IBM Power Systems customers have higher risk environments that require extremely low-risk infrastructure. For decades, Power systems have met this need with trusted, enterprise class, on premise performance. Now, in response to advances in cloud computing, customers needed a way to extend their workloads into secure, flexible hybrid environments at their own pace and price point. Enter: PowerVS.
I joined the product team at its inception in February 2019 and contributed design research expertise on over a dozen landmark features for the next three years.
In early 2019, I worked with a crew of approximately a dozen offering managers, developers, and designers to build the product that would become PowerVS. When the product formally launched on the IBM Cloud Platform a year later, PowerVS had grown from a few dozen users to the landmark product of the IBM Power portfolio.
As a design researcher on PowerVS, I addressed a three-pronged research plan to create a sustainable and agile growth trajectory to keep pace with tight development deadlines and increasing business demands.
We replaced the beta product's rudimentary command line with a user-friendly, accessible, and visually effective interface. This involved wireframe construction, formative information architecture, front-end development, stakeholder buy-in, and hundreds of prototypes. Of course, all designs were informed first and foremost by research.
I conducted over 60 generative and targeted research interviews with the small population of early adopters. Structured interviews were conducted over video call and focused on use cases, feature requirements, and experiential feedback. Developers and offering managers were invited to listen in and contribute.
The 4-person design team focused on creating a sustainable user ecosystem with robust lines of communication between sponsor users and the product team. This initiative continues today and is now a network of 100+ partners, users, and contributors. Design partners participate in usability tests, surveys, interviews, and workshops to co-create a truly user-focused product.
Carbon Design Interface
The final user interface was developed using IBM's Carbon Design System, leveraging both stock and customized components for a robust, accessible workflow.
"Just get IBM Power Systems Virtual Server. It's worth it! The main point of us using IBM power systems is to extend and expand our IBM product workloads. It has been instrumental in solving some major business problems at my company."
- Matt Steggell, Bamboo HR
Everything's a Prototype
Design is an ongoing process. After PowerVS launched on the IBM Cloud Platform, I continued to work on the product as a design consultant and subject matter expert. I led a growing team of over 40 cross-disciplinary experts on design research initiatives to continually improve the evolving PowerVS user experience.
The following case studies outline my work as PowerVS matured into the go-to turnkey product for hybrid Power workloads.
Case Study: Networking Redesign
The original layout for networking on PowerVS was adopted from PowerVC, a related precursor platform for virtualization management. As PowerVS reached a progressively wider breadth and depth of workloads, networking capabilities fell short of customer requirements. I was brought in to conduct research on user needs and formulate a redesign of PowerVS’s networking services and flows.
I conducted market research, competitive analyses, and user feedback sessions to generate a ranked shortlist of feature requirements. I used lo-fi wireframes to present preliminary findings to subject matter experts and internal management to inform sizing and feasibility for each area of focus.
I worked with a front-end developer and engineering team to create and iterate on a design prototype of the user flow. Research also identified subnets as a problematic dependency, which we were able to fold into our roadmap planning and correct at the ground level. Progressive prototypes were tested with 9 users via contextual research interviews hosted on Invision. Insights were synthesized on sticky notes, Mural, and Microsoft Excel.
I played back research methods, key insights, and final designs to the larger team of stakeholders and executives up to the VP level. (Get in touch if you'd like to see the full presentation). The work was met with resounding approval and the feature was slated for development the following quarter. Following its release, PowerVS experienced a significant uptick in edge use cases in several unexplored market sectors and a boosted NPS rating.
Case Study: Loadtimes
I realized that we had an issue with PowerVS page loading times, with several pages taking almost a full minute to completely appear. I surfaced this issue to the product team in a presentation explaining the user metrics associated with load lag and correlational user dropoff, which gained viral status and ultimately led to a significant overhaul of the product’s backend.
Research + Data Collection
After a feature kickoff meeting, we began research using Amplitude for digital optimization analytics. With Amplitude, I was able to connect the PowerVS backend to generate easy-to-understand data visualizations with anonymized user metrics. I segmented user groups to identify common flows, associated wait times, and user attrition. Results were consistent with my projections.
I scaled back the product-wide data collection to focus on several high traffic user flows. I migrated this (frankly, overwhelming) amount of data to spreadsheets and synthesized key data points and trends onto charts. Storytelling was key. Collaboration with the development team was key at this stage as we explored triage options for the near-insurmountable scale of the problem at hand.
With realistic constraints in mind, I proposed a minimum viable product (MVP) to address load time reduction on the pages that we identified as most pivotal to the user experience.
Since we lacked the time and resources necessary to overhaul the product’s backend, we proposed a more realistic, immediate plan to add loading indicators and improve error messaging. In user testing, these steps were proven to restore user confidence and make for a more robust event management UI.