K-12 budgeting and HR product suite unified into a single, cohesive experience under a refreshed design system with scalable, accessible components.
K-12 budgeting and HR product suite unified into a single, cohesive experience under a refreshed design system with scalable, accessible components.
K-12 budgeting and HR product suite unified into a single, cohesive experience under a refreshed design system with scalable, accessible components.
✦
Senior Product Designer
Senior Product Designer
✢
2023-2024
2023-2024
✱
Design System
Design System
❖
Edtech, Finance, HR
Edtech, Finance, HR



The problem
Allovue’s apps exist as three siloed products instead of one cohesive unit.
Allovue's product suite was designed to help schools and districts plan, manage, and track their financial and HR data — but each app was built in isolation, and resulted in a user experience that felt both disjointed and difficult to navigate. Users often felt confused and had their expectations unmet by similar, but not identical, UX patterns and functionality.

The problem
Allovue’s apps exist as three siloed products instead of one cohesive unit.
Allovue's product suite was designed to help schools and districts plan, manage, and track their financial and HR data — but each app was built in isolation, and resulted in a user experience that felt both disjointed and difficult to navigate. Users often felt confused and had their expectations unmet by similar, but not identical, UX patterns and functionality.

The problem
Allovue’s apps exist as three siloed products instead of one cohesive unit.
Allovue's product suite was designed to help schools and districts plan, manage, and track their financial and HR data — but each app was built in isolation, and resulted in a user experience that felt both disjointed and difficult to navigate. Users often felt confused and had their expectations unmet by similar, but not identical, UX patterns and functionality.
By the numbers
What makes developing the product without a design system difficult?
80+
Duplicate table components
Across the app, more than 80 variations were copy-pasted, resulting in an amassing tech debt for table settings and features (filtering, grouping, etc).
3-4+
code variations per component
In addition to copy-pasting similar code, devs would often write their own version of a component, leading to multiple helper files for a single widget.
~16-20
active developers
Devs were often split into smaller teams for different features, resulting in limited working knowledge of usable code components and overarching UX.
60+
Figma files
With no central source of truth, design patterns were scattered across a growing number of feature files, each with their own type, layout, and component guidelines.
By the numbers
What makes developing the product without a design system difficult?
80+
Duplicate table components
Across the app, more than 80 variations were copy-pasted, resulting in an amassing tech debt for table settings and features (filtering, grouping, etc).
3-4+
code variations per component
In addition to copy-pasting similar code, devs would often write their own version of a component, leading to multiple helper files for a single widget.
~16-20
active developers
Devs were often split into smaller teams for different features, resulting in limited working knowledge of usable code components and overarching UX.
60+
Figma files
With no central source of truth, design patterns were scattered across a growing number of feature files, each with their own type, layout, and component guidelines.
By the numbers
What makes developing the product without a design system difficult?
80+
Duplicate table components
Across the app, more than 80 variations were copy-pasted, resulting in an amassing tech debt for table settings and features (filtering, grouping, etc).
3-4+
code variations per component
In addition to copy-pasting similar code, devs would often write their own version of a component, leading to multiple helper files for a single widget.
~16-20
active developers
Devs were often split into smaller teams for different features, resulting in limited working knowledge of usable code components and overarching UX.
60+
Figma files
With no central source of truth, design patterns were scattered across a growing number of feature files, each with their own type, layout, and component guidelines.
Who we're helping
Users and team members alike struggle with finding consistent user experiences.
Principals
“
I don’t understand why I can’t filter on this table before export. I can do it on this other page — why not here?
In charge of
Budgeting for ~1-5 institutions, including hiring and staffing changes, across multiple apps in the suite
Current frustrations
Inability to have the same basic functionality across tables for filtering, sorting, grouping, etc. means being unable to rely on the app to export consistent reports for budget reviews
Superintendents
“
My dashboard shows me a summary of my district’s progress, but individual pages don’t show me the same charts.
In charge of
Managing and reviewing budgets for all schools/institutions within a district, and submitting for board approval
Current frustrations
Confusing expectations for data visualization make it difficult to understand how individual budgets contribute to the district’s whole progress, and which are in current need of attention
Engineers
“
I don’t know how to figure out if we already have a similar component, other than asking in Slack and hoping somebody online knows.
In charge of
Developing features and components according to new design specs, while also trying to re-use/copy code from existing functionality
Current frustrations
No central place to find reusable code components, and no way of knowing which components are most up-to-date in terms of design and functionality
Designers
“
I don’t know why we made this design decision, but there might be a clue in one of the 20 pages of this old feature file?
In charge of
Designing and maintaining a consistent user experience, including building off of established design patterns and features
Current frustrations
Mismatches between designs and in-app features cause feature drift; no way of tracking decision-making leads to time redesigning similar features without understanding implementation barriers
Who we're helping
Users and team members alike struggle with finding consistent user experiences.
Principals
“
I don’t understand why I can’t filter on this table before export. I can do it on this other page — why not here?
In charge of
Budgeting for ~1-5 institutions, including hiring and staffing changes, across multiple apps in the suite
Current frustrations
Inability to have the same basic functionality across tables for filtering, sorting, grouping, etc. means being unable to rely on the app to export consistent reports for budget reviews
Superintendents
“
My dashboard shows me a summary of my district’s progress, but individual pages don’t show me the same charts.
In charge of
Managing and reviewing budgets for all schools/institutions within a district, and submitting for board approval
Current frustrations
Confusing expectations for data visualization make it difficult to understand how individual budgets contribute to the district’s whole progress, and which are in current need of attention
Engineers
“
I don’t know how to figure out if we already have a similar component, other than asking in Slack and hoping somebody online knows.
In charge of
Developing features and components according to new design specs, while also trying to re-use/copy code from existing functionality
Current frustrations
No central place to find reusable code components, and no way of knowing which components are most up-to-date in terms of design and functionality
Designers
“
I don’t know why we made this design decision, but there might be a clue in one of the 20 pages of this old feature file?
In charge of
Designing and maintaining a consistent user experience, including building off of established design patterns and features
Current frustrations
Mismatches between designs and in-app features cause feature drift; no way of tracking decision-making leads to time redesigning similar features without understanding implementation barriers
Who we're helping
Users and team members alike struggle with finding consistent user experiences.
Principals
“
I don’t understand why I can’t filter on this table before export. I can do it on this other page — why not here?
In charge of
Budgeting for ~1-5 institutions, including hiring and staffing changes, across multiple apps in the suite
Current frustrations
Inability to have the same basic functionality across tables for filtering, sorting, grouping, etc. means being unable to rely on the app to export consistent reports for budget reviews
Superintendents
“
My dashboard shows me a summary of my district’s progress, but individual pages don’t show me the same charts.
In charge of
Managing and reviewing budgets for all schools/institutions within a district, and submitting for board approval
Current frustrations
Confusing expectations for data visualization make it difficult to understand how individual budgets contribute to the district’s whole progress, and which are in current need of attention
Engineers
“
I don’t know how to figure out if we already have a similar component, other than asking in Slack and hoping somebody online knows.
In charge of
Developing features and components according to new design specs, while also trying to re-use/copy code from existing functionality
Current frustrations
No central place to find reusable code components, and no way of knowing which components are most up-to-date in terms of design and functionality
Designers
“
I don’t know why we made this design decision, but there might be a clue in one of the 20 pages of this old feature file?
In charge of
Designing and maintaining a consistent user experience, including building off of established design patterns and features
Current frustrations
Mismatches between designs and in-app features cause feature drift; no way of tracking decision-making leads to time redesigning similar features without understanding implementation barriers
Current State
Three systems that should be one easy, seamless experience.
The original Allovue product suite had been built 10 years earlier and consisted of three separate codebases. In an effort to bring the apps into a more modern tech stack, they were being rewritten using Phoenix and Elixir. As each of the apps were initially built by siloed teams, they used similar (but ultimately separate) design languages, had differing UX patterns for components, and had features that couldn’t ultimately be duplicated easily due to differing tech stacks. As team members changed, it became even more difficult to maintain consistent patterns across design and development.

Budget
Plan and forecast budget changes for the upcoming year.
Used by
Budgeters and HR to plan non-personnel expenses and staffing costs for the upcoming fiscal year.

Allocate
Tie long-term financial goals to budget spending.
Used by
District leaders to allocate district-wide resources based on enrollment projections and funding formulas.

Manage
See current budget health and school year spending.
Used by
District leaders to review and manage the current state of their budget spending, including staffing changes and incoming funds.
Current State
Three systems that should be one easy, seamless experience.
The original Allovue product suite had been built 10 years earlier and consisted of three separate codebases. In an effort to bring the apps into a more modern tech stack, they were being rewritten using Phoenix and Elixir. As each of the apps were initially built by siloed teams, they used similar (but ultimately separate) design languages, had differing UX patterns for components, and had features that couldn’t ultimately be duplicated easily due to differing tech stacks. As team members changed, it became even more difficult to maintain consistent patterns across design and development.

Budget
Plan and forecast budget changes for the upcoming year.
Used by
Budgeters and HR to plan non-personnel expenses and staffing costs for the upcoming fiscal year.

Allocate
Tie long-term financial goals to budget spending.
Used by
District leaders to allocate district-wide resources based on enrollment projections and funding formulas.

Manage
See current budget health and school year spending.
Used by
District leaders to review and manage the current state of their budget spending, including staffing changes and incoming funds.
Current State
Three systems that should be one easy, seamless experience.
The original Allovue product suite had been built 10 years earlier and consisted of three separate codebases. In an effort to bring the apps into a more modern tech stack, they were being rewritten using Phoenix and Elixir. As each of the apps were initially built by siloed teams, they used similar (but ultimately separate) design languages, had differing UX patterns for components, and had features that couldn’t ultimately be duplicated easily due to differing tech stacks. As team members changed, it became even more difficult to maintain consistent patterns across design and development.

Budget
Plan and forecast budget changes for the upcoming year.
Used by
Budgeters and HR to plan non-personnel expenses and staffing costs for the upcoming fiscal year.

Allocate
Tie long-term financial goals to budget spending.
Used by
District leaders to allocate district-wide resources based on enrollment projections and funding formulas.

Manage
See current budget health and school year spending.
Used by
District leaders to review and manage the current state of their budget spending, including staffing changes and incoming funds.
Making the plan
Running a component audit: accessibility, visual consistency, and tech debt.
To begin to understand where we were with our current design language, we started a full product audit of the components used across the new codebase. I worked alongside some folks in engineering to create a spreadsheet in Airtable, cataloguing the name, location in app, Figma links, expected functionality, and notes on tech debt for every component. Once we had a working list, it was easy to start grouping components that should be merged into a single unit, or components that had shared functionality that would benefit from smaller building blocks of code.
We also made a separate sheet for tables, specifically — given that most of our app was based around data tables, we wanted to make sure that the differences in data management UX were clearly documented.
The team
Senior Product Designer
x1
Senior Engineer
x1
Engineer
x1

Making the plan
Running a component audit: accessibility, visual consistency, and tech debt.
To begin to understand where we were with our current design language, we started a full product audit of the components used across the new codebase. I worked alongside some folks in engineering to create a spreadsheet in Airtable, cataloguing the name, location in app, Figma links, expected functionality, and notes on tech debt for every component. Once we had a working list, it was easy to start grouping components that should be merged into a single unit, or components that had shared functionality that would benefit from smaller building blocks of code.
We also made a separate sheet for tables, specifically — given that most of our app was based around data tables, we wanted to make sure that the differences in data management UX were clearly documented.
The team
Senior Product Designer
x1
Senior Engineer
x1
Engineer
x1

Making the plan
Running a component audit: accessibility, visual consistency, and tech debt.
To begin to understand where we were with our current design language, we started a full product audit of the components used across the new codebase. I worked alongside some folks in engineering to create a spreadsheet in Airtable, cataloguing the name, location in app, Figma links, expected functionality, and notes on tech debt for every component. Once we had a working list, it was easy to start grouping components that should be merged into a single unit, or components that had shared functionality that would benefit from smaller building blocks of code.
We also made a separate sheet for tables, specifically — given that most of our app was based around data tables, we wanted to make sure that the differences in data management UX were clearly documented.
The team
Senior Product Designer
x1
Senior Engineer
x1
Engineer
x1

Building a foundation
Working with our current designs to modernize and refine our system.
After our initial audit, the design team started to pull examples from key pages of the existing apps in order to produce a fuller picture of the product suite. While each of the apps had similar design foundations, they weren’t the same — and needed to be redefined using modern type scales, accessible colors, and responsive design methodologies.
Instead of reinventing the wheel, we identified the visual elements and rules that best supported the main interfaces — key type sizes/weights, color language, and component functionality — and made sure to extend those systems rather than fully replace them. We made use of tools like Huetone and Accessible Palette to extend our shades in a system similar to Radix UI’s, and Typescale and The Good Line Height to keep our type looking good on the web.
We ended up with a modern, accessible, refreshed version of our older systems that helped the products feel like a single experience. Below are some examples of how the foundations system changes started to shape our new visual language.
Color palette

What we kept
Key color accents pulled from branding, extended into a 12-step color scale system.
What we updated
Introduced color tokens and an extended scale system, and guidelines for which shades could be paired together to ensure meeting WCAG contrast requirements.
Typography

What we kept
A few key sizes and weights from the original dashboards and main layouts of the app.
What we updated
An extended type scale, with defined type styles (H1/H2/H3, paragraph, link, etc) and pairing guidelines.
Icons

What we kept
Some custom icons, updated to match the visual style of the new icon set.
What we updated
Moved from three separate icon sets to Tabler icons (and a few custom ones for our needs!)
Charts + Data

What we kept
Usage of different line weights and secondary textures to differentiate charts (instead of just color)
What we updated
Accessible chart colors, tooltips, and more refined type layouts to help cognitive load for users
Components

What we built
Reusable component library, complete with design decision logs, usage guidelines, development specs, and accessibility notes for implementation.
Patterns

What we built
Design pattern library, consisting of standardized flows for features like table filtering/grouping, upload form flows, page layouts, and more.
Inputs + Forms

What we built
Standardized library of input components with optional features, and accessible form patterns for creating, modifying, and uploading data.
Illustrations

What we built
Custom vector illustration library for in-app usage, as well as onboarding and marketing materials, based on our newly refined palette.
Building a foundation
Working with our current designs to modernize and refine our system.
After our initial audit, the design team started to pull examples from key pages of the existing apps in order to produce a fuller picture of the product suite. While each of the apps had similar design foundations, they weren’t the same — and needed to be redefined using modern type scales, accessible colors, and responsive design methodologies.
Instead of reinventing the wheel, we identified the visual elements and rules that best supported the main interfaces — key type sizes/weights, color language, and component functionality — and made sure to extend those systems rather than fully replace them. We made use of tools like Huetone and Accessible Palette to extend our shades in a system similar to Radix UI’s, and Typescale and The Good Line Height to keep our type looking good on the web.
We ended up with a modern, accessible, refreshed version of our older systems that helped the products feel like a single experience. Below are some examples of how the foundations system changes started to shape our new visual language.
Color palette

What we kept
Key color accents pulled from branding, extended into a 12-step color scale system.
What we updated
Introduced color tokens and an extended scale system, and guidelines for which shades could be paired together to ensure meeting WCAG contrast requirements.
Typography

What we kept
A few key sizes and weights from the original dashboards and main layouts of the app.
What we updated
An extended type scale, with defined type styles (H1/H2/H3, paragraph, link, etc) and pairing guidelines.
Icons

What we kept
Some custom icons, updated to match the visual style of the new icon set.
What we updated
Moved from three separate icon sets to Tabler icons (and a few custom ones for our needs!)
Charts + Data

What we kept
Usage of different line weights and secondary textures to differentiate charts (instead of just color)
What we updated
Accessible chart colors, tooltips, and more refined type layouts to help cognitive load for users
Components

What we built
Reusable component library, complete with design decision logs, usage guidelines, development specs, and accessibility notes for implementation.
Patterns

What we built
Design pattern library, consisting of standardized flows for features like table filtering/grouping, upload form flows, page layouts, and more.
Inputs + Forms

What we built
Standardized library of input components with optional features, and accessible form patterns for creating, modifying, and uploading data.
Illustrations

What we built
Custom vector illustration library for in-app usage, as well as onboarding and marketing materials, based on our newly refined palette.
Building a foundation
Working with our current designs to modernize and refine our system.
After our initial audit, the design team started to pull examples from key pages of the existing apps in order to produce a fuller picture of the product suite. While each of the apps had similar design foundations, they weren’t the same — and needed to be redefined using modern type scales, accessible colors, and responsive design methodologies.
Instead of reinventing the wheel, we identified the visual elements and rules that best supported the main interfaces — key type sizes/weights, color language, and component functionality — and made sure to extend those systems rather than fully replace them. We made use of tools like Huetone and Accessible Palette to extend our shades in a system similar to Radix UI’s, and Typescale and The Good Line Height to keep our type looking good on the web.
We ended up with a modern, accessible, refreshed version of our older systems that helped the products feel like a single experience. Below are some examples of how the foundations system changes started to shape our new visual language.
Color palette

What we kept
Key color accents pulled from branding, extended into a 12-step color scale system.
What we updated
Introduced color tokens and an extended scale system, and guidelines for which shades could be paired together to ensure meeting WCAG contrast requirements.
Typography

What we kept
A few key sizes and weights from the original dashboards and main layouts of the app.
What we updated
An extended type scale, with defined type styles (H1/H2/H3, paragraph, link, etc) and pairing guidelines.
Icons

What we kept
Some custom icons, updated to match the visual style of the new icon set.
What we updated
Moved from three separate icon sets to Tabler icons (and a few custom ones for our needs!)
Charts + Data

What we kept
Usage of different line weights and secondary textures to differentiate charts (instead of just color)
What we updated
Accessible chart colors, tooltips, and more refined type layouts to help cognitive load for users
Components

What we built
Reusable component library, complete with design decision logs, usage guidelines, development specs, and accessibility notes for implementation.
Patterns

What we built
Design pattern library, consisting of standardized flows for features like table filtering/grouping, upload form flows, page layouts, and more.
Inputs + Forms

What we built
Standardized library of input components with optional features, and accessible form patterns for creating, modifying, and uploading data.
Illustrations

What we built
Custom vector illustration library for in-app usage, as well as onboarding and marketing materials, based on our newly refined palette.
Maintaining consistency
Keeping everything in sync with our tokens system.
When we started work on updating the design system, it became important to us to have a workflow that ensured both design and engineering had easy and quick access to the most up-to-date design language, ideally in one central location. Although Storybook was our usual first pull, the product suite was being re-written in Elixir, so we ended up trialing a few other options before settling on (funnily enough!) Phoenix Livebook, an Elixir-based version of Storybook.
While our initial component audit took place in Airtable, we began to find that updating component status there, in Figma, and in Livebook became too much of a hassle. Our first solution was to write a custom Figma plugin to sync design token updates with Livebook, but as the Livebook community gained more traction and tooling, we instead moved to using automated updates via Slack and Github integrations to keep our team in sync.


Maintaining consistency
Keeping everything in sync with our tokens system.
When we started work on updating the design system, it became important to us to have a workflow that ensured both design and engineering had easy and quick access to the most up-to-date design language, ideally in one central location. Although Storybook was our usual first pull, the product suite was being re-written in Elixir, so we ended up trialing a few other options before settling on (funnily enough!) Phoenix Livebook, an Elixir-based version of Storybook.
While our initial component audit took place in Airtable, we began to find that updating component status there, in Figma, and in Livebook became too much of a hassle. Our first solution was to write a custom Figma plugin to sync design token updates with Livebook, but as the Livebook community gained more traction and tooling, we instead moved to using automated updates via Slack and Github integrations to keep our team in sync.


Maintaining consistency
Keeping everything in sync with our tokens system.
When we started work on updating the design system, it became important to us to have a workflow that ensured both design and engineering had easy and quick access to the most up-to-date design language, ideally in one central location. Although Storybook was our usual first pull, the product suite was being re-written in Elixir, so we ended up trialing a few other options before settling on (funnily enough!) Phoenix Livebook, an Elixir-based version of Storybook.
While our initial component audit took place in Airtable, we began to find that updating component status there, in Figma, and in Livebook became too much of a hassle. Our first solution was to write a custom Figma plugin to sync design token updates with Livebook, but as the Livebook community gained more traction and tooling, we instead moved to using automated updates via Slack and Github integrations to keep our team in sync.


Setting expectations
Standardizing our component behaviors across the products.
Approach
Identify similar components, break them down into the smallest atoms, and build responsive, accessible components from stacking building blocks.
Impact
Shared functionality, increased development speed, and reduced avenues for bugs and vulnerabilities.
As we began design work on individual components, we wanted to make sure that expected functionality and accessibility considerations were accounted for at the start of the process. It was important to us not only to ensure visual compliancy with WCAG regulations (updating our color contrast using APCA compliancy, adjusting visual hierarchy to help colorblind users), but also to ensure that the code version of the component was accessible.
Each component had a small spec written for it, outlining its use cases, relevant development patterns on W3C, necessary ARIA roles, keyboard functionality, and other best practices from our peers in the industry. I recommend checking out Heydon Pickering’s work on Inclusive Components, as well as Adrian Roselli’s articles about modern table design — they’re both great resources for providing clean, robust, accessible web components without sacrificing on some of the quality of life features that modern components require.
In this stage, I worked as a design engineer alongside the engineering team — taking components from my initial design work in Figma and developing, testing, and documenting them in Elixir. Our components were all written in isolation and demoed first in our version of Storybook — and brought into the main app in incremental updates.
Buttons

What we gained
Button with optional parameters for size, emphasis, leading/following icons, interaction states, loading animations, and special disabled states (with tooltips!).
Breadcrumbs

What we gained
Responsive link system with text-truncation (screen reader friendly!), smooth hover-to-reveal animations, and optional page icons.
Inputs

What we gained
Stacking component system of titles, error feedback, and inputs (text/textarea, number, select, radio, checkbox, etc) — each with multiple parameters for icons, loading states, descriptions, and more.
Tab navigation

What we gained
Keyboard-friendly tab-link component with responsive, collapsible menu for mobile.
Table settings

What we gained
Easy-to-implement filtering, sorting, and grouping modules for all data tables; a standard table toolbar with hooks ready for implementing exporting, fullscreen viewing, showing/hiding columns, and more.
Data tables

What we gained
Accessible data-table component, with expandable sub-tables, frozen columns, column types, adjustable density, and more.
Setting expectations
Standardizing our component behaviors across the products.
Approach
Identify similar components, break them down into the smallest atoms, and build responsive, accessible components from stacking building blocks.
Impact
Shared functionality, increased development speed, and reduced avenues for bugs and vulnerabilities.
As we began design work on individual components, we wanted to make sure that expected functionality and accessibility considerations were accounted for at the start of the process. It was important to us not only to ensure visual compliancy with WCAG regulations (updating our color contrast using APCA compliancy, adjusting visual hierarchy to help colorblind users), but also to ensure that the code version of the component was accessible.
Each component had a small spec written for it, outlining its use cases, relevant development patterns on W3C, necessary ARIA roles, keyboard functionality, and other best practices from our peers in the industry. I recommend checking out Heydon Pickering’s work on Inclusive Components, as well as Adrian Roselli’s articles about modern table design — they’re both great resources for providing clean, robust, accessible web components without sacrificing on some of the quality of life features that modern components require.
In this stage, I worked as a design engineer alongside the engineering team — taking components from my initial design work in Figma and developing, testing, and documenting them in Elixir. Our components were all written in isolation and demoed first in our version of Storybook — and brought into the main app in incremental updates.
Buttons

What we gained
Button with optional parameters for size, emphasis, leading/following icons, interaction states, loading animations, and special disabled states (with tooltips!).
Breadcrumbs

What we gained
Responsive link system with text-truncation (screen reader friendly!), smooth hover-to-reveal animations, and optional page icons.
Inputs

What we gained
Stacking component system of titles, error feedback, and inputs (text/textarea, number, select, radio, checkbox, etc) — each with multiple parameters for icons, loading states, descriptions, and more.
Tab navigation

What we gained
Keyboard-friendly tab-link component with responsive, collapsible menu for mobile.
Table settings

What we gained
Easy-to-implement filtering, sorting, and grouping modules for all data tables; a standard table toolbar with hooks ready for implementing exporting, fullscreen viewing, showing/hiding columns, and more.
Data tables

What we gained
Accessible data-table component, with expandable sub-tables, frozen columns, column types, adjustable density, and more.
Setting expectations
Standardizing our component behaviors across the products.
Approach
Identify similar components, break them down into the smallest atoms, and build responsive, accessible components from stacking building blocks.
Impact
Shared functionality, increased development speed, and reduced avenues for bugs and vulnerabilities.
As we began design work on individual components, we wanted to make sure that expected functionality and accessibility considerations were accounted for at the start of the process. It was important to us not only to ensure visual compliancy with WCAG regulations (updating our color contrast using APCA compliancy, adjusting visual hierarchy to help colorblind users), but also to ensure that the code version of the component was accessible.
Each component had a small spec written for it, outlining its use cases, relevant development patterns on W3C, necessary ARIA roles, keyboard functionality, and other best practices from our peers in the industry. I recommend checking out Heydon Pickering’s work on Inclusive Components, as well as Adrian Roselli’s articles about modern table design — they’re both great resources for providing clean, robust, accessible web components without sacrificing on some of the quality of life features that modern components require.
In this stage, I worked as a design engineer alongside the engineering team — taking components from my initial design work in Figma and developing, testing, and documenting them in Elixir. Our components were all written in isolation and demoed first in our version of Storybook — and brought into the main app in incremental updates.
Buttons

What we gained
Button with optional parameters for size, emphasis, leading/following icons, interaction states, loading animations, and special disabled states (with tooltips!).
Breadcrumbs

What we gained
Responsive link system with text-truncation (screen reader friendly!), smooth hover-to-reveal animations, and optional page icons.
Inputs

What we gained
Stacking component system of titles, error feedback, and inputs (text/textarea, number, select, radio, checkbox, etc) — each with multiple parameters for icons, loading states, descriptions, and more.
Tab navigation

What we gained
Keyboard-friendly tab-link component with responsive, collapsible menu for mobile.
Table settings

What we gained
Easy-to-implement filtering, sorting, and grouping modules for all data tables; a standard table toolbar with hooks ready for implementing exporting, fullscreen viewing, showing/hiding columns, and more.
Data tables

What we gained
Accessible data-table component, with expandable sub-tables, frozen columns, column types, adjustable density, and more.
Expanding our patterns
Using our building blocks to create a new, seamless navigation.
Approach
Unify the product suite and organize navigation by process needs: managing your current budgets, or planning for the future.
Impact
Reduced time for users to find, manage, and plan for multiple budgets — and allowed users to easily share and review individual financial reports.
As part of the unification of our three apps, we looked to update our main sidebar navigation to reflect our new information architecture and design language. At the time, we were still maintaining three separate sidebars, each of which had similar page names and visual treatments, but not identical — leading users to feel confused as they moved between each application during their budgeting process.
The system we had set up with our new Foundations and Components libraries made it easy to prototype a new, tabbed navigation for our users not just in Figma, but in-app. We were able to iterate quickly on navigation changes based on our feedback sessions with districts, and launch a closed beta program for a select few to test the new changes live (behind a feature flag). You can read more about the full process of redesigning the IA in Redesigning an information architecture (coming soon!).

Expanding our patterns
Using our building blocks to create a new, seamless navigation.
Approach
Unify the product suite and organize navigation by process needs: managing your current budgets, or planning for the future.
Impact
Reduced time for users to find, manage, and plan for multiple budgets — and allowed users to easily share and review individual financial reports.
As part of the unification of our three apps, we looked to update our main sidebar navigation to reflect our new information architecture and design language. At the time, we were still maintaining three separate sidebars, each of which had similar page names and visual treatments, but not identical — leading users to feel confused as they moved between each application during their budgeting process.
The system we had set up with our new Foundations and Components libraries made it easy to prototype a new, tabbed navigation for our users not just in Figma, but in-app. We were able to iterate quickly on navigation changes based on our feedback sessions with districts, and launch a closed beta program for a select few to test the new changes live (behind a feature flag). You can read more about the full process of redesigning the IA in Redesigning an information architecture (coming soon!).

Expanding our patterns
Using our building blocks to create a new, seamless navigation.
Approach
Unify the product suite and organize navigation by process needs: managing your current budgets, or planning for the future.
Impact
Reduced time for users to find, manage, and plan for multiple budgets — and allowed users to easily share and review individual financial reports.
As part of the unification of our three apps, we looked to update our main sidebar navigation to reflect our new information architecture and design language. At the time, we were still maintaining three separate sidebars, each of which had similar page names and visual treatments, but not identical — leading users to feel confused as they moved between each application during their budgeting process.
The system we had set up with our new Foundations and Components libraries made it easy to prototype a new, tabbed navigation for our users not just in Figma, but in-app. We were able to iterate quickly on navigation changes based on our feedback sessions with districts, and launch a closed beta program for a select few to test the new changes live (behind a feature flag). You can read more about the full process of redesigning the IA in Redesigning an information architecture (coming soon!).

Documenting our work
Accessible components are part of our quality assurance, every time.
Even though accessibility considerations were included in our initial design and spec work, it was important to us to make sure that we double-checked our work throughout the process. Instead of relying on accessibility checklists in our PR descriptions, we added accessibility reviews directly into our automated PR checks, which flagged any WCAG errors and offered suggestions for potentially missed elements like ARIA roles.
We also made sure that design was included as an optional group on all Design System-related PRs — so that design was never a blocker for moving ahead on critical components, but the team was always kept in the loop about the status of component changes, and able to offer input when needed.
Documenting our work
Accessible components are part of our quality assurance, every time.
Even though accessibility considerations were included in our initial design and spec work, it was important to us to make sure that we double-checked our work throughout the process. Instead of relying on accessibility checklists in our PR descriptions, we added accessibility reviews directly into our automated PR checks, which flagged any WCAG errors and offered suggestions for potentially missed elements like ARIA roles.
We also made sure that design was included as an optional group on all Design System-related PRs — so that design was never a blocker for moving ahead on critical components, but the team was always kept in the loop about the status of component changes, and able to offer input when needed.
Documenting our work
Accessible components are part of our quality assurance, every time.
Even though accessibility considerations were included in our initial design and spec work, it was important to us to make sure that we double-checked our work throughout the process. Instead of relying on accessibility checklists in our PR descriptions, we added accessibility reviews directly into our automated PR checks, which flagged any WCAG errors and offered suggestions for potentially missed elements like ARIA roles.
We also made sure that design was included as an optional group on all Design System-related PRs — so that design was never a blocker for moving ahead on critical components, but the team was always kept in the loop about the status of component changes, and able to offer input when needed.
Documenting our work
Making sure that the next folks understand, too.
A problem we kept running into across departments was that knowledge eventually became siloed — a single person might have championed an accessibility initiative, or written a great feature update, but due to a lack of documentation, it went unused (and eventually rebuilt). Design decisions were often documented across a variety of sources, including the original feature specs they were built from, Slack comments, and Jira ticket updates. There was no real central source of information that either team could point to in order to help new folks joining the team get a good understanding of the app ecosystem.
In addition to the new Livebook documentation, we updated how we structured our Figma projects — creating individual library files for Foundations, Components, Patterns, Charts/Data, and Illustrations. Each library shared a page template outlining its purpose, relevant links in-app, usage guidelines, design specs, decision logs, and more. Both the code and design resources were pinned in all of our shared spaces, and added into our onboarding processes for new teammates.

Documenting our work
Making sure that the next folks understand, too.
A problem we kept running into across departments was that knowledge eventually became siloed — a single person might have championed an accessibility initiative, or written a great feature update, but due to a lack of documentation, it went unused (and eventually rebuilt). Design decisions were often documented across a variety of sources, including the original feature specs they were built from, Slack comments, and Jira ticket updates. There was no real central source of information that either team could point to in order to help new folks joining the team get a good understanding of the app ecosystem.
In addition to the new Livebook documentation, we updated how we structured our Figma projects — creating individual library files for Foundations, Components, Patterns, Charts/Data, and Illustrations. Each library shared a page template outlining its purpose, relevant links in-app, usage guidelines, design specs, decision logs, and more. Both the code and design resources were pinned in all of our shared spaces, and added into our onboarding processes for new teammates.

Documenting our work
Making sure that the next folks understand, too.
A problem we kept running into across departments was that knowledge eventually became siloed — a single person might have championed an accessibility initiative, or written a great feature update, but due to a lack of documentation, it went unused (and eventually rebuilt). Design decisions were often documented across a variety of sources, including the original feature specs they were built from, Slack comments, and Jira ticket updates. There was no real central source of information that either team could point to in order to help new folks joining the team get a good understanding of the app ecosystem.
In addition to the new Livebook documentation, we updated how we structured our Figma projects — creating individual library files for Foundations, Components, Patterns, Charts/Data, and Illustrations. Each library shared a page template outlining its purpose, relevant links in-app, usage guidelines, design specs, decision logs, and more. Both the code and design resources were pinned in all of our shared spaces, and added into our onboarding processes for new teammates.

Building the team
Built, maintained, and championed by a core team.
The folks working on the design system updates were volunteers from the team who had already been advocating for consistent component design within their individual scrum teams. As interest in the project grew, I started hosting weekly Design System meetings, where design and engineering came together to talk about the progress of component work and give feedback on design and implementation. Eventually, we came to have a team of a few core members who were interested in building and supporting the system.
As the team grew, I also acted as project manager for the Design System — developing a feature roadmap, creating and updating tasks in Jira, maintaining a backlog, and updating key stakeholders on our progress.
Building the team
Built, maintained, and championed by a core team.
The folks working on the design system updates were volunteers from the team who had already been advocating for consistent component design within their individual scrum teams. As interest in the project grew, I started hosting weekly Design System meetings, where design and engineering came together to talk about the progress of component work and give feedback on design and implementation. Eventually, we came to have a team of a few core members who were interested in building and supporting the system.
As the team grew, I also acted as project manager for the Design System — developing a feature roadmap, creating and updating tasks in Jira, maintaining a backlog, and updating key stakeholders on our progress.
Building the team
Built, maintained, and championed by a core team.
The folks working on the design system updates were volunteers from the team who had already been advocating for consistent component design within their individual scrum teams. As interest in the project grew, I started hosting weekly Design System meetings, where design and engineering came together to talk about the progress of component work and give feedback on design and implementation. Eventually, we came to have a team of a few core members who were interested in building and supporting the system.
As the team grew, I also acted as project manager for the Design System — developing a feature roadmap, creating and updating tasks in Jira, maintaining a backlog, and updating key stakeholders on our progress.
Keeping the momentum
Maintaining buy-in from developers and the team at large.
In order to help fight design system burnout, I started with educating others about the system and its benefits. I started up optional Lunch-and-Learn sessions for fun CSS tricks, design demos in Figma, and code walkthroughs for upcoming components — extending invitations to both engineering and the team at large. I also gave presentations on the development of the system to the full team to help folks understand the initiatives design was trying to tackle, and how they benefitted both our users and the internal team’s process.
Team members were also hesitant to learn another development process where they might have to stare at a wall of text to digest what’s going on — so we started scheduling open office hours for the Design System, where at least one member of the design team would be present to help folks as they learned to use the new component library and offer support where needed.
Creating our component spreadsheets early on and then developing a project management system for the Design System also made it easy to provide tangible metrics for stakeholders, such as: reduced lines of code, number of bugs, hours saved developing functionality, etc.

Keeping the momentum
Maintaining buy-in from developers and the team at large.
In order to help fight design system burnout, I started with educating others about the system and its benefits. I started up optional Lunch-and-Learn sessions for fun CSS tricks, design demos in Figma, and code walkthroughs for upcoming components — extending invitations to both engineering and the team at large. I also gave presentations on the development of the system to the full team to help folks understand the initiatives design was trying to tackle, and how they benefitted both our users and the internal team’s process.
Team members were also hesitant to learn another development process where they might have to stare at a wall of text to digest what’s going on — so we started scheduling open office hours for the Design System, where at least one member of the design team would be present to help folks as they learned to use the new component library and offer support where needed.
Creating our component spreadsheets early on and then developing a project management system for the Design System also made it easy to provide tangible metrics for stakeholders, such as: reduced lines of code, number of bugs, hours saved developing functionality, etc.

Keeping the momentum
Maintaining buy-in from developers and the team at large.
In order to help fight design system burnout, I started with educating others about the system and its benefits. I started up optional Lunch-and-Learn sessions for fun CSS tricks, design demos in Figma, and code walkthroughs for upcoming components — extending invitations to both engineering and the team at large. I also gave presentations on the development of the system to the full team to help folks understand the initiatives design was trying to tackle, and how they benefitted both our users and the internal team’s process.
Team members were also hesitant to learn another development process where they might have to stare at a wall of text to digest what’s going on — so we started scheduling open office hours for the Design System, where at least one member of the design team would be present to help folks as they learned to use the new component library and offer support where needed.
Creating our component spreadsheets early on and then developing a project management system for the Design System also made it easy to provide tangible metrics for stakeholders, such as: reduced lines of code, number of bugs, hours saved developing functionality, etc.

Impact
We built trust with our users and teammates by giving them a consistent, accessible, and easily understandable user experience.
Wins for our users
Clear and consistent user experiences reduced friction across the product suite.
A new, easy-to-use app navigation, with clear categories for budgeting tasks
Increased table functionality across the app, including sorting, filtering, grouping, and more
Keyboard-accessible components, for faster data manipulation at a large scale
Modern, legible dashboards and data visualization, ready for presentation to district officials
Wins for the team
Increased cross-collaboration and built a scalable, reusable component library.
Interactive library of reusable code components alongside documentation and usage guidelines
Accessible component patterns with built-in testing during the PR process
Reduced time for developing new features and extending functionality to existing features
Increased team bonding and product ownership for design, engineering, and product at large
Impact
We built trust with our users and teammates by giving them a consistent, accessible, and easily understandable user experience.
Wins for our users
Clear and consistent user experiences reduced friction across the product suite.
A new, easy-to-use app navigation, with clear categories for budgeting tasks
Increased table functionality across the app, including sorting, filtering, grouping, and more
Keyboard-accessible components, for faster data manipulation at a large scale
Modern, legible dashboards and data visualization, ready for presentation to district officials
Wins for the team
Increased cross-collaboration and built a scalable, reusable component library.
Interactive library of reusable code components alongside documentation and usage guidelines
Accessible component patterns with built-in testing during the PR process
Reduced time for developing new features and extending functionality to existing features
Increased team bonding and product ownership for design, engineering, and product at large
Impact
We built trust with our users and teammates by giving them a consistent, accessible, and easily understandable user experience.
Wins for our users
Clear and consistent user experiences reduced friction across the product suite.
A new, easy-to-use app navigation, with clear categories for budgeting tasks
Increased table functionality across the app, including sorting, filtering, grouping, and more
Keyboard-accessible components, for faster data manipulation at a large scale
Modern, legible dashboards and data visualization, ready for presentation to district officials
Wins for the team
Increased cross-collaboration and built a scalable, reusable component library.
Interactive library of reusable code components alongside documentation and usage guidelines
Accessible component patterns with built-in testing during the PR process
Reduced time for developing new features and extending functionality to existing features
Increased team bonding and product ownership for design, engineering, and product at large
Final thoughts
Redesign doesn’t have to mean “start from scratch”.
Redesigning an application sounds like a lot of work, and it is — but it doesn’t have to mean throwing away everything you’ve worked on for years and starting completely from scratch. By nature, people are often resistant to change; even though a new interface might mean a sleeker experience with better functionality, users sometimes lament any change to an app simply because they’ve gotten used to the old design and its quirks.
That’s why redesigning an app shouldn’t have to mean starting from zero. Pick the elements that define your app now (a single heading, a widget on your dashboard) — identify the parts of it that work, how it could be better, and make small, incremental changes to it. Change can (and often should!) be gradual; it helps users get accustomed to the new flow, and doesn’t leave them confused or distressed that the tool they use for their day-to-day work has shifted dramatically overnight.
Make a plan, make it visible, keep folks engaged, and celebrate every little win along the way — your team and your users will thank you.
Further resources for building your design system
Color, Type, + Spacing
Code components
Accessibility testing
Design systems
Final thoughts
Redesign doesn’t have to mean “start from scratch”.
Redesigning an application sounds like a lot of work, and it is — but it doesn’t have to mean throwing away everything you’ve worked on for years and starting completely from scratch. By nature, people are often resistant to change; even though a new interface might mean a sleeker experience with better functionality, users sometimes lament any change to an app simply because they’ve gotten used to the old design and its quirks.
That’s why redesigning an app shouldn’t have to mean starting from zero. Pick the elements that define your app now (a single heading, a widget on your dashboard) — identify the parts of it that work, how it could be better, and make small, incremental changes to it. Change can (and often should!) be gradual; it helps users get accustomed to the new flow, and doesn’t leave them confused or distressed that the tool they use for their day-to-day work has shifted dramatically overnight.
Make a plan, make it visible, keep folks engaged, and celebrate every little win along the way — your team and your users will thank you.
Further resources for building your design system
Color, Type, + Spacing
Code components
Accessibility testing
Design systems
Final thoughts
Redesign doesn’t have to mean “start from scratch”.
Redesigning an application sounds like a lot of work, and it is — but it doesn’t have to mean throwing away everything you’ve worked on for years and starting completely from scratch. By nature, people are often resistant to change; even though a new interface might mean a sleeker experience with better functionality, users sometimes lament any change to an app simply because they’ve gotten used to the old design and its quirks.
That’s why redesigning an app shouldn’t have to mean starting from zero. Pick the elements that define your app now (a single heading, a widget on your dashboard) — identify the parts of it that work, how it could be better, and make small, incremental changes to it. Change can (and often should!) be gradual; it helps users get accustomed to the new flow, and doesn’t leave them confused or distressed that the tool they use for their day-to-day work has shifted dramatically overnight.
Make a plan, make it visible, keep folks engaged, and celebrate every little win along the way — your team and your users will thank you.
Further resources for building your design system
Color, Type, + Spacing
Code components
Accessibility testing
Design systems
Case studies
All content is prohibited from being used in any manner pertaining to AI asset generation, AI training, or machine learning
© Rizzo Daly, 2019–2026
Designed by Rizzo with ❤️🔥
Case studies
All content is prohibited from being used in any manner pertaining to AI asset generation, AI training, or machine learning
© Rizzo Daly, 2019–2026
Designed by Rizzo with ❤️🔥
Case studies
All content is prohibited from being used in any manner pertaining to AI asset generation, AI training, or machine learning
© Rizzo Daly, 2019–2026
Designed by Rizzo with ❤️🔥