Took first-run workflows from frustration to flow - elevating success rates from 14.9% to over 70% through user-driven design

The Azure Quantum product team had a usability problem. Telemetry insights indicated a steady pace of attrition through onboarding first-run workflows with over 85% abandonment rate.😖 However, through telemetry the product team couldn’t explain why users were dropping at key points throughout this process, that’s where my UX research comes in. My hiring manager informed me that my design direction and UX outcome was to mitigate abandonment in onboarding “first-run workflows” — experiences that allow users to set up a workspace and submit their first quantum computing or optimization job.

I assisted with a product management user research study for first-run workflows within the first two weeks. This was a great introduction to the current experience journey with 9/9 research participants – spanning quantum Developers, Researchers, Decision Makers, and Consultants – failing to set up a workspace 😖 utilizing the existing workflows and supporting documentation. Without setting up a workspace, users wouldn’t be able to submit a first quantum job or access any of these other experiences.

It was my job to dive into the details, explore points of friction, identify opportunity and design a more intuitive and effective end-to-end first-run workflow outcome.

From this initial study, I broke down onboarding “first-run workflows” into four projects and conducted targeted research studies to isolate the key decision points for each user journey. These projects together constituted onboarding “first-run workflows".

Through iterative user research, cross-functional collaboration and data-driven design, I designed holistic solutions that drove the success rate of first-run workflows from 14.9% to over 70% within a six month period. The design decisions are data-driven solutions informed by telemetry (quantifiable usage data) and qualitative experiential research.

My role:

User research | Design | Design socialization and workshop leader | Cross-collaboration with Quantum Researchers, Product Management & Engineering teams | Alignment with Azure Product Management, Branding, Marketing, User Research and Design teams.


1 - Simplified Azure Quantum workspace creation

CHALLENGE:

There are 4 main pages to Azure Quantum Workspace creation: Basic Information, Providers, Tags, Review + Create, with a number of incremental steps within each page. Telemetry insights indicated consistent attrition at key points during the workflow, but the product team was unsure what was motivating decisions to abandon the workflow.

 
 
Is it meant to look complicated? Because it does.
— Research participant
 

GOAL:

Gain user trust, harness engagement and mitigate abandonment. Standardize workflows and components to the Azure framework design system. Design a more efficient, intuitive experience utlizing common interaction patterns, known data, providing clear calls-to-action, and enabling a more accessible and comprehensive understanding through documentation of benefits, requirements, and process to use Azure Quantum computing/optimization solutions.

PROCESS:

User research confirmed that creating an Azure Quantum workspace was difficult and confusing, costly decisions ultimately led to user frustration and a high level of workflow abandonment.

  • Conducted baseline user-research focused on the current experience journeys of quantum thought leaders, researchers, developers, enthusiasts and practitioners.

  • Synthesized research insights and journey mapped the user’s experience.

  • Socialized user journey insights and led cross-functionally design workshops, promoting visibility, design participation, and alignment on a common understanding of the user’s current experience, and user’s intentions, expectations, and goals using Azure Quantum.

    • Users

Defined the UX Outcome - to provide a clear path for the quickest method to create an Azure Quantum workspace… then iteratively design to meet this goal.

  • Partnered with Azure Fluent framework PM and Design leadership teams and drove alignment with Fluent patterns and system initiatives.

  • Worked with engineering to identify opportunity to optimize the workflow

  • Worked with Microsoft Legal counsel to streamline provider Terms of Conditions experience.

  • Workshopped page messaging, titles, taglines to be more inviting, informative and effective calls-to-actions

  • Fluent Design system component creation, evolution, socialization, and documentation.

  • Created fully functional prototypes for design discussions to enable research and design iteration, stakeholder feedback, and drive alignment.

  • Revised documentation to align with and reflect design decisions and workflow enhancements

OUTCOME:

  • Refined the end-to-end workflow with better alignment with Azure Fluent design specifications

  • Optimized user data input - Identified a number of inputs that could be prefilled and others that could use smart-defaults or suggestive placeholders.

  • Consolidated multiple Terms of Service into a single input to accept

  • Improved Documentation

Documentation improvements, consolidating Terms and Conditions, and aligning with Fluent’s common usability patterns alone improved workspace creation and first quantum job submission success from 14.9% to 42%. - We named this workflow the Advance Create path

* User research revealed a strong preference for less complex choices during onboarding and a quicker low-coding experience that provided a conceptual understanding before diving into optimization of their workspace, after their first job submission.

Taking this into account, in addition to the Advanced Create workflow, I designed a Quick-Create workflow to reduce heavy pricing and provider decisions presented within the Advanced create workflow. In addition to utilizing the above Advanced Create optimizations, the Quick-Create workflow also selects all FREE providers and enables a much more streamlined, less complex workflow, requiring ONLY a unique “Workspace name” to create a workspace, with the ability to edit all workspace details post-creation.

Now, having two workflows options, we needed a way to select each path and provide a choice for users. I created an initial page that included a choice between Advanced Create path and the Quick Create path, and designed the Subscription input to check the user's currently logged-in subscription and pre-select it.

  • If the user left the Subscription select blank, or was not the owner of the Subscription (with no pre-defined subscription, region and storage settings) then the Advanced create button was the only option enabled.

  • If the user was the owner of the subscription, then the Quick create path was pre-selected, with the freedom to choose Advanced create if desired.

Here are the complete workflows you can click through.

Advanced Create path

 

Workspace Quick-Create workflow

Quick create, only two steps, with a single keyboard input required for “Workspace name”. This approach enabled an instant workspace creation and boosted workspace creation completion rates from 42% to over 70%

 

In closing…

I leveraged user research insights to deliver a smoother onboarding experience, introducing friendlier workflow messaging, a simplified Terms of Service, clearer calls-to-action, and alignment with Azure Fluent design system standards. Partnering cross-functionally to lead workshops, align stakeholders, and design a low-friction workspace creation workflow that improved success rates by nearly 5x.

* Solutions for Azure Quantum workspace creation, Azure Quantum Credits program, first-run workflows, and Azure Quantum Sample Gallery all evolved from this initial research and design socialization.


2 - Clarifying Azure Quantum Overview page and Getting Started workflows

Overview page served as a central user management hub, with a launching point for Quantum services, including Getting Started.

CHALLENGE:

User research showed that the Overview page and Quick Start options were confusing and overwhelming, making it difficult for first-time users to determine the easiest path to begin.

On page load, users were presented two Quick Start options Quantum Computing Quickstart and Optimization Quickstart and a lackluster page with no excitement or driving call-to-action.

The two options were not “quick”, but instead complex workflows that immediately frustrated and overwhelmed users who were looking for an easy win to take back to their organizations. Both options involved setting up your local environment in VS code, adding the Azure Quantum Developer Kit (QDK) extension, include Python and Jupyter Notebook support if needed, install any Python updates, then connect Jupyter notebooks through an Azure Quantum workspace to submit programs to a remote quantum service. If user wanted to do both Quantum computing AND Quantum Optimization, then they had to go through both of these setup configurations.

Users were starting then abandoning these workflows due to lack of engaging calls-to-action, and the high complexity and lack of overall knowledge and trust in Azure Quantum services and pricing options.

GOAL:

  • Provide a launching point that focuses new users on the easiest pathway to submit a first job, while enabling education and experiences for returning users

  • Through better documentation, educate users about all quantum job submission options

  • Ensure design direction supports both new and returning users

  • Align closely with Azure Fluent design patterns

PROCESS:

Led user-centered research, presentations, design workshops and design discussions to enable feedback, iteration, and solidification of design and vision.

  • Conducted user research studies with Quantum thought leaders, researchers, developers, enthusiasts, and practitioners to better understand and drive redesign discussions from a user-centered perspective.

  • Synthesized and Socialized research and design through design workshops and presentations, enabling visibility and participation to cross-functionally understand, define and promote a common understanding of the user’s intentions, expectations, and goals.

    • A number of research participants made suggestions for easier getting started options and better support documentation. One specific research participants suggestion was to integrate Jupyter notebooks (interactive environments (details)) into the getting-started exercises. Doing so would enable Azure Quantum to provide pre-made notebooks to use as a starting point for learning and exploration.

  • I led a cross-functional figjam workshop to explore repositioning the portal as a “Getting Started” experience with easier options to begin. The workshop exercise enabled visibility, participation, and alignment on a path forward. Both the repositioning of the page and the Jupyter Notebook integration gained traction and support.

    • However, initially, I was unable to convince Engineering leadership that the notebook option would be the easiest workflow and be prioritized first, so I couldn’t rename it to “Quickstart”😖

  • I created a proposal that broke down the Getting Started experience into a three-panel design and catered to both new and returning users, with the Jupyter Notebook option, the more advanced Local configuration options and a “Learn More” option for more useful documentation.

  • I ran another round of usability research on this design and found that the most prominent choice was still Local configuration with 4/5 quantum researcher study participants selected a more difficult and advanced method over a much simpler option of using templated premade Sample Jupyter notebooks within Azure Quantum… when asked why, participants said “because it’s labeled quickstart”. 😃

  • After another short redesign iteration, I A/B tested a Getting Started proposal that renamed the Jupyter Notebooks workflow as the “Quickstart” option and prioritized it first on the page, while presenting the Local configuration option as a secondary, more advanced option, with Learn More as a third option that linked to documentation, for users that found the two options difficult.

  • I also provided a variation that completely removed all other calls to action on the page and highlighted the Jupyter Notebook option as “Quickstart without install” option

The “Quickstart without install” variation WAS A HIT! Users responded enthusiastically to its simplicity and clear call-to-action to “Submit your first job using Notebooks within minutes”.

  • Users sailed past this step and on to Sample Gallery and using Jupyter Notebooks to submit their first job.

  • More experienced users were still able to see and access the more advanced Local Configuration option and were not bothered by this repositioning

  • Partnered with Azure Fluent design system teams drove consistency with Fluent design system initiatives.

OUTCOME:

I drove cross-functional alignment and drove impact through user research and socializing the user sentiment throughout the product team. User research insights enabled initial short-term wins by correcting unintended implementation inconsistencies and updating user documentation, while shaping long-term product innovation and strategic vision.

  • User research revealed the easiest pathway through Azure Quantum to submit a first job was using pre-built Jupyter notebook

  • Initial short-term win - Partnering with Fluent design system team component and pattern usage helped drive consistency and alignment with Azure design leadership and their Fluent design system initiatives, with the added benefit of not reinventing the wheel.

  • Initial short-term win - Text revisions and providing workflow guidance enabled a higher user confidence, better comprehension and more success navigating through Getting Started exercises.

  • More concise and effective documentation enabled users and lowered the support cost on Azure Quantum team.

  • Renamed the Jupyter Notebooks workflow as the “Quickstart” option and prioritized it first on the page, while presenting the Local configuration option as a secondary, more advanced option, with Learn More as a third option that linked to documentation, for users that found the two options difficult.

  • Quota usage was moved to the Operations left side menu and co-located with credit usage and Job usage content. This presented a more intuitive ingress point and holistic view of usage, and removed unneeded clutter from the getting started workflow.

In closing…

Rapid research and design iterations shaped the Getting Started experience to serve the needs of both first-time and advanced users, offering a clear path to submit a first quantum job, advanced local configuration options, and on-demand learning opportunities. This work sparked the creation of the notebook Gallery, enabling users to explore templated Jupyter Notebooks to learn and driving a 3x increase in successful job submissions


3 - Empowering User Education by Enabling Better Sample Gallery UX

CHALLENGE:

The Notebook Gallery showcased a collection of pre-built Jupyter Notebook interactive environments (click for details), that provided interactive computing, data analysis, and visualizations in one place, combining the functionality of a document within a code editor. Designed to help users explore, learn, build, and offer a seamless entry point into quantum computing and optimization through hands-on, partner-driven examples.

User research indicated that users found Notebook Gallery difficult to use, find advanced samples, understand how to manipulate between various quantum providers, and navigate the workflow to submit quantum jobs. This led to low engagement, more abandonment, and fewer job submissions.

 
The page appears visually monotonous and is not optimized for specific use cases
— Quantum Educator
 
 
 

GOAL:

Create an intuitive, inviting Sample Gallery Experience that drives engagement and user education. Help users learn taxonomy, structure and process while guiding them through sample-driven task completion and quantum job submission.

PROCESS:

I led and fostered a cross-functional UX research and data-driven design process. This enabled design socialization and inclusion through usability studies and design-led workshops and weekly UX office hours. Iteration backed by strong user-centered data-driven design decisions, alleviated short-term friction and shaped long-term Sample Gallery vision.

  • Viewed telemetry over a 20-day observation period:

    • 58% of users who accessed the Notebook Gallery, cloned a notebook to use

    • 50% of these users actually utilized the notebook and ran at least one cell (executed a task). So thats 25% of users that accessed Notebook Gallery, cloned and actually tried to use a notebook.

    • Only 17% of of users who landed on Notebooks Gallery, cloned a notebook then completed the notebook sample workflow and actually submitted a quantum job.

  • Conducted and synthesized a user research study with 6 Quantum researchers, to test their ability to submit a Jupyter notebook sample.

  • Research exposed that Notebook Gallery was unclear and undervalued

  • 3/6 research participants commented that they didn’t know what a Notebook was

  • 2/6 were able to complete the task and submit a Jupyter Notebook within an hour.

  • 2/6 participants mentioned they would like to see more of a Hello world introduction to the gallery.

  • Socialized research and design through cross-functional research studies, design workshops, and presentations. Enabling participation and visibility, to define and cross-functionally promote a common understanding of the user’s intentions, expectations and goals.

  • Standardized on common usability patterns with basic alignment, more informative titles and helpful call-to-action taglines

  • Partnered cross-functional and cross-team with Azure Fluent design system PM and Design teams to ensure alignment with Fluent Design patterns and system initiatives

  • Design system component creation & evolution, socialization, and documentation

  • Created fully functional prototypes for design discussions to enable feedback, iteration, and solidification of design and vision.

OUTCOME:

Redesigned Notebook Gallery enabled user education and engagement through a streamlined, inviting, and familiar experience, by reducing clutter, clarifying calls-to-action, and aligning with Azure Fluent standards.

  • Enabled a more consistent and intuitive end-user experience while enabling quicker implementation by aligning with Azure Fluent design system and benefitting from known card patterns.

    • Added a variation to Fluent’s card pattern with the inline provider selection that highlighted Azure Quantum’s provider-agnostic and kernel agnostic approach.

  • Enhanced categorization provided order to the page and surfaced targeted content across Quantum Computing and Optimization, with Hello world samples for each.

  • Hello World notebook samples were a success!

    • Provided a welcomed, introductory experience

    • Enabled the easiest path for quantum job submissions

    • Showcased Azure Quantum’s provider and language-agnostic flexibility through interactive variations.

  • Post-submission, the Hello World samples collapsed to elevate relevance and surface notebooks users had previously engaged with or were actively working on.

 
They’re super super stripped down, like the simplest possible kind of thing, really just showing how it all just fits which is perfect. Because it gets you started, so it does what it’s intended to do, which is getting going and started so it’s really great.
— Quantum explorer
 

In closing…

88% of users agreed with the statement for the redesigned experience: “Sample Gallery is integral to learning, exploration, and understanding of Azure Quantum's offerings”.

Rapid research and design iterations balanced first-time and advanced users needs, offering clear concise call-to-action for the easiest and quickest path to submit a first quantum job, with the ability to learn more if additional information was required.


4 - Enabling clarity and trust through Azure Quantum Credits and Quota experience

CHALLENGE:

User research for Credits and Quotas experience revealed that many users were experiencing difficult conversations within their organizations for funding quantum research. Users claimed the Azure Quantum Credits and Quotas experience and its data visualizations was confusing, unhelpful, with a lack of concrete usage and funding data that researchers could take back to their organizations.

All participants in two separate mixed content studies (12 participants) expressed 6 key points of dissatisfaction with the Credits and quotas experience.

  • Frustration over the lack of standardization in a unit of measurement from provider to provider

  • The use of subjective data visualizations

  • Lack of Subscription and Quota data

  • Lack of clarity around how credits are applied for and used

  • Lack of clarity around associated costs.

  • Lack of structure that users could take back to their organization

The lack of clarity around the limited credit program and the absence of control over increases or upgrades within the experience drove users to call Azure Quantum support, incurring costs on our support and product management teams.

GOAL:

Provide users with clear visibility into credit and quota usage at the workspace and subscription level, contextualized by their selected provider plan, with easy access to upgrade or increase allocations.

I championed and led a cross-functional UX research and data-driven design process enabling design socialization and inclusion through usability studies and design-led, user-centered workshops and weekly UX office hours.

  • Partnered with Azure Fluent design teams to standardize and drive alignment

  • Usability study to get a user-centered pulse.

  • Iterated on design from research insights

  • Socialized research and design through cross-functional research studies, design workshops, presentations, and weekly UX office hours. This enabled participation and visibility, allowing for the definition and cross-functional promotion of a common understanding of the user’s intentions, expectations, and goals.

  • Design system component creation & evolution, socialization, and documentation

  • Created fully functional prototypes for design discussions to enable feedback, iteration, and solidification of design and vision.

PROCESS:

OUTCOME:

  • Created the Quantum credits program to remove financial barriers, where users could apply for credits and grants to alleviate some of the initial financial strain on their organization. Users’ funding conversations were simplified and enabled with data. - click for details

  • Redesigned the Credits and Quota dashboard experience within a standardized table presentation

  • Standardized on total quantity of units for measurements and amount used per Azure subscription

  • Simple bar graph data visualizations, per quantum provider subscription and displayed usage and credit availablility as a percentage of that whole

  • Presented associated costs and the ability to increase and upgrade plans as needed.

  • Users could now account for quota availability and credit usage, apply for grants, request upgrades, and more easily plan for future projected costs and funding discussions within their organizations.

 
What really made the difference was these credits, this credit program that allowed us to experiment with what we were doing without using our own budget, which is very limited, so that was for us a game changer.
— Quantum Innovator
 
 

In closing…

The Redesigned Credits and Quotas experience provided valuable user insight, inspired trust in Azure Quantum, and lowered the financial burden of learning quantum computing and quantum optimization within the Azure Quantum platform.

 

 

Quantum computing and quantum optimization made easy! 😜

Collectively, these design outcomes enabled Azure Quantum users to achieve their short-term quantum goals, plan for funding conversations and long-term quantum strategy within their organizations. Through user research and data-driven design outcomes, these solutions drastically improved people’s confidence, engagement, and productivity in Azure Quantum. Standardization on familiar usability patterns, and a user-centered focus drove innovative design, effective data visualization, insightful and empowering workflows, and holistically informed, effective user-centered experiences.

I’ll leave you with one last quote for my final round of user research.

 
I think Azure Quantum is really cool. Its pretty straight forward to setup and use. So I definitely think that it could get even cooler and I would love to keep using it.
— Quantum Innovator
 
 

Thank you for reading! 😃

I’d love to hear your thoughts! Connect with me on Linkedin or send me a message below.