclusive to everyone and optimal for our users in whatever situation they find themselves.

If you want good data then think about the questions you are asking. Make sure they are about user experience and not marketing or diversity, and make sure that you track changes over time with statistically significant user data so you can prove the impact of your accessibility programme, and give weight to the statement that?Accessibility Benefits Everyone.

#accesibility #a11y #UXdesign #data