Noticing that two data points are corresponding
During the autumn of 2022, still not half a year in my career at Emarsys, my then team lead asked me to do a UX audit on the features our team owns. Not because he wanted to have new information; he thought it would help me become more familiar with our features. He wasn’t wrong.
But I noticed a few things.
One in particular is that the workplace switcher of the software was buried under a menu, which, considering established patterns, made no sense at all. I asked him why that is and what we can do about it, and after that discussion, I decided to check in with my Product Manager (PM) and my UX Researcher pair to gather information to learn what users think about it.
After looking into the written user feedback already gathered (and gathering digital dust), I realised that not only did my UX audit point out this problem, but we had quite a few complaints about it from users. However, nobody thought it a problem because the assumption was that very few people used it and not so frequently.
Analysing what causes the problem on the user’s side
After synthesising complaints, looking for patterns in the data, it became very clear that the assumption was wrong: the workspace switcher was used by more people than assumed, and also more frequently, while the switcher itself was buried under a user profile menu, making it hard to find and use. I analysed the findings with the PM and the UX researcher, and we found the following things.
For new users, it was hard to find and easy to forget where it was.
But it created more problems for everyone, not only for new users:
- More clicks were needed (3 instead of a maximum of 2). Though this does not seem like a big issue, considering the points below, it was.
- When users checked which workplace they were on by clicking on the profile menu, a lot of times they couldn’t see the whole name of the workplace, which was also an issue. It turned out that if a company used more than one workplaces, they usually seperated them by adding a few letters to the end of the name, which wasn’t visible!
- And more importantly, because it was hard to check on which workspace they worked, they made mistakes: creating marketing campaigns in the wrong workspace, or worse, sending out campaigns from the wrong workspace, which messed up their billing analytics, their marketing analytics, and sometimes even their brand image!
After compiling the insights, I instantly had buy-in from the PM, but not from the product leadership, so the project got deprioritised.
Figuring out what to do with this issue
But we didn’t give up, and prioritised the problems and defined the expected outcomes. We hoped that this would convince the leadership. And it did. We defined the following three expected outcomes:
- The workplace switcher has to be easily findable anywhere, so people know which workplace they are working in.
- The switcher has to let people navigate between workplaces easily and quickly.
- The names of the workplaces should be easily identifiable.
We later measured this by comparing the two versions by conducting usability tests and A/B testing.
Concerns, ups and downs while designing the solution
I looked for established patterns in other products to find best practices, and organised an ideation workshop with relevant people: the PM on the topic, the UX researcher, designers and developers from the Design System team, and a few designers from other teams.
The ideation workshop gave us a lot of ideas, but the team wasn’t convinced that either idea was good enough or safe to try out, even with testing.
After consulting with my team lead, who told me (since Emarsys was bought by SAP ~1,5-2 years before this story) that the central SAP Design System team might have a pattern for that, so I should look into it.
I did, and found a suitable pattern (which was almost the same we already sketched out during the workshop), then talked with a few designers at SAP about it and the issue we want to solve. It turned out that we were headed in the right direction!
This was the reinforcement my PM needed, so we proceeded with it: I redesigned by moving it0 to the always visible upper bar in my Figma files, and started to define some details: the naming pattern and the known use cases and error cases, while the PM and the Developers started to look into the not yet known error cases.
Our users had feedback on it
With the concept designed, we started to test it. We used two methods:
- We showed it to internal users who used or might have used the switcher a lot to gather feedback from them. (Note: we have an internal admin tool, so they do not always rely only on what is part of the public UI.)
- And we did an A/B testing using the designs in Figma to measure how the pace of navigation between workspaces changes.
The outcomes of testing
It turned out that:
- The colleagues were very happy with the new solution and relieved because they were very frustrated with the current state. Curiously enough, they never complained beforehand, because they started to rely on the internal admin tool for this task as well.
- From the A/B testing, we learned that the new solution was more than 50% faster to use, and from the written feedback, we knew they welcomed the change.
So I finalised the design details and the documentation, while the Engineering team started to work on the implementation of the new solution.
Life with the new workspace switcher
We stopped getting complaints about that part of the navigation.
Though the company took a different direction in 2024, so the navigation designed in 2023 will be deprecated soon, I see that this pattern could be reused once again for a technologically different workplace switcher at a later point.
I learned so much from this project, even if it took more than half a year to complete
This project was something that went through a few deprioritisations, and it had a lot of concerns around it, and I worked on other projects during the ~9 months of this project, but learned so much from it.
The two most important learnings are:
- Focusing on the target outcome helps a lot when we work together, because otherwise, anyone can start to follow their agenda.
- Design is even messier than how I experienced it before. So I learned to embrace the chaos and trust in the process, the team, myself and the tools I have.
