web navigation redesign
Role
UX Researcher
Research methods
Tree test
Skills
Synthesis
Output
Usability report to be presented to the team
Collaboration
Worked along with another UX Researcher, Product Manager
Tools
UserZoom, Google Slides, Confluence
Background
The structure of the web navigation is cumbersome and not well organized​
​
Objective
Evaluate the findability of content within the information architecture of an e-commerce website
​
How can we increase the findability of the navigation menu on flixbus.com and flixbus.de?
​
​
We focused on understanding the findability of finding these items from the homepage:
-
Tracking bus status
-
Get info about additional luggage policy
-
Discover destinations
-
Check for nearby stations close the departure
-
Contact CS for filing complaints
​
​
Tree testing is a technique for evaluating the discoverability of items in a website and is used for optimising the navigational structure. It is also known as a reverse card sorting.
We did tree test as the UX Research methods in order to evaluate:
-
Can users find items from the "flixbus.com" and "flixbus.de" page?
-
How confident are users that they have correctly located the items?
-
Are there any "Disaster" navigation tasks?
​
My research process:
1. Kick off
The Product Manager came up with the request to evaluate the navigation menu on the website. I asked him his hypotheses about which items that is confusing to find. He came up with list of items backed up with data from tracking too. We target 15 people to participate in the tree test for both .com and .de (30 participants in total). We were using UserZoom tool, so we recruited participants from their pool.
​
2. Prepare tasks
We first ask:
-
"Imagine you'd like to check your bus whether it has been delayed. Where would you click to find it?"
-
"Imagine you're thinking to contact FlixBus CS about a missing item on the bus. . Where would you click to find it?"
-
"You would like to get informed about the size of luggage you are allowed to bring. Where would you click to get this information?"
​
​After attempting each task, the participants were asked to rate how confident they were that they had found the correct location (1, Not at all confident - 7 Extremely confident)
​
We also did benchmarking with another bus travel brand sites in the US.
​
We run the test for a day, the next day we have received the results.
​
3. Synthesis
Using Disaster Ratings
By combining the percent of participants that found an item along with how confident they were, it is possible to create a 'four-block' diagram.
Ideally the tasks would fall into the upper right quadrant (high success and high confidence).
​
Each of the other three quadrants represent failure, however, tasks where the majority of participants fail to complete them (i.e. Success Rates >50%), but are confident they have done (i.e, Confidence Ratings >4.0) are classed as a 'Disaster'.
​
​
​
4. Findings and Recommendations
Web navigation is not intuitive enough, readability & accessibility should be improved as majority mentioned aspects like colors, contrasts, font (US benchmarking supports this findings). Item wordings and categories should be more self explanatory & differentiable.
After, we redesigned the navigation on .de and we achieved these outcomes:
-
Findability of tracking the bus status improved from 0% success rate (old web navigation) to 81% success rate
→ wordings & category are more self explanatory now (Echtzeit-Info) -
Customer service findability improved from 17% to 35%
→ still not clear enough as the majority has been looking at “mehr Themen” -
Stations nearby success rate improved from 73% to 87%
-
Additional Luggage success rate slightly dropped from 80% to 71%
-
Discover network success rate decreased from 80% to 74%
5. Learnings
I found the tree test is very good in evaluating the structure of a navigation especially snythesising the findings using the Disaster Ratings, which was really pragmatic and eye opening for us, UX Researchers and Product Manager to take action for the next steps. We also have a follow up test for CS contact in another study.