Took initiative and transformed the Information Architecture of an Enterprise SD-WAN Network Product, which significantly improves the entire product content taxonomy, discoverability, and usability and drives the product vision.
Lead the end-to-end design:
Product Vision
Workshop
Heuristic evaluation
User research
Design
Design QA
1 Product Manager
1 Junior Designer
2 UI engs
Design and research: 3 months
Created in 2021
Releasing in March 2022
Would like to skip the design process? Check out the final design and recap directly.
Prisma SD WAN, formally CloudGenix, was acquired by Palo Alto Networks in mid 2020. SD-WAN is a software-defined approach to managing the wide-area network. It takes user and application centric approach to connect users to the internet and applications. It is an autonomous network system, leveraging ML and data science, reducing manual configuration and troubleshooting.
Since I was new to the product when I started the IA redesign, I firstly met with product experts to understand the target users and their product usage and tasks.
From heuristic evaluation and user interviews, I found the illogical organization and multi-layer hierarchy create usability problems. The information architecture problems become more and more obvious as the product scales over years, the new features are difficult to fit into the existing model anymore.
From user research, I found that users are struggled with the poor Information Architecture. Based on research insights, I summarized a few user goals for designing an IA with better usability.
Since Prisma SD WAN is an acquisition by Palo Alto Networks, the navigation structure is also different from the Design System standard and other products. for better product consistency, Prisma SD WAN should update the navigation and try to align with the standard.
I created the site map and conducted card sorting exercises with PM, TME, and SE teams, this exercise really helped me gain more product knowledge and understand how the SME (Subject Matter Experts) mental model of the Information Architecture. Below are some results of the card sorting exercise. Option 2, which contains 2 major categories: Monitor and Manage, is well received since it takes into account the defined target users and the respective solutions to their problems and motivations as well as the objectives of the organization.
Simply regrouping the existing navigation items is not enough, user workflows and Information Architecture should work in tandem and dig into the information grouping under each page is important. From previous user feedback, I discovered that the “Monitor” related information grouping is not efficiently assisting the user to complete their tasks and goals.
To better understand the existing "Monitor" related user cases, I facilitated cross-functional teams interviews with SEs, TMEs, and Internal users, gathered user stories and asked them to walk me through their current workflows in the existing UI, then I identify pain points and iterated to the ideal workflows.
Here is one of the examples:
As a Network Operation Engineer, when someone reported a problem from a site, I need to check all the circuits and paths performance (along with other site level metrics) on the site, so I can determine the cause of the issue.
- Insight from cross-functional teams and user interview
The information architecture seems simple in the existing workflow, however, the interaction under the "Activity" view is overwhelming and requires a lot of clicks on the filters (filtering each application, circuit, path, device, interface) to get a comprehensive view of a site.
The information hierarchy is clear and logical. All metrics about the site is summarized at one place, with drill down capability to details.
Similar to the example above, I mapped out other typical ideal monitor workflows and iterated the Information Architecture accordingly.
Since the changes impact the IA and user workflow, to validate, I created prototype and run usability testing with users, and compared the time on tasks. Both qualitative and quantitative data shows that the new "monitor" taxonomy has better usability and efficiency.
After the card sorting and user flow mapping, I created a tree testing study with 8 tasks based on various use cases on the portal to test the new Information Architecture. The tree testing study was sent to 500+ recent portal login users, 12 responses gathered:
The new Information Architecture has clear and logical grouping and taxonomy based on user tasks and scenarios. The visual hierarchy and content is clear and straightforward. Some terminology is also updated to be more understandable. The navigation component is also aligned with the Design System.
Provide guides for the first login after the release, help user better understand and quickly get onboard with the new IA.
This project was not an initiative from the PM team nor on the roadmap, I started this project in order to achieve better consistency in navigation component with other products. However, I found lots of usability and discoverability issues with the existing Information Architecture. So I took the initiative and drove the efforts through research, ideation, and testing, which leading to a product level Information Architecture and content taxonomy makeover. It also took me couple rounds of discussion to have the leaderships buy-in this new direction.
The leadership later realized that this IA is essential for the long-term product vision- SASE Integration (Merging Security and Network products) , and it eventually becomes the top priority on the product roadmap.
The product team was not familiar with user research, I educated the team about the benefit and importance of user research, drove the research process and involved the cross functional teams closely, and proved that research would result in better outcome.
The old IA was designed by the VP of Engineer of the product, it took a lot of persuasion and pitch to have him agreed on this new IA. Instead of technology and features, I drive the conversation to focus on users and their workflow, along with user research insights to support.
