2023, UX INTERNSHIP

Creating Commvault's first design system site for PMs, developers, and designers

7 Minute reading time

Overview

I designed a new internal site that empowered consistent design output at the level of the individual contributor. The project unified dozens of conflicting design paradigms, established a much needed source-of-truth for design systems, and set the company farther along on its goal to make Commvault's product more cohesive.

My role

I contributed to the overall design process including user research, ideation, prototyping, and user testing. This design system site is an ongoing project that I helped jumpstart.

Project scope

1 UX designer, 1 developer, 1 design director, and 1 manager
May 2022 - May 2023 (1 year)

Tools

Figma, Illustrator, Photoshop

A customized designed system site for PMs, developers, and designers ensures that teams are better aligned during the product development cycle.

Setting goals

User research goals

1. Identify teams' pain points during projects

By conducting user interviews with PMs, developers, and designers.

2. Identify key features in design system site

By user testing the previous design system site.

Stakeholder interviews

Preliminary interviews with teams confirms hypothesis that early scope definement ... (add a section clarifying, when teams say they want a clear user experience defined ahead of time, what do they mean? (a visual guide), early communication,

We began our research process by taking a step back and first collecting stories from PMs, developers, and designers about frustrating and successful experiences they have had during projects. We interviewed a total of XX people and .... We summarized our stakeholder's stories and mapped out their general experience during the product development cycle to vizualize where in the process people were experiencing the contention.

A visual guide

🧾

Users enjoyed receiving a physical token of their experience.  

Formal user flow

👭

Users felt that the information they shared and received were an appropriate level of personal.

Early communication

⏱️

The line moved more slowly than expected; the more succinct the kiosk is, the more users it can reach.

A clearly defined user experience means...

HMW align UX expectations early in the product development cycle?

User testing #1

Preliminary user testing a previously designed MVP tool designed to solve the issue, to see if the tool was useful to people.

The template tool was the most developed feature on the site designed to identify scope requirements and spit out existing templates. At the time, It used a wizard to gather those requirements and had a basic results page that provided links to existing examples within Commvault that users could click on and navigate to.

User testing insights

What we learned...

1. The wizard may be obsolete

people generally know what workflows they're looking for and the wizard is more of a confusing hurdle in their way.

2. High cognitive load

A little honesty and vulnerability goes a long way when creating a meaningful interaction with someone else.

3. Has potential to be useful but it has confusing marketing

A little honesty and vulnerability goes a long way when creating a meaningful interaction with someone else.

Ideations

Playing around with ways to modify the template tool to better suit user needs.

Based on our insights from our surveys, we hypothesized that music might be a good gateway to bond people who are unfamiliar with one another since music is such a universal media and gives the right level of personal exposure without invading privacy.

Compiling insights

What's working and what's not

What's working:

It's easier to strike up a conversation with someone when they share a common interest because it's an easy icebreaker.

What's not:

A little honesty and vulnerability goes a long way when creating a meaningful interaction with someone else.

User testing #1

Then we conducted a site-wide (previous site + reworked template tool) user testing, which revealed that the tool we put the most effort into people didn't even notice, but the testing revealed important insights about how people naturally navigate the site.

Since our goal was to create positive interactions between strangers on a commute, we asked people to tell us their stories about when they felt a positive interaction with a stranger.

Use testing insights

What we learned...

2. No one gravitates to the template tool

A little honesty and vulnerability goes a long way when creating a meaningful interaction with someone else.

1. Example page is most important

It's easier to strike up a conversation with someone when they share a common interest because it's an easy icebreaker.

Testing hypothesis through field study

Prioritization matrix helps up narrow down essential features for site MVP.

During our interviews, we heard a large variety of needs and wants from people. Knowing that we couldn't satisfy everyone, based on our interviews and user tests, we sorted ideated features into a matrix of most desired and highest effort to help us narrow down the most necessary features.  

Setting goals

Design goals

1. Figure out what to do about template tool and example pages

By conducting user interviews with PMs, developers, and designers.

2. Build out components section

By user testing the previous design system site.

3. Build out design guidelines section

By user testing the previous design system site.

Testing hypothesis through field study

Bringing the template tool's intentions to example pages

Since our user testing revealed that people gravitated towards example pages when using the design system site, we brought the intentions of the template tool to example pages. Since people generally came to the site with a general idea of what workflows they were looking for, we got rid of the requirements wizard and instead put all the example workflows we had in the side navigation so people could flip through them faster.

Ideation

Building out components section, prioritizing most necessary features

We then began ideating some potential ideas and brought to the table sketches and wireframes. We came up with both physical and digital solutions, and ultimately settled on two main possible directions that we could take.

Reiterations

User testing to improve navigation and legibility

We wanted to test people’s reactions to the interactions we were facilitating in our first idea. To do so, we used google slides to create deck of music profiles and google forms to moderate the likes, recommendations, and comments being sent.

Testing hypothesis through field study

Six rounds of card sorting to determine design guidelines navigation

Our card sorting activity taught us that PMs, developers, and designers all categorized our design guiedlines a little differently. We extracted the most common patterns in categorization and from there created our categories for the landing page. and as always, the full list is expandable under the side navigation.

Project Takeaways

There are no dumb questions

This project was such a fun learning experience because

Project Takeaways

If I could redo this project, I would change...

1. Test the more creative option first
2. tackle it from a business standpoint: spotify collaboration
2. furter prioritize efficiency: we identifies that the piec of the experience that people liked the most was the printing and receiving on a tracks card, if we can make the adding sog process even more efficient i would go back and do that (identifying the key feature of the product and maximizing it's reach)

The problem

Commvault's user experience was inconsistent.

Why does this modal use radio buttons while another uses checkboxes? Or why does the configuring workflow launch a wizard here, but a form elsewhere? There were inconsistencies across Commvault’s product from a component to a workflow level, and Commvault’s teams were looking for ways to ensure their projects were consistent.

The problem's source

Commvault's teams had no centralized source of truth that explained the “why”

Developers used storybook to reference for components, designers had figma files they could reference (but this was a very new thing for them), but there was no reference for workflows or patterns. No documentation that explained why certain workflows and patterns were the way they were, and, their product already had so many inconsistencies that no one knew what to reference when working on a new project.