UX Design

UI Redesign +
Achieving
WCAG 2.0

Improving accessibility and readibility
of Knowledge Articles


The problem:
Users were dissatisfied with overall readability and the business wanted to maintain brand consistency


The solution:
On the surface, this was a simple UI redesign.
However, it was made complex due to several accessibility issues and legacy code


The challenge:
Addressing the accessibility issues alone took several sprints.


Initial assessment

UI inconsistencies

branding, visual heirarchy, layout, line heights etc



The site in 2022

Initial research

I looked at articles, blogs and content across our platforms (product, main website, support, community etc).

I accessed PowerBI where I had the latest year's worth of customer feedback to look at.

I then interviewed the main stakeholder to understand how/why the articles are formatted currently and the challenges they faced.

Observations/Takeaways

Years of design debt
.

The current UI + components weren't up-to-date with brand guidelines.
Custom coded components manually added in

Easier way to show relevant content?

The articles were long because they covered multiple scenarios under a topic. This made it hard to read and it was up to the user to find the right info

Internal tool vs Front-end

Although the KB team formatted the article properly the global CSS would overwrite their code.


The KB site in 2022

Design from 2022, currently being built

New priority

WCAG 2.0

Accessibility improvements take precedence


The challenge:
Limited Dev resources and time meant we had to move UI changes to the backlog and take up Accessibility fixes.




The solution:
We worked with UXR who identified issues and we split them across our sprints.

I took on the role of accessibility SME within customer support.


JAWS/NVDA User Testing + Outcomes

23+ JIRA stories | 7+ Months

The current UI + components weren't up-to-date with brand guidelines.
Custom coded components manually added in

4 Check-ins to assess progress

We had to make quarterly ADA improvements and achieve WCAG 2.0
by year-end.

8 FABLE users + session recordings

We worked with visually impaired users who would provide feedback on what wasn't working. I was able to joining the last set of these sessions and invited my Dev team to sit it too.


Margaret Kuo walking us through observed behavior.

Certain issues took months to rectify

We missed UAT cases or new errors cropped up.

Some went through all 3 rounds of testing before being deemed usable. Like the chatbot, contact support flow and nested FAQs.

Some were easy fixes like test contrast, font size, DOM order.

Tab order followed the code DOM structure. A decision was made to skip the filter entirely.

The project was a crash course in WCAG 2.0 principles.

It was great collaborating with JAWs users and a revelatory experience as a designer.

Conclusion

Lessons Learned

Lessons learned:

CSAT scores aren't always a reliable indicator of quality

90% of our users left no feedback. The articles had a deflection rate of 80%. A small subset of detractors left -ve reviews.

Workshop outcomes depend on the group invited

A workshop I conducted to gather ideas from different SMEs (outside KB) meant I was talking with people who didn't have in-depth knowledge of the problem.

Accessibility should be baked in from the start

It is harder to retroactively make ADA compliant product/website fixes.
We kept this in mind while designing future pages/components

It was great collaborating with JAWs users and a
revelatory experience as a designer.

This project helped me establish a great rapport with Devs

The back-and-forth and daily conversations helped me grow as a designer and helped share the knowledge with the rest of the team.

Conclusion

What I would do differently

What I would do differently now:

Pitch UI improvements better

Outling impact before submitting a project request so it doesn't fall under 'cosmetic changes'

Call for simpler solutions to unsalvageable code

The contact support flow was ultimately too flawed to be improved by a significant margin.

Next project