Co-op Insurance talked about usability testing with people who are visually impaired last week on the Digital blog.   

Improving and influencing better accessibility where we can is important. This post describes how we prepared for the sessions. We hope it encourages more product teams to test with people with a range of access needs.  

1. Charities can help you recruit

We’d found it challenging and time-consuming to find participants who are visually impaired through recruitment agencies so UX designer Paul Braddock made direct contact with the Royal National Institute of Blind People (RNIB). Although it took a while to get approval for our post which asked for participants on the RNIB’s social media page, the number of respondents was worth it. Charities and specialist organisations that have a vested interest in – and access to – a group of people you’re trying to find seem to be very willing to collaborate.  

 2. Ask participants to bring their own device 

Observing someone using your product or service on their own device gives a more accurate indication of how they would interact with it outside a research session.  

When we asked our participants to bring their own tech, we learnt a lot about the additional software they used too. For example, one participant brought their laptop and showed us Dolphin Supernova – a magnifier and screen reader they use to zoom in on a page, read it aloud, and replace colours that are difficult for them to differentiate between. They told us they “can’t function without it”. But, if we hadn’t asked them to bring their own device, we would most likely have asked them to use one of our Macs which Dolphin Supernova isn’t compatible with. In that situation, we’d have missed out on seeing our service in a realistic context.  

3. Send digital consent forms

We sent out digital consent forms through Consent Kit before the sessions so that participants could take their time reading them with their assisted technology and understanding what they were signing up for. We knew that paper forms would likely be more time-consuming and less preferable. We also couldn’t anticipate what problems there may be with talking through and then signing digitally in the sessions so it felt important to sort out consent beforehand. 

 4. Talk about travel arrangements  

If you ask visually impaired participants to get to a venue, find out how they plan to get there and whether they’d like you to meet them off public transport. Paul met one of our participants at Manchester Victoria station. She’d never been to Manchester on her own before and told us she found big cities a bit overwhelming. They navigated the short walk to Federation House together and the chat on the way worked as an extra warm up to the session.
 

5. Allow for extra time

Factor in extra time for practicalities like travelling, accessing the building and keep in mind that participants might not instantly feel comfortable in an unfamiliar venue so your introduction may take a little longer while you help them to relax. 

A participant’s personal device may take longer to load or update than you’d expect too. One of the participants we met had specialist screen reader equipment that took a little while to set up on their mobile phone. Around half way through the session, they felt that because they were using the zoom so much, it would be easier for them to switch to a desktop device – they said this is what they would have done at home. Changing over and setting up again also took a little extra time. Seeing this sort of thing is really insightful though, so scheduling extra time means you won’t be tempted to feel like it’s an inconvenience. 

 

6. Go off script – you might learn more

Sometimes the thing you’re testing just won’t meet a participant’s accessibility needs and as demoralising as that is, it’s better to see those problems early. So, as with any usability testing, be prepared to change direction if a participant is struggling with a task because you’re still likely to learn a lot. 

We often found that participants used the service in ways we hadn’t anticipated so if an accessibility issue came up it made sense to discuss straight away, learn from it, and then move back to the script. For example, one of the participants we spoke to zoomed into pages by default. A lot of what we discussed in the session wasn’t in our discussion guide, but we were still getting useful insights. 

Testing with people with other types of access issues

So far, we’ve only run sessions with people who are visually impaired. Of course, there are many more types of vulnerable user and testing with a range of needs is important. This is a good start though.  

If you’ve tested your product or service with people who are visually or hearing impaired, or have a motor or cognitive disability, we’d like to find out what considerations you had before running your sessions. Share it in a comment below or tweet @CoopDigital. We’ll keep it in mind.  

Catherine Malpass 
Lead user researcher 

This is a guest post about Co-op Insurance and their website, co-opinsurance.co.uk. The author is Paul Braddock, a user experience designer who works on it.

In the summer, we blogged about how and why we asked AbilityNet to carry out an accessibility audit on the Co-op Insurance website. The post explained the content, design and technical improvements we made off the back of it. To continue our work around better accessibility, the Insurance design team recently did usability testing with visually impaired users. 

In this post we share what we learnt from the testing and how we’re making improvements. 

What we tested

The Insurance team is responsible for co-opinsurance.co.uk. This includes managing the product information and financial promotions for each type of insurance. Buying journeys begin on our site but once a user chooses ‘Get a quote’, their online journey passes over to a partner’s site.

We can’t necessarily fix the accessibility issues we identify on a partner sites, but we can influence them. We’re actually in the process of creating a set of accessibility guidelines that cover our expectations of our partners.

Catherine Malpass from the Co-op Digital team and Louisa Robinson from Insurance ran the usability sessions. We focussed the testing on the travel section of the website including the buying journey. We wanted to look at how easily visually impaired users could:

• navigate the journey
• understand information – particularly in tables
• interact with a pop up

We kept what we’d learnt from the audit in mind when we were considering what to test. 

Positive things that came out of the testing

During the testing, we heard positive feedback about:

• the in-page navigation and clear labelling which helped users read and identify content quickly. One participant said it’s “fairly easy to find your way around and there’s a decent menu I can use to navigate to other insurance products.”
• the colours on the homepage because – along with icons – they help users differentiate between the 2 products. “I like that the colours for car and home insurance are different,” said one participant. “I like having the icons next to the product names because they make it obvious and easier to navigate.”
• the consistency of the page layout. A participant said that “the pages are similar all the way through so you can memorise where things like the next and back button are.”

Areas that need more work 

However, the testing also showed us that some features weren’t compatible with screen readers. We noted them along with recommendations for how to improve them. 

Things we’ve iterated on already include:

1. How a screen reader reads out the prices in tables. We’ve changed the aria labels so that monetary values are read out how we’d naturally say them, for example, “thirty-three pounds eighty-four” instead of “pound symbol three three full stop eight four”. We don’t feel we should be asking people with visual impairments to work harder to use our service and we think that making this change will reduce some of the cognitive load. 
2. How well we use alternative text (alt text). Testing threw up instances where images were tagged with the type of product or policy they were there to support. We’ve changed them so they fulfil their role of describing the content of the image. This makes the service more inclusive. 
3. Making images clickable. The screen grab of our travel product page below shows 3 insurance products (multi-trip annual, single trip and backpacker). Each has an image, the name of the product with a link that takes them to a more detailed product page, and a line of copy to help the user decide if it’s suitable for their needs. Testing showed us that when the user zooms in on the page (common for someone who is visually impaired), or if they’re using a screen reader, the link is difficult to navigate to and then difficult to access. There was an expectation that images should be clickable, so we’ve now made each image clickable which makes the information quicker to find. 

Things in the backlog we hope to complete before Christmas include:

1. How a screen reader reads icons in tables. At the moment it reads “tick” or “cross” but we’ve added aria labels so tick symbols will read as “included”, and crosses as “not included” on our next release.
2. Test alternatives to the pop up. We saw the screen reader struggle and the participants become confused with a pop up on the single trip online journey, shown in the screen grab below. We’ll be looking at how we can give users the same information in a more accessible way. For example, we may embed the content at different points in the journey instead. 

Things we’re encouraging our third-party partners to look into include:

1. The live chat feature.  This feature wasn’t compatible with screen readers so thinking about how to improve it or even considering alternative ways to communicate with users instantly is one of our recommendations. At the moment, the feature also covers up content on the page so we recommend looking into how to help the user feel more in control – this might mean giving an option to remove / minimise /block the live chat. 
2. How we present information. In the testing we saw users become a bit overwhelmed when comparing information. We recommend a review of how we display content when comparing price and cover in insurance tiers.
3. The postcode finder. Right now, when a user starts to type in a postcode, the screen reader repeats “there are zero results for this postcode” until the field is complete. This makes it difficult for users to hear – and check – which letter or number they’ve just typed. We recommend looking into how this can be improved.

Carrying on testing, iterating, improving and influencing

Accessibility should be a consideration right from the start when we design products and services. At Co-op Insurance there’s a lot linked to our site that we cannot control but we can influence. We’ll continue to test with people with access needs and we’ll keep trying to improve the experience for everyone so it’s as inclusive as possible. 

We’ll also share our research findings, audit feedback and blog posts with our external insurance partners to help raise awareness of the importance of accessibility.

Paul Braddock
UX Designer from Co-op Insurance