
On July 14, 2020, Accessibility New York City hosted a meetup entitled “Using Task-Oriented Walk Throughs for Accessibility Reviews.” Presenter Sarah Pulis, digital accessibility and inclusive design expert and co-founder of Intopia, described a novel process to use when evaluating tasks in the user interfaces of various information technologies.
Here are some of our highlights from her presentation:
- Prepare your personas
- Identify user goals
- Create action lists for tasks
- Perform the walk throughs
- Discuss solutions to potential barriers
- Take-aways
Transcript: Using Task-Oriented Walk Throughs for Accessibility Reviews - Sarah Pulis
THOMAS: And with that, I'm gonna hand it over to Sarah, to give her presentation for tonight. Thanks, Sarah.
SARAH: Thank you so much, Thomas. It is absolutely lovely to be here with you all. And to talk today about task oriented walkthroughs, and how they can be used for accessibility reviews. Anyone who knows me knows I am absolutely terrible at introducing myself.
So I have a little note here that says I probably should tell you a little bit about me. So I've worked in accessibility now for over ten years, actually. Just being able to focus on accessibility and inclusive design. So that's been a really, really cool sort of experience.
And over that time, I've worked with lots of different organizations, everything from sort of helping them make their websites more accessible, through to how, at an organizational level, do we actually make accessibility business as usual, and all the things that we might need to do, like looking at policies and frameworks, and processes, and building up capability through training, and you know, I've probably done it. As Thomas mentioned, I'm also quite involved with the community.
Co-organizing the Sydney accessibility Meetup, and I've got an amazing team of people that are just second to none. It's really cool. I am watching the chat. So when I suddenly laugh out loud, that's probably what's coming.
Glen was saying -- yay a11ybytes. He was one of our presenters this year, the first time we moved a11ybytes online, for obvious reasons. He was an awesome presenter. So let me start by sharing, I guess, how I came to think about using task oriented walkthroughs.
I'm going to put a little ask out here for you. I'm going to share this story, it is about a real organization, and I would really appreciate it if you didn't jump on social media and say: Oh my gosh, so and so organization has done X. I've shared this before, and unfortunately, everyone misses the: This organization was really sort of focused on accessibility, and they just focus on the one little bit that I'm using to demonstrate how I came to think about task oriented walkthroughs.
So if I could just ask you to be a bit sensitive about that, and not push that out publicly, without context, that would be awesome. So here in Australia, we have two main supermarkets. Coles and Woolworth's. And we were working with Coles on a rebuild of their website.
So Coles, you can go and buy your groceries, either in-store or online. And my colleague was actually embedded in the Cole's team. So when they started to rebuild their website, they had a real focus on accessibility. And they wanted to do accessibility by design, and make sure it was considered from right at the start of the project all the way through.
That included both formal technical testing against WCAG, the Web Content Accessibility Guidelines, the standards for accessibility, as well as usability testing. So they were doing all of the right things. The website came to me -- so my colleague was embedded in that team, providing that expertise, and working with them, and they wanted to do a WCAG review, and I hadn't worked on the project. So I had no context, other than we were helping with this particular website.
And so I actually went and did a formal WCAG review on the Coles website. And here I've got a picture of the homepage. It's got sort of a banner at the top, your usual, it's got the logo, a way to search for products, a few navigation elements, like everything for specials, and a few more menu options, and sort of in the main center here, it's got some of those tiles, it's got your proverbial advertisement -- or actually a few advertisements -- and if we can see down a little bit, which is outside of the screenshot, it then has some of the common items that people might want.
And lastly, then, on the right hand side, we have four buttons. Trolley and checkout, choose a delivery time, log in or sign up, and help or support. Now, I have a question for you. How many tabs do you think it takes to get to the trolley and checkout button?
Where that trolley and checkout button was the one I just mentioned visually, on the right hand side. I'm gonna open it to chat. Normally I have people shout out, but obviously we're virtual. How many tabs do you think it took to get to trolley and checkout?
Charles Hall. 87. All right. I have 87.
I have an ideally 1. 10, 14, 26, 5, 15, 11, 14. So it looks like a lot of people think it should be up to 20-ish or so. Charles has gone way out there at 87.
You can't reach it? No, that wasn't a problem, I'm happy to say. So it actually took. And -- have you seen this presentation before, Charles?
It took 87 tabs. If I was just tabbing through this page linearly, to actually get to that trolley and checkout button. And I'm like... Wow.
This is a pretty important thing. Not just for users. You know, when you want to actually... You put all your items in your basket, you want to go to trolley and checkout, that obviously is something that you absolutely want to get to, it should be easy to get to.
And also, you know, from a business point of view, I mean... Being able to easily purchase something is core to a business. And so for me, this is like... Wow.
This is potentially a really major problem for someone who is just progressing linearly on this site, either by tabbing or reading through the content with a screen reader. Although I know screen readers have lots of different shortcuts they can use to get to sections of content. And so interestingly, for those that are involved in accessibility, and know WCAG, I'm a bit of a WCAG wonk, I kind of know it back to front and inside out, our criteria around keyboard, so mainly 2.1.1 keyboard, talks about: Can you operate something with a keyboard?
Can you do everything with a keyboard that you can do with a mouse, for instance? But it doesn't say anything about this. I don't think I found one keyboard failure on this website. You could tab to everything, activate it with keys, you could use keys for all standard interactions, which is certainly important.
But couldn't actually fail them. So when we're doing reviews like this, if we come across an issue like this, it might not be a direct fail of WCAG, but we're like... Hey. This could have a significant impact on users, and we're gonna raise it as what we call an expert observation.
And we pop it down and we pop it down, why we think it might be an issue and what we think the impact on the user might be, and what have you. And so I did that. And to Coles's credit, they went... Hey.
I can see exactly how this could be a problem. But of course, we were post-development. Everything had kind of been done. And so the overhead of actually changing the order in which someone tabbed through this page was going to be too much to do.
And so what Coles did is they put a few extra skip links, so we can see in the top left hand corner of the screenshot, we've got a view trolley and check out skip link, and that took you directly to the trolley and checkout button, so you can get there and progress. And put a few extras to the login and sign-up and what have you as well, to make sure it was easy to get to those menu items.
But this process made me reflect on a few things. The first was: When we think about WCAG, WCAG is really focused on what I call functional requirements. WCAG has to be designed in a way, because we are testing against it, it has to be designed in a way that you can do yes/no, pass/fail sort of outcomes. So can I access everything with a keyboard?
Yes/no. We can get a fairly clear outcome, to paraphrase that success criteria. But can I easily get to something that I want? Is quite subjective.
What's considered easy? We had a lot of guesses there. Should it be 10? Or should it be 15?
That's quite hard to measure, outside of context. So I totally get why WCAG doesn't cover this. In its current structure. But it does mean that if we're just focusing on a WCAG review, then really, we're just focusing on those functional requirements, and we're probably more focusing on a page by page assessment.
We're assessing each individual elemental component or information on a page. We're not looking at it from a task perspective. I want to buy something. So that was my first insight.
My second one was... I did this about four years ago, actually. I know, because I was wrapping this up just before my wedding, which always sticks in my mind. And I realized at that time that I was actually thinking...
More outside of WCAG. So I wasn't just following a process or a checklist. Is it keyboard accessible? Yes/no.
I was thinking about what a user wanted to do. But I was thinking about it in a bit of an ad hoc way. I didn't have a formal process behind me, to sort of go... Okay.
I'm going to review it in this way, or have this particular user focus, or what have you. And the last insight was: Despite the fact that Coles had taken an accessible by design approach, my colleague, who is actually one of my co-founders, really knows his stuff -- it was still something that was overlooked. It wasn't something that was picked up earlier in the process. And by the time I picked it up, obviously, as I mentioned, Coles at least did put in a fix, but that fix was not to redevelop the website or to change the focus order.
And so it got me thinking about how we could... I guess formalize this process a little bit, to look at more the task path, the path that someone would take to complete a task. And secondly, how we could identify this issue much, much earlier on in the process. So that we could make some informed decisions about -- whether it's the design or the order in which someone tabs through the page, or that kind of thing.
We actually considered that earlier on. And this is how I came to sort of think about this task oriented walkthrough process. It's quite similar for anyone who is a UXer. You've probably heard of a cognitive walkthrough process.
And it is extremely similar to that, and I guess I've been lucky enough to work in UX teams before, and sort of have exposure to different tools and techniques, and I think this is where I sort of drew on that knowledge of other tools and methodologies that were already out there. And just adapted it a little bit for, in particular, for accessibility. So when we talk about a task oriented walkthrough, it is a method, and usually what happens is a person like myself works through tasks as if they were a user.
And the ultimate outcome is to better understand if the user can or might be able to easily carry out tasks within a system. Now, I've got the big red flag here. I am not saying task oriented walkthroughs are a substitute for usability testing, for testing with real users. I love talking about all the different tools in your tool kit that you have, when you do accessibility.
You know, one is WCAG. Another could be a task oriented walkthrough like this. Doing user research, usability testing, that kind of thing. That's another tool.
So for me, this concept of a task oriented walkthrough is just another tool in your tool box. But we are not trying to say that by just doing this, we actually do understand what a real user's experience will be. So let's unpack a little bit about how you go about performing a task oriented walkthrough. We've got a few steps here.
The first is to prepare your personas. They're the people you're going to adopt to step into the shoes of these users, as you walk through your system, product, whatever it is. With this particular user in mind, and with a particular task in mind. The second thing is to identify your user goals.
What the user wants to do on your system. The third step is going to be to create a list of actions that the user might take to achieve each of those tasks or each of those goals. The fourth, to perform the walkthrough, and then five is: Discuss the solutions to any potential problems that have been identified. So let's first look at preparing your personas.
So first of all, what's a persona. A persona is a fake person, a fictitious person, who represents a real user that will traditionally be using your product or service. Usually personas are created from primary research. So actually talking to different users or customers, if you've sort of got customers -- and taking all of that research, and kind of creating -- there's no hard and fast rules for how many personas you have, usually.
Up to eight, I would say, I've seen, in organizations. And those personas are kind of representative of different sort of people that might be using your system. On the screen, I've just got a sample persona. Just found on the web.
Nothing in particular. Personas can vary as to what information they might include. Generally speaking, they are given a name, so that they can be referred to. They're given sort of some background information in this particular persona.
It's got a bit of a background to the person, and a few sort of factual pieces of information, like age, family, location. This one also talks about personality sort of elements. Introvert versus extrovert, analytic versus creative, loyal versus fickle, passive versus active. They can be particular to the individual, and maybe what the organization wants to focus on.
And preferred channels. How they like to, say, access content. Whether it be through web browsers, mobile phones, traditional ads, that kind of thing. Quite common to personas, though, are the goals and frustrations section.
So in this particular scenario, we have goals. This is around travel booking. So the goals of her persona is to spend less time on booking. Her frustrations was: Too much time spent on booking, too many visits per trip, and not terribly tech savvy, so she doesn't like the process.
So it's very common to see goals and frustrations in relation to what you're looking at. Now, from a task oriented walkthrough perspective, we've really got two options here, at least in my view. Where we can use some existing personas that are disability-specific. So being created, I guess, particularly to highlight or create personas that are based around different disabilities.
The second option we have is to enhance your own organization-specific personas. So like what we saw in the previous screen. With specific user needs. And in this case, since we're looking at the sort of focus around disability and the ways in which people might actually interact with our product and service, then those user needs might be specific to that.
So if you go down the disability-related persona path, the good news is: There are a few out there that we can just start with. You can rip off the shelf. The image here is gov.uk has seven personas, that you can access online. The other ones that I really like are personas from the book A Web For Everyone.
And they're also available online. They've got a lot of commonalities in them. But there's a few I'd love to combine together. But they tend to...
You can pick them up. They do tend to tell the person's story. They do have those goals and frustrations, but they tend to be related to the individual's goals and frustrations, rather than something they wanted to do, like we saw in those other personas. So the second way we can approach this is to enhance your existing organizational personas.
So let's say we take this sample persona again, Jill, and we could layer on top some different user needs or technology uses. So ability. Jill may be profoundly Deaf, fluent in Sign Language, but not English. She may like to use video chat to interact with Sign Language friends.
And chat to -- or interact with businesses like this as well. And in the technology space, she uses captions and transcripts on media. Now, there are definite benefits, I think, to taking the time to enhance your existing personas like this. The disability-specific personas are mainly focused on the person as a person with a disability.
And so they don't always have... Often the goals and frustrations of an individual -- let's say Jill wanting to book travel -- can often be exactly the same, irrespective of who a person is, whether they are Deaf, or are blind, or -- and so on. So when you do enhance your existing personas, it reinforces that, at the core, a person is a person. And it might just happen that they find themselves with a permanent disability, or situational or even a temporary disability, to, I guess, leverage what the Microsoft Inclusive Design Tool kit does really well -- which is outline that when we're dealing with these different things, we find ourselves in different situations.
And often when we are designing things, we are designing -- and it will help anyone in these different situations. So I really love this. But I know for someone who's just starting out that it might be just one of those accessibility personas, off the shelf. Which is why we've sort of done -- for us, when we're just quickly working with individuals, we might start with that.
And enhancing those personas, like that... Organizational personas, might be something that they do over time. Denis and Kaitlyn have an awesome video, online webinar, that talks about personas and unpacks that in a little bit more detail. So if you are interested in that, I would definitely recommend that you go and have a look at that.
So to help with this process, because I've used this in this process -- also in designer training, when I'm working with designers to build their knowledge in sort of accessible design, and teaching them how to do these task oriented walkthroughs, what we did is we created some persona cards based off the gov.uk accessibility personas. So on the front side of those cards, it has some of that factual information that we've pulled directly from the UK personas.
So here we've got Ashley. 24 years old. Her abilities are -- severely sight impaired, lost most of her sight in her teens, and it talks a little bit about the technology that she uses, her goals, and her frustrations. Then on the back, we've actually crafted these ourselves.
And popped down sort of a list of things that could improve Ashley's experience. So descriptive headings. There's a lot more text on the screen, but descriptive headings, clear form labels, meaningful links. Some of those design elements that we would look for to sort of look for in a design to say -- is this improving Ashley's experience as a screen reader user?
And we did that for all of the seven gov.uk personas. So we've got our persona group. We now identify user goals. So often these next two steps might actually have been already done, as part of your general UX practice.
Listing what user goals are. For example, a user on an ecommerce website. Their goals might first of all be to create an account, if they're new to the website. They might want to buy items.
But they might be different sort of paths within here. They might want to buy items -- do they want delivery? Do they want click and collect, where they order it online, but they can go and pick it up from the store? And it's all ready for them?
They might want to find their local store. If they want to go in and buy in person. They might want to contact the organization with feedback. Good or bad.
And I've got an extra task here. Find out about accessibility options. So given that we are looking at this through an accessibility lens, what if someone does want to find out about accessibility options? Does the organization have any accessibility options?
Perhaps if someone is having trouble with a website, often the first thing they might look for is an accessibility link to see -- you know, what commitment the organization has made to accessibility and also how they might want to contact the organization, if they're having any difficulties. Once we've defined those goals, we create an action list for each of those tasks or for those goals. So a very simple example here. Certainly not complete.
But if my goal is to buy items, then I might go to in this case -- we'll say the Coles website. We want to enter "butter" into the search field and select search. We're then going to find Flora butter, because that's what I'm looking for. I want to view the details of Flora butter.
I want to add that product to my cart. I then want to navigate to the trolley, and check out, and I want to proceed through the checkout, and complete that. So there's a bit of -- there's quite a lot of high level tasks here. If you were doing it, you would probably be a bit more granular.
You would flesh out the steps that you would need to do, to go through the checkout process. Within a goal like this as well, there might be different ways or different paths that the user might take. So here I've got -- I'm searching for butter, but maybe I want to browse through categories to find butter. That might be another path that someone might take.
So having those different paths that sort of users might take also makes sure that we're looking at all those different ways, and making sure all users have that choice, and in this particular task, reflecting choice in the action list. And our next step is to perform the walkthrough. Pretty simple. Adopt a persona.
Choose the goal that you're wanting to look at. And then walk through the actions, to achieve that goal as if you were that person. As if you were that adopted persona. And then repeat for other personas, for other goals.
Now, this particular process can be done. It can be done solo. I have done that before. Just sort of as a sanity check, to go through and actually step through some of these particular options.
The second way that you can do this is you can do it with a partner or partners. So you might actually want to grab a colleague, and be instructing them through this. So say: Hey, here's your persona that you're adopting. Here's this goal.
Now you walk through this process. And I'll sit here and observe what you say. So you would still want to use the same methods, like you do in usability testing. Asking the person to talk out loud.
About their experience. And if they're unsure about things. What they do at a particular point in that kind of thing. But that can be useful, because then you're more observing someone else doing that.
And you can get multiple people to do it. So that you can get multiple perspectives on this. The third way you could actually use this particular process is with an actual -- almost like a real person. And instead of adopting the persona, actually walking through the scenarios with them.
And I've done this in a semi-related way, where I worked with someone who was both -- had sort of quite a lot of accessibility knowledge themselves, but was also a screen reader. And the organization I was working with, they actually had some designs. And we would sort of walk through the designs with that person. At the time, because the designs were just flat, they had no semantics.
The person I was working with obviously couldn't fire up their screen reader, and go through the designs. But I could almost act as the sort of -- describing the interface, and asking questions, that kind of thing. So it's a little bit of -- I guess it's using this concept. It's not exactly the same as the sort of task oriented cognitive walkthrough.
But it was a really great sort of successful process. That I was involved in. As you walk through your process, one of the parts of the traditional cognitive walkthrough is also asking questions. So sort of...
Questions that will prompt you to think about: What was that experience like for the user? Was it clear to do X? I've got a few sample or example questions here. I've kind of morphed these questions along the way, and I do think that they sometimes -- they can apply to all users, but you do have to have a bit of a knowledge for what you're looking for, when you actually perform this process.
My sample questions are a combination of those that are in your traditional sort of cognitive walkthrough process, but also I've taken parts of, say, universal design principles as well, and trying to interweave those two things together. Because the great thing about principles like the universal design principles, or even our POUR principles from WCAG, perceivable, operable, understandable, and robust is: They're certainly more broad, than the WCAG which is sort of pass/fail. We can say: Is something easy?
Or is the information clear? And answers to those questions I think can be a bit subjective, particularly if we as practitioners are doing this process. So my sample questions here are: Was there an obvious path the user would take to complete the task? So examples here might be an obvious path.
Okay. Was it easy to see the button? So we could link that to contrast, for instance. For someone with low vision.
Obvious path... Was the control easy to find? And depending on which persona that we've adopted, does that mean that perhaps -- was the button easy to find within, say, someone's field of view? If they were a screen magnifier user, and zoomed in on the screen, and sort of could only see maybe a quarter of the screen at any point in time?
Was it easy to find that particular, say, button that we might be looking for? Was all necessary information available to the user for them to identify actions available? And which one was the correct action? So necessary information.
Stock standard stuff. Does your button have a label? Does your form field have a label? Did you provide information about, say, if someone was creating an account, and was filling in the password field, did you provide information about what the password restrictions are, or requirements are, before they filled it out?
Was that necessary information there available to identify the action and to actually complete that action? Then if that correct action was taken, will the user realize and understand that this is progress towards their desired goal? So is it clear that they have done something, and they have progressed, and whatever the next step is, okay... I'm still on the right track.
I've got that confirmation that... Yep, it's been successful. I'm now doing the next step towards whatever my ultimate goal was. And could the task be completed efficiently, with minimum effort?
And if we think back to our keyboard example, this is where... Okay. Efficiently, I'm not really sure that 87 tabs, if that's how you were navigating through a website, would be considered efficient. If we took the idea of, say, a persona that a screen reader user like our Ashley persona -- efficiently could be if Ashley likes to, say, navigate by headings, for instance, to quickly sort of work through a particular -- work through the page and find a particular section of content, or that kind of thing.
Then making sure that our headings are clearly described, and also marked up correctly, is going to lead to a more efficient experience for Ashley. So if we were doing this, with our persona, Ashley, this is the way we might go about it. So Ashley is 24 years old. Legally blind.
And uses a screen reader. This is from the gov.uk personas. Tips that we included on the back of our card for Ashley, descriptive headings, clear form information, having meaningful links, making sure any new information... So if you've got sort of alerts, or that is announced...
That we have text alternatives for images. That interactive components are accessed in a logical manner. A logical flow. Content flow is also intuitive.
So those two things always tend to be tied together. There are transcripts for videos. And modals work. Quote-unquote.
So that was just things that we popped down for people to keep in mind. So then if we did make it look at, say, our Coles website with Ashley in mind, our first step was to find the search field. So we would -- we've got the search field here. It's in the banner area.
Things we might check. Okay. Ashley. We talked about having clear form labels.
Is this form clearly labeled? And also, this is where, if you're doing this at the design stage, you can make sure that you don't have requirement steps within your designs. So when we're sort of talking about accessibility, there's a lot that we feel actually lives sort of with the designer. So things like making sure that form fields have descriptive form labels, how they linked up in the code -- that's more a dev thing, but making sure that those form labels are there and that they are clear.
So that could be considered a requirement. Depending on how much documentation your organization tends to like -- might be how much effort you put into writing down those requirements. But that could be something. So, you know, what's the order in which the content is represented on the page?
When it goes into dev? How many... What do I need to get through to get to this particular search field? So Ashley -- we look at here.
She's found the search field, we verify that we've got a place holder here that searches for projects. On the requirement side, as just mentioned, we want to make sure the form label is accessible, it's linked up in code, so that might be a requirements gap, if we haven't actually listed that, or at least there isn't a common understanding between the designer and developer, that that needs to happen. So right. We've found our search field.
We then want to enter "butter" into the search field. This is sort of an autocomplete, or a type-ahead, depending on what you want to call it. I've typed in butter, and there are a number of different options. Butter, peanut butter, unsalted butter, buttermilk, and butter chicken.
So how is Ashley going to interact with this? Is it clear what keyboard interactions we're going to use for this particular control? Knowing that there are standard keyboard interactions for controls like this, are we using things that are logical? That are consistent with Ashley's previous experience?
So let's say I can successfully select butter, press the down arrow, it reads out, and off we go. I now can see the next screen, where we now have our list of butter items. So we've got our first tile is an advertisement, our second talks about Coles unsalted butter, Coles butter pats, Coles unsalted butter, different weight, and so on and so forth. So Ashley was looking for Flora butter.
So how does she find Flora butter? Perhaps -- and this is where -- and I'll sort of talk about this a bit later -- is that one of the downsides of this particular process is knowledge of your persona, and in particular, in this case, knowledge of how, in this case, Ashley uses her particular technology, a screen reader, in order to navigate through the page. When I was sort of guiding people through this particular scenario, we've often done a lot of prework, to sort of talk about different ways that, say, in this case, Ashley as a screen reader user, navigates.
So although for most people, this wouldn't be the case, for me, my very, very first thing is: If you were just going through the content linearly on the page, to get to a particular point, Flora butter, what would that experience be like? As I acknowledged before, though, I do know that for many screen reader users, they use all those wonderful shortcut keys to jump to headings and buttons and links and that kind of thing.
So at least from my experience, working with screen reader users, I know headings are one of those core things. So I would say: Okay. If Ashley was navigating through this page, using that headings shortcut, then how would she go about this? And that would then lead me to question...
Okay. We've got all of these products. Would it be right to say that, say, Coles unsalted butter, which is one of our products, should that be a heading? Has that been marked up as a heading?
And is that heading clear, so that someone can navigate through? So even I -- and I am not a Native screen reader user. I do not have a disability. Even I am making certain assumptions, when I use this process.
I try, though, to base it on sort of my previous experience. Working or doing research, or usability testing with users. And some of those really common fundamental ways that someone might actually use to navigate through this website. So we found our Flora butter.
As I said, I'd be testing this. Okay? If I was going through headings, is that clear? What other ways might a screen reader user get to this particular information?
We then have clicked through to Flora butter, because our next step was that Ashley wanted to find out more about Flora butter. More about the product. So we've been taken to a product page that has Flora butter, the price, the fact that it's made in Australia, and a bit about Flora butter. A rich, delicious blend with buttermilk.
So I'm Ashley. I've got to this page. I've read through the information. Can I read through the information?
Is it in a logical order? And then of course, she wants to buy that product. So is it clear, then, from the interface how I'm going to buy the product? And we've got a button.
Add one for $3. And we can actually select that button and it's going to add to the cart. Things we might ask ourselves. Again, is that label clear, and in particular, in this case, is it clear for screen reader users?
Have we defined the role? So it is a button. It actually has a dropdown next to it, which I could use to add more, but I'm just adding one at this stage. So have I actually defined that this is a button?
And had passed those requirements on through to the developer as well? And as I mentioned, there's that extra control. So I could actually open up and add two for $6, four for $12, et cetera. So it's kind of a standard menu option, if I wanted to add more than one to my cart.
So thinking about that efficiency question that we asked. If we had that scenario where, actually, Ashley has a really big family and she wants to buy four Flora butters at a time, because she goes through that butter really, really quickly, how efficient is that? In this case, we could open that menu, choose those four for $12 fairly easily, making sure that it has all that lovely accessibility stuff that we need for screen reader users.
And go on. As opposed to -- not that I would hope this would happen, but having to click on the "add 1 for $3" four times. It might confirm that that has been added, and might need to refined this button again to add another one to my cart. That's clearly where these questions of efficiency will fail dismally.
So once we've done that process, and as I mentioned, you can do it as many times with as many different personas, you might get other users, or people within your organization, to do this, to get different views. We of course need to discuss solutions to these potential barriers. And I use the words "potential barriers" really deliberately. They are still for me potential barriers.
We are still using this process not as users. Just as professionals, wanting to use this as one tool in our tool kits. So right now, for me, they are potential barriers. So in this stage, and particularly if you've done in process, together with others, you could come together and think...
Okay. What are the potential barriers? For me, my questions, when I do this particular process, or even in doing reviews with teams, more design reviews, is: How confident are you about the potential barrier? And I say this because...
And yes, I'm drawing on quite a lot of experience here... But when I look at a design, and I see a potential barrier, sometimes I'm like... Yeah. I'm pretty sure this is going to be a real issue for a user.
Sometimes I'm like... Look. I don't feel like I have enough knowledge here to really know if this is going to be a real issue for the user or not. So I actually use that kind of gut check with myself.
To sort of say... What level of confidence do I have that this might be a potential barrier? And then of course, what do we want to do about it? So my ideal in this particular situation is: If you've got a barrier where you're pretty confident it's going to affect the experience, then I would absolutely be looking at how we can change things.
So depending on whether you're using this process in design, or even -- you can be using it in development, or for any existing site, even, that might dictate sort of what your action is. What you might be able to change. If I've got a potential barrier, but I'm really unsure about it, my biggest message that I say to everyone that I work with is: Let's put it up for usability testing. And so in those particular scenarios, sometimes I could go two ways about how we improve the experience.
If I'm really unsure, I might go... Look, let's just leave this as-is. Because even if we change it, I would still probably be guessing about the impact that's going to have on the experience. And we really, really need to get this tested with real users.
For me, that approach is -- it's a pragmatic approach. I don't want to make someone do something, change something, if the confidence level isn't there. So that's one possibility. The other one could be...
Look, I have a bit more confidence, but it might not be in that category of "pretty sure about this", and maybe you do make changes to the design at that particular time. So what are my takeaways from sort of coming through this process? Coming through to how we can use this task oriented walkthrough process in order to help us walk through our interfaces, to detect some of those issues early? So for me, the key thing about this process -- it is user focused.
So when we do adopt those personas, we are clearly keeping -- even if it's a fake user, we are still keeping that user in mind. We are trying to understand what the experience might be like. You know, for that user. The other key thing, I think, is it's task-focused.
When we do a WCAG review, it tends to be, for me, at least, very much more page-focused. Components on this page. Page-level information. We're not always thinking about the transition from one page to another.
So for me, they're the two callouts of this particular process. The second one is: It can be done on designs. As well as on existing systems. So if we go back to our Coles example, we could have used this process on the Coles designs, and realize that if we just followed kind of the natural visual order, and that's kind of how we ended up with that right bar being so far down in the content order that we had all those tabs to get through, we could have identified this a lot earlier.
And we could have said: Okay. We're going to define the focus order, and perhaps we actually go through the banner. We would then go through those right hand items of delivery times and login and those kinds of things, and then we're going to go to main nav. Because when you think about it, a lot of those items there that were in the right hand side are often visually in the top banner.
And if they were, we probably wouldn't have had a problem. So that just forces us to sort of take that time to think through this particular process. It can be done when you don't have access to real users. So as I mentioned, when you've got a flat design, you may not be able to do usability testing with a flat design at this point in time, although there are other methods like this, or co-design, that you can be doing.
But you can still do this. And then once it has gone into development, then you can test with real users. It can identify those requirement gaps. So if you haven't clearly defined what the focus order is, where someone is going to tab through, within those designs, then you've probably got a requirements gap.
For us, this is part of good interaction design. Which is why we actually say it sits with the designers. Rather than the developers. So that we can pick up these kinds of issues, and we can think about it from: What does the user want to do on this particular -- in this particular application?
And it's often said, this process, to be cost effective, fast, and easy to conduct. I would say yes and no. It can be -- you know, and that's in comparison to, say, other usability techniques that we might be using. It's definitely...
It probably doesn't take quite as much effort to do. And you could do it fairly quickly. You might want to, for instance, focus on personas, where you're a little bit worried about the experience, maybe, that you had missed something. And limit the personas that you're using, and really sort of focus on.
But when you think about it, if you had seven personas, and you had six key goals, that's still a fair few combinations to get through. So I think that "cost effective, fast, and easy to conduct" is definitely a relative term. But -- and I've said this, and I will say this again -- this is not a replacement for usability testing. I do not want anyone to walk away and think that.
You know, ultimate goal is always: Can a real user use this? Can they do things easily, in a way that they are satisfied with? Again, one tool in your tool kit. Not your only tool.
I still think it can be time-consuming. As I mentioned. When you do that sort of combination of personas versus goals. And particularly -- you'll definitely get a bit more robustness if you maybe do this process with a few different people.
If you're doing that sort of pair review. That sort of cuts down on some of the bias, or can cut down on some of the bias. So I still think it could be time-consuming. But I also think it's an invaluable way of actually looking at an interface.
And the big, big callout for this, which I did touch on, when I mentioned about -- I'm sort of drawing on my knowledge of accessibility and the different ways that, say, these personas might actually use their technology. Or things that might impact them. Your insights can be impacted. By the people who are actually doing that walkthrough.
So you want to keep that in mind. And you want to sort of... If you're using this within your own organizations, think about how you can build up different people's skill and knowledge, or provide some instructions at the start, that might help with this process. Some are easier than others.
Keyboard-only testing probably is -- a persona that's a keyboard-only user is probably a little easier. Obviously your screen reader user is a bit more of a complex persona. And that's just your technology-based ones. In those gov.uk personas, we've got Paul, who in the description from gov.uk has Aspergers, so that's going to be less about understanding technology, and more about the design of the interface as well.
So there are different layers and levels, of course, to those different personas, and understanding that as well. But ultimately, everything that I do is about sharing my knowledge, and people becoming accessibility champions. We were really sad not to make it to the CSUN conference. Or at least, not all of us to make it to the CSUN conference this year, from Intopia.
CSUN runs every year, over in Anaheim now. And we tend to go and always meet all our accessibility friends. And we started a trend of bringing badges. And this year, we had a Star Wars theme, and so I thought it was appropriate that for everyone here, there's Yoda saying: Powerful you have become.
The a11y I sense in you. Thank you.
THOMAS: Wow. Thank you so much, Sarah. That was an awesome presentation. And I guess I meant to call it out before you started talking, but I always love listening to you.
And your positive attitude toward accessibility. Very engaging. Thank you so much. So for everyone on the YouTube stream, and also here in our Zoom chat, this is now our Q and A period.
So for those of you on the Zoom, you can ask your question, or request to ask a question, and go off of mute, if you want to ask your question. With your video. Or audio. That's all great.
We love participation. I will also be monitoring the YouTube channel, and ask those questions to Sarah on your behalf. For those of you on YouTube. I wanted to kick it off, while we're waiting.
But I was curious, Sarah, for the section about actually writing out the steps for a user task, do you have a heuristic, once it gets to this number of steps, you should make multiple tasks, or I'm curious how you approach the number of steps.
SARAH: Yes. We don't have a hard and fast rule about that. I think that you can sort of... I sometimes do break it up into subtasks.
So for instance, when you're putting something into your cart, if we use that example, you do have to log in first. So you might break up the login task as maybe a discrete step that you could test. That is within the context. And then what you would want to make sure is just that handoff, though, between -- I have logged in, and now I'm going about, and, say, searching for a product, or that you would want to make sure that if they were separate tasks or goals, that that handoff process was still tested.
But those modules could almost like -- in my mind, I'm seeing them as little modules, and they're kind of back to back. It's making sure that that connection -- you know. Between one sort of module to the other is right.
THOMAS: Great. Thank you. We've got Andy McBride on YouTube saying: Great presentation, Sarah.
SARAH: Thank you!
THOMAS: Would anyone like to ask a question?
Hi. It's Aurelia. I'm just sorry if I just don't know this any better, but... Screen reader users versus keyboard-only users...
Can you speak a little bit more to that, please?
SARAH: Yep, absolutely. So in my view, both are keyboard users. So when I talk about a keyboard-only user not in the context of a screen reader user, I'm more talking about a sighted keyboard user. Someone who can see the screen, but predominantly uses the keyboard in order to navigate around.
By comparison, and if we're just focusing on sort of desktop screen reader usage -- although you could hook up the Bluetooth keyboard to a mobile device as well -- then it's a keyboard-only user -- is the input method, but I guess the interaction is coming through the screen reader. So obviously information is being announced. Form labels, form fields, all that lovely sort of semantic stuff. That a screen reader user needs, in order to be able to navigate around a website.
So it's not anything to do with mobility? You were talking about vision?
SARAH: So a sighted keyboard user, generally speaking, is probably to do with mobility. So they might have a mobility impairment. But still be looking at the screen. Whereas a screen reader user tends to be someone who is blind or has low vision, and is using that technology.
Okay. Thank you very much.
SARAH: No problem. Andrew has asked: Should I be testing my sight with eye tracking software or other similar assistive technology? Well... That's an interesting one.
When we're looking at sight, it's a little requirement that people often miss with WCAG, which is to get technical conformance requirement, for which I paraphrase as: It works with commonly used assistive technologies. And that's actually part of a WCAG review that often people get caught up just sort of looking at the WCAG success criteria. I know of a few users that actually used eye tracking software, in order to navigate. As opposed to keyboard-only.
Because -- and/or voice recognition. I personally haven't tested things with eye tracking software before. And that's just because I haven't had the opportunity to do so. So I think that anything that we consider a technology, and particularly if we felt that there might be certain things that don't work well for that particular technology, that we should be testing with.
THOMAS: Okay, Sarah. We have a question from YouTube, from Micronaut. As a designer, we're often in Sketch. Any tips on incorporating accessibility thinking into the design process, outside of contrast?
Basically what can be done when the product isn't live?
SARAH: Yeah, absolutely. So I think there's probably two parts. There's the sketch part. And there's the -- what can we do just as designers, irrespective of whether you're using Sketch or InVision or any of those particular tools.
I know that Sketch has some plugins, as do some of the other designer tools that might allow you to -- they do have, yeah, inbuilt color contrast checkers. But there are others that allow you to annotate your designs with things like content order, tab order, and so on. And if not, you can kind of make your own sort of pseudo-way of doing that. And actually communicating that over to your developers.
So from a tool perspective, that's probably certainly one thing that can help those sort of annotations. Having said that, you have to think through, of course, the content order and the tab order, and that kind of thing. Like in the way that we've kind of been talking about, to make sure that it does -- it is logical and it's probably working for the user, not against the user. So that's kind of one thing that you can do.
Other annotations, like even with content designers, however it is that you like to do it, making sure that they provide the alt text, that shouldn't in my view be left to the developers. The developers should just be able to say: Yep, I've got this image. I've got alt text that's been written. With sort of -- what's the purpose of the image in mind?
That kind of thing. Or flagged as decorative. And then they just take that, and implement that crafting of the alt text, to me, sits in design land. Whether it's context design, UX design, depending on the size and roles within your team.
THOMAS: Great. We've got time for one more question. If anyone's got a question. Kate?
SARAH: Kate, you're on mute. Can you unmute yourself? There we go.
So I thought something you said was really interesting, about how a user -- if a person with a -- hits an accessibility problem, their next step might be to look for information about the site, and whether or not it's accessible. Do you have any recommendations for a minimum level of accessibility? What someone should put in that kind of situation? Is that in the footer?
Is it someplace else? What are some ideas of how to handle that?
SARAH: Yeah, absolutely. I think there's two parts to that. There's the: Where is the information provided? Traditionally, and there might be people on the call that might have more personal insight to this, it does seem like the accessibility link is often included in the footer.
That's a very common pattern. And I would imagine people have probably gotten used to that. I have seen the odd site that actually does include it in the skip links. So a skip link, directly to the accessibility page.
Because those skip links tend to be more for sort of accessibility-related matters, having said that, a skip link will only surface if you're a keyboard-only user. So it's certainly not a foolproof scenario, from that point of view. In terms of what information is included in an accessibility page, there's a few templates out there. So the web accessibility initiative has sort of a rough template of what you might include in an accessibility statement.
There's two accessibility statement generators that can sort of start you off with sort of -- what's our commitment to accessibility? Making sure that there is clear contact for people who might want to contact the organization. Good or bad. I'm always really clear.
It's always lovely when someone actually messages you through a channel like that. And says: Hey, actually, your site was really great to use! Thanks so much. So there is sort of some fundamentals that you rarely include in accessibility statements.
The ones that I love, though, are the live ones. And actually, Coles is an example of this. Where they do list known accessibility issues. So a recent example that I could use, which is happening right now, but is not actually -- unfortunately the organization I was working with doesn't have a live statement.
At the moment, there is a very particular bug when JAWS 2020 users are using Firefox. So one of the recent updates to Firefox created this massive bug, where on certain websites, the tabs crash it. And it's really not something that's in control of the organization. It has to be fixed by Firefox.
That would be the ideal thing for me. To include in an accessibility statement. Of going... Hey, we know that there's this issue.
And in this case, it's not in our control. But use a different version of... You know, say, if you do use multiple screen readers, use a different screen reader. Use different browsers.
Because it's not happening with Chrome, for instance, and JAWS 2020. So you could provide that advice. The other thing you could do is, if you, again, have an issue where -- and this particularly, when the site is being designed to be accessible, and certain things just crop up all of a sudden, unknowns, or suddenly JAWS releases a new version, and someone discovers an issue with it, and you could list known issues, and then workarounds, so it might be a known issue, but perhaps it's gonna take you a week to action it, for whatever reason, or two weeks, you certainly don't want to leave that too long.
But in the mean time, you could be including that information in your accessibility statement, to say: Hey, we know about this. Here is a potential workaround for you. You might be able to do this in a different way. So for me, you know, ideally, I love accessibility statements that do that.
I often have -- not a lot of traction, though. Because legal is like "oh my gosh!" Sometimes getting an organization to even commit to putting up an accessibility statement... Because they have the fear that if they pop it up, it kind of opens them up to people suddenly...
It's almost like a focus on accessibility, and at least, all organizations I've worked with are still on that path of improving their accessibility... I don't think that should stop them, but obviously in a large corporate -- that can often be a bit more tricky.
Thank you.
SARAH: Thanks, Kate!
THOMAS: All right. So I think we will end the Q and A portion of the Meetup tonight. We can hang out in this room a little bit longer, if Sarah has time, and we can have some conversations, like we would normally have. If we were all meeting in person.
But we want to thank everyone that's joined us tonight. We want to thank our presenter, Sarah Pulis, so much, for the presentation tonight. So if everyone could clap.
SARAH: Thank you.