Google Cloud Platform
Self driving cloud
Cloud technology is complex. We wanted to leverage the power of Google Cloud's data, machine learning, and artificial intelligence to simplify over 5 million user's entire cloud experience.
To comply with my non-disclosure agreement, I have omitted and obfuscated confidential information in this case study. All information in this case study is my own and does not necessarily reflect the views of Google.
UX design contract onsite with Google Cloud
Rapid prototyping, wire frames, redlining, storytelling, cross-functional collaboration, design project management
I helped design and launch a new tool on Platform Intelligence, a cross-functional product team, consisting of designers, engineers, PMs, and researchers.
Learning the lay of the land
Before diving into work, I started by setting up lunch or coffee with people on our team to get an understanding of what they had tried before, what went well or didn't, and who I could go to for different types of assistance. This helped me to understand the team’s style and build relationships and mutual understanding for the coming months.
And during these conversations, our challenge quickly became clear.
We knew that the cloud computing industry had some serious problems.
Across all cloud users:
Cloud users' time spent on network troubleshooting
Permissions are over granted
VMs are over provisioned
We were asking people to make decisions that computers would be better suited to calculate
“Almost half of C-suite executives cited cloud complexity (47%) as the factor that will have the most negative impact on cloud computing’s ROI over the next five years”
-Deloitte Cloud Complexity Management Survey Results
But we knew it could be better than this
Our team functioned with one full-time UX researcher, two full-time engineers, one full-time PM, and two UX designers, along with a dozen folks augmenting those roles in part-time ways.
The senior UX designer focused on information architecture and user journeys. I focused on user flows, prototypes, user interface, and visual design.
Our project touched on almost every aspect of Google Cloud's platform. Our partner list was immense. Over the year I sat in on meetings with billing, security, analytics, storage, and dozens of specific product teams. Building trust and enthusiasm for the project was as important as understanding technical needs in each new connection.
I also sat in live testing with software engineers, systems administrators, heads of IT, and data scientists. Learning to understand the scope of their work was central to making sense of the features that would be relevant to them.
I dived into research and analyzed competitors and third-party tools by looking into their UIs, published reviews, internal user feedback, and studies Google had already conducted to prioritize the critical needs and most-demanded features for an automated recommendation system.
Terraform, Ansible, Chef, and Puppet were all examined in addition to Azure and AWS.
Design and prototype
I began to lay out rough UI patterns and worked with engineers to prototype the basic functionality early in the project. Breaking features into must have and nice to have was key. I discussed the overall product vision with PMs and senior designer, as well as prototyping and tweaking UI treatments with developers.
A couple dozen card layouts were tested with our users
(Redacted) results from 180 users
Partnering with our UX researcher
Our dedicated UX researcher helped me build tests that looked at different types of users and what tasks they wanted Recommendations to simplify for them. Variations of the how, where, and when of these recommendations were explored and documented.
Rapid prototyping let me take in one round of test results and adapt the UI to get further clarity out of the next round. Sitting in on the majority of the tests myself gave me nuanced insight into user's sentiments that spreadsheets alone couldn't capture.
The core user experience takes place over a recommendation hub, a list view, and a detail panel.
The main landing page features cards that bundle together similar recommendations.
Clicking into a card shows a list view of the recommendations bundled into that card.
Clicking into a recommendation opens a detail panel with more info and the chance to customize, enact, or dismiss the recommendation.
Advance to the list view
Narrows results down
Next / Previous
Advances cards or list
This prototype lays out the basic interactions for a user
Give it a try
Return to parent screen
Call to action
Make it so
Opening Recommendations to select customers for our Alpha release allowed us to iron out further gaps between customer's expectations and our performance. But the essential response from these key customers was awesome.
Million unused permissions removed
Major music streaming service
“This is by far one of the best things we’ve seen in Google Cloud.”
Major news publisher
“Networking security insights without us having to specify anything is magic, real opportunity for Google to differentiate.”
Here's my coworkers Chris Law and Dima Melnyk rolling out the results of over a year of my team's hard work.