Helping people get jobs by
scaling positive outcomes
Helping people get jobs by scaling positive outcomes
Helping people get jobs by scaling positive outcomes
A DESIGN SYSTEM STORY
A DESIGN SYSTEM STORY
background
The sea of blue problem
The sea of blue problem
In 2021, Indeed's Creative team introduced Aurora, a new brand identity, to modernize Indeed’s visual identity. By 2024, however, user feedback signaled that Aurora was no longer meeting the moment.
External hiring macro-trends had cooled from the post-pandemic peak in 2022. In parallel, our internal product experience had become an obstacle; employers felt overwhelmed by a journey cluttered with competing priorities.
Data showed that our UI had become too noisy, with excessive, primary call-to-actions diluting the user’s path to success. Users were telling us, "Indeed has turned into a sea of blue" and "I feel like every page is trying to take me away from the one thing I want to do, make a hire."
In 2023, Aurora Uplift was launched as a cross-functional initiative between Product Design, Brand, and Design Systems to simplify the interface and lead users to perceived value.
In 2021, Indeed's Creative team introduced Aurora, a new brand identity, to modernize Indeed’s visual identity. By 2024, however, user feedback signaled that Aurora was no longer meeting the moment.
External hiring macro-trends had cooled from the post-pandemic peak in 2022. In parallel, our internal product experience had become an obstacle; employers felt overwhelmed by a journey cluttered with competing priorities.
Data showed that our UI had become too noisy, with excessive, primary call-to-actions diluting the user’s path to success. Users were telling us, "Indeed has turned into a sea of blue" and "I feel like every page is trying to take me away from the one thing I want to do, make a hire."
In 2023, Aurora Uplift was launched as a cross-functional initiative between Product Design, Brand, and Design Systems to simplify the interface and lead users to perceived value.
objectives
Align design system goals with business goals
Align design system goals with business goals
Aurora Uplift objectives
Reduce cognitive load and steer users away from platform distractions and directly toward positive outcomes (e.g. higher quality candidate-to-employer matches) and sponsored apply clicks.
Design System objectives
Deploy the Aurora Uplift theme in our next major, IFL 7.
Migrate 85% of Indeed's product surfaces to the Aurora Uplift theme.
Improve positive outcomes and sponsored apply clicks metrics.
Aurora Uplift objectives
Reduce cognitive load and steer users away from platform distractions and directly toward positive outcomes (e.g. higher quality candidate-to-employer matches) and sponsored apply clicks.
Design System objectives
Deploy the Aurora Uplift theme in our next major, IFL 7.
Migrate 85% of Indeed's product surfaces to the Aurora Uplift theme.
Improve positive outcomes and sponsored apply clicks metrics.
challenges
The catch
The catch
Product teams were under intense pressure to deliver features during a hiring slowdown; they viewed a design system migration as 'extra work' that didn't immediately move their specific business metrics.
IFL version drift
50% of current adopters were still on Indeed Foundations Library (IFL) 4.x or 5.x while current was 6.3.0. This meant we had to support multiple upgrade paths for our next major release, IFL 7.
Creative misalignment
Finding alignment with the creative direction was a bigger problem than it should have been. Not having a clear decision maker led to a shorter runway and wasteful meetings.
Shifting roadmap
I made decisions on trade-offs for priorities I roadmapped at the beginning of the year, before Uplift was announced. Luckily I had already planned a liberal amount of padding that I could dip into if needed.
Product teams were under intense pressure to deliver features during a hiring slowdown; they viewed a design system migration as 'extra work' that didn't immediately move their specific business metrics.
IFL version drift
50% of current adopters were still on Indeed Foundations Library (IFL) 4.x or 5.x while current was 6.3.0. This meant we had to support multiple upgrade paths for our next major release, IFL 7.
Creative misalignment
Finding alignment with the creative direction was a bigger problem than it should have been. Not having a clear decision maker led to a shorter runway and wasteful meetings.
Shifting roadmap
I made decisions on trade-offs for priorities I roadmapped at the beginning of the year, before Uplift was announced. Luckily I had already planned a liberal amount of padding that I could dip into if needed.
Lesson learned
Lesson learned
I had finalized our roadmap in February which was too early. Indeed had just officially transitioned to a fiscal calendar that aligned with it's parent company, Recruit. This meant Q1 technically started April 1 rather than January 1. I immediately course corrected which I go into more detail down below.
I had finalized our roadmap in February which was too early. Indeed had just officially transitioned to a fiscal calendar that aligned with it's parent company, Recruit. This meant Q1 technically started April 1 rather than January 1. I immediately course corrected which I go into more detail down below.
approach
Building the design system strategy
Building the design system strategy
Key design decisions
Restrict most components to a neutral color palette and reserve Indeed Blue for high-value interactions.
Roll out our new custom font type, Indeed Sans.
Add new tokens with gradient values that will indicate Indeed's messaging feature.
Add an AI icon to call out AI-assisted features within our product.
Release Aurora Uplift in two phases.
Fix adoption
I created a top-down directive by influencing senior leadership (Sr Dirs + VPs) to prioritize package upgrades.
The team educated, trained, and shared ownership/impact with stakeholders.
We empathized and found shared goals with product, design, and engineering management.
Added a 6 month migration timeline to our governance model.
Systematize the decision-making and task management processes
I introduced the RAPID framework for decision-making and RACI for task management. This provided clear ownership of decisions and roles within tasks. Used for the entirety of the initiative.
We accepted this is going to get a lil' bit scrappy. With an initiative this large that touched every surface of our business, we're going to have to rely on our design principles to navigate the ambiguity.
Maintain the stability of our system as well as keep healthy, transparent relationships with our current users
Even with a team heads-down on Uplift, I ensured folks had the capacity to continue to monitor our support channels and host office hours.
I was transparent with the teams impacted by sidelined priorities. I provided reasoning as well as alternative solutions like following our contribution process or exploring work-arounds.
Key design decisions
Restrict most components to a neutral color palette and reserve Indeed Blue for high-value interactions.
Roll out our new custom font type, Indeed Sans.
Add new tokens with gradient values that will indicate Indeed's messaging feature.
Add an AI icon to call out AI-assisted features within our product.
Release Aurora Uplift in two phases.
Fix adoption
I created a top-down directive by influencing senior leadership (Sr Dirs + VPs) to prioritize package upgrades.
The team educated, trained, and shared ownership/impact with stakeholders.
We empathized and found shared goals with product, design, and engineering management.
Added a 6 month migration timeline to our governance model.
Systematize the decision-making and task management processes
I introduced the RAPID framework for decision-making and RACI for task management. This provided clear ownership of decisions and roles within tasks. Used for the entirety of the initiative.
We accepted this is going to get a lil' bit scrappy. With an initiative this large that touched every surface of our business, we're going to have to rely on our design principles to navigate the ambiguity.
Maintain the stability of our system as well as keep healthy, transparent relationships with our current users
Even with a team heads-down on Uplift, I ensured folks had the capacity to continue to monitor our support channels and host office hours.
I was transparent with the teams impacted by sidelined priorities. I provided reasoning as well as alternative solutions like following our contribution process or exploring work-arounds.
FY2024 Initiative timeline
process
Getting into a rhythm
Getting into a rhythm
We met with Brand, UX, and Engineering leadership to build a repeatable workflow to ideate, create, validate, and deliver changes in our token and component libraries. This process allowed the design system team to work asynchronously as design decisions were being finalized. If iterations were needed, Leads would review and decide async so that we could focus on moving forward in our design critique/feedback meetings.
delivered
Theming Capabilities
Theming Capabilities
Design
Product design and prototypes utilizing our token library in Figma can toggle between themes within seconds.
Development
With the release of IFL 7, you simply manage a single string in your application state that matches the tokens.
Design
Product design and prototypes utilizing our token library in Figma can toggle between themes within seconds.
Development
With the release of IFL 7, you simply manage a single string in your application state that matches the tokens.
<div data-theme="aurora_uplift" className="app-wrapper">
delivered
Introduced a neutral color palette
Introduced a neutral color palette
Adjusting a variety of components to a neutral palette allowed the primary blue to have more prominence.
Adjusting a variety of components to a neutral palette allowed the primary blue to have more prominence.
delivered
Brought energy back to our primary blue
Brought energy back to our primary blue
Increased the prominence of core actions and reduced friction by providing a clearer visual hierarchy toward the user's end goal. We put these changes behind a 60-day user test to see how they performed against the control version.
Increased the prominence of core actions and reduced friction by providing a clearer visual hierarchy toward the user's end goal. We put these changes behind a 60-day user test to see how they performed against the control version.
delivered
Indeed Sans
Indeed Sans
Designed for scale
Indeed Sans supports 136 languages covering all Latin-based scripts.
Unexpected results
2.5% increase in positive outcomes and 2.6% increase in sponsored apply starts. $10 million a month increase in revenue.
Designed for scale
Indeed Sans supports 136 languages covering all Latin-based scripts.
Unexpected results
2.5% increase in positive outcomes and 2.6% increase in sponsored apply starts. $10 million a month increase in revenue.
delivered
AI feature identifier
AI feature identifier
Preparing for the release of Indeed's AI Assistants, the team designed and deployed a unique logo to help users identify agentic experiences throughout our products.
Preparing for the release of Indeed's AI Assistants, the team designed and deployed a unique logo to help users identify agentic experiences throughout our products.
delivered
Messaging-specific gradients
Messaging-specific gradients
We added gradient background as additional semantic tokens to our library that brings warmth and energy to Indeed's messenger experience.
We added gradient background as additional semantic tokens to our library that brings warmth and energy to Indeed's messenger experience.
delivered
Aurora Uplift
Aurora Uplift
Our plan was to deliver Aurora Uplift into you two separate phases. Release the theme first so that we could test as close to production as possible. Then, we'd release the breaking changes in our next major release. We actually released in three phases.
The unexpected lift in revenue that Indeed Sans was showing in it's test had executlive leadership Slacking me directly to deploy it as soon as possible. We threw it in a patch and released it the next day.
Our plan was to deliver Aurora Uplift into you two separate phases. Release the theme first so that we could test as close to production as possible. Then, we'd release the breaking changes in our next major release. We actually released in three phases.
The unexpected lift in revenue that Indeed Sans was showing in it's test had executlive leadership Slacking me directly to deploy it as soon as possible. We threw it in a patch and released it the next day.
results
We were on the right track
We were on the right track
Business impact
Our users were starting to see more value from Indeed. The test group generated $18 more per revenue-generating advertiser, totaling a $8.8M revenue lift during a 53-day time period (from April 16th - June 7th)
We increased Positive Outcomes by 5.4%. 2.5% of that coming from our typeface change alone.
On the Job Seeker side, we increased sponsored apply starts by 2.6%.
We observed a +2% revenue increase (stat sig, powered) which led to an annualized incremental revenue win of $140MM.
Design System impact
Reduced IFL version drift and reached an 80% adoption rate of IFL 7. For reference, we peaked at a 65% adoption rate for IFL 6.
Design system's visibility increase exponentially.
Business impact
Our users were starting to see more value from Indeed. The test group generated $18 more per revenue-generating advertiser, totaling a $8.8M revenue lift during a 53-day time period (from April 16th - June 7th)
We increased Positive Outcomes by 5.4%. 2.5% of that coming from our typeface change alone.
On the Job Seeker side, we increased sponsored apply starts by 2.6%.
We observed a +2% revenue increase (stat sig, powered) which led to an annualized incremental revenue win of $140MM.
Design System impact
Reduced IFL version drift and reached an 80% adoption rate of IFL 7. For reference, we peaked at a 65% adoption rate for IFL 6.
Design system's visibility increase exponentially.
Thanks for reading.
Thanks for reading.
A DESIGN SYSTEM STORY