Typing

Cloud Automation Experience

My Role: Strategy, User Research, UX/UI Design, Interaction Design

Welcome - Blank State_2x.jpg
The Problem

To receive the maximum cost savings benefits from CloudHealth, you needed to be a subject matter expert in cloud management. These Cloud Admins would spend hours each week creating reports that would visualize their savings opportunities and then manually take action on those opportunities on a separate cloud provider platform. For a novice user, or someone is not up to speed with the latest cloud technologies, achieving cost savings was painful.

Project Goal

We wanted to take an automated, proactive approach to continuously manage cost for our customers' cloud resources for them so that they can hit their cloud KPIs and related business goals while maximizing cost savings.

Research & Discovery Methods

This project was entirely remote. I worked with a Product Manager and a team of engineers all based out of India. 

Video Conference
  • Competitive Analysis

  • Customer Interviews

  • Personas

  • Design Workshops

  • Usability Testing

Research Approach

We knew automation would involve a high level of trust between the product and the customer. We started by contacting 6 customers to ask them about their potential concerns with automation, as well as their current pain points with managing their cloud cost.

  • Walk me through your current process for managing cost.

  • What are your struggles and challenges? 

  • Do you have any concerns with automation?

  • What would you need to see in the product that would ease any potential blockers from automating the cost savings experience.

  • Who is performing the activity of cost management at your company?

Mac Desktop
Research Insights
  • “Onboarding team members to CloudHealth is nearly impossible with the learning curve.”​

  • “I want my teams to be able to manage their own cloud cost.”

  • “Identifying cost savings opportunities and acting on those opportunities is my full time job” 

  • “Automation sounds like a dream, but we would need a lot of data details. I would want to test it out beforehand so  that I feel comfortable with the actions taken on my behalf.”​

  • “I'm considered an SME for cloud savings and I believe my process works well, but I would be interested in comparing numbers with your tool." 

Personas

Based on these interviews as well as input from our internal CAM/TAM team, I identified two personas for this project. We referred to them throughout the entire product development process.

Primary Persona

Jason, a Cloud Admin, spends countless hours trying to optimize cloud spend for his business unit. It's a manual task that he would like to pass on to the Cloud Analysts for each team, however they spend more time asking him questions on the complex workflow than what it's worth. He is frustrated but continues to do this task himself, and hopes that it's accurate. 

Personas - Automator.jpg
Design Studio Workshop

I organized a design workshop where I presented the problem & pains from customers, and we generated fresh ideas as a group. We focused on the goal of simplifying the onboarding experience

 

Given that the team was fully remote, I utilized Miro to organize everyone's ideas and our project goals.

Brainstorming.jpg
Sketches to High-Fidelity Prototype

I quickly sketched an onboarding workflow that aligned with our project goals and received feedback from both the project team and my design team. After that I jumped right into Figma and utilizing our design system, I was able to arrive at a high-fidelity prototype that was ready for user testing. 

PXL_20220405_172203874.jpg
Welcome - Blank State_2x.jpg
Onboarding - Step 2_2x.png
User Testing & Feedback

I contacted 6 customers that were interested in the tool and willing to give feedback. I drafted a script and asked each customer to walk me through the onboarding setup.

Overall, customers were excited to get this feature in their hands. However, based on our user testing we identified a few key changes that would improve the experience. 

prototype.jpg

Key Research Insights

  • Wizard setup was intuitive, but too lengthy.

  • Wireframes for the charts caused a lot of questions. Include dynamic charting visuals in next iteration.

  • Users found it cumbersome to edit their settings within a wizard. Limit to a single page form.

  • The informational banners caused hesitancy & concern. I changed these to standard tooltips. 

Improved Designs

I presented the key take-aways to my team, as well as improved design solutions to reflect the feedback. 

Onboarding Page

Welcome - Blank State_2x.png

3 Step Configuration

Settings_2x.png

Try Before You Buy - Dashboard before automation

SavingsAutomator.png
Measuring Success

A major theme of this project was rapid time to market. At the time of writing this, we have 10 customers in the Beta program for this feature. Based on customer interviews as well as metrics tracked directly from our Savings Automator feature, we were able to determine the following results.

  • Customers were able to decrease their overall spend by $40,000/month on average. 

  • 7/10 customers have turned on full automation.

  • Over 1,000+ actions have been automatically performed by CloudHealth through Savings Automator.

  • 6/10 customers have onboarded at least 1 other team member to manage their account's cost savings.

  • On average, Cloud Admins are saving approximately 10 hours/week on administrative tasks that are now automated.

Lessons Learned

Communication

This project was completed start to finish on a fully remote team. Being based in Boston and my team in India was a bit of a challenge in terms of time zones and getting quick turnarounds. Shortly after the project started we agreed on an ideal meeting time that worked for everyone and stuck with that time for the remainder of the project.

Gathering & Prioritizing Feedback

We operated as a scrappy agile team and focused on the goal of getting users onboarded quickly. However there were potential improvements for the dashboard experience that were overlooked in order to produce an MVP. We achieved the goal but could have simultaneously gathered feedback for the dashboard. Thankfully these initiatives are now on the backlog!