This case study is protected

Available on request to recruiters and hiring managers. Enter the password to continue.

No password? Request access.

Back to work

Design Systems at Meta

Accessibility as a property of the system

Turning accessibility from a product afterthought into a built-in property of how Facebook's design system ships, across five years, 400+ components, and a global user base of billions.

Visual — hero Abstract component-library motif with accessibility layers overlaid (focus order, contrast, screen-reader labels).
Role
Senior Product Designer, Accessibility
Timeline
2020 – Present
Apps
Facebook, Instagram, Messenger, WhatsApp
Scope
Accessibility, Design Systems, Cross-platform

When I joined the accessibility team, Facebook's design system, the foundation thousands of designers and engineers build on, had a weak accessibility foundation. This left product teams to solve most of the accessibility work (WCAG AA) on their own, producing uneven experiences for people who rely on assistive technology at the scale of billions. I led the multi-year effort to close that gap by fixing accessibility once, in the components and primitives everyone already used, rather than endlessly product by product.

Impact at a glance

Coverage

Raised accessibility compliance from a small minority of the library to the overwhelming majority of components, across every framework.

Process

Made accessibility a launch requirement in the system's definition of done, and leveraged component adoption goals to scale that impact across the app.

Business

Shipped accessibility changes to the production app that also increased core engagement metrics.

Scale

Trained the design system's own designers to own accessibility, moving myself out of the critical path.

The arc

Four phases took this from an idea to an owned practice: identify the gap, operationalize it, drive the numbers, then institutionalize it. I started by creating an accessibility source of truth for the component library, built the process and leadership mandate that made it routine, drove compliance and proved the business case, and finally made it durable enough to run without me.

Visual — diagram Maturity staircase rising from "no accessibility layer" to "institutionalized," with the four phases as steps.

Five years, on a timeline

Select a year to see what shipped and which phase of the arc it belonged to.

Identify

2020 · Found the gap

Discovered that the core component library had no accessibility spec, and initiated the first source of truth for how components should behave when used with a screen reader.

01Earning leadership buy-in

For years, accessibility lived at the end of the process, a best-effort pass after the real decisions were made, and the first thing cut when timelines tightened. The opportunity was to move it from optional to mandatory. I worked to secure leadership backing to make accessibility a requirement for component launch, writing it into the definition of done so it couldn't be skipped. Later I made the case in leadership's own language of product performance, showing that accessibility improvements also benefited the broader product, which shifted the conversation from "accessibility as compliance" to "accessibility as a growth lever."

ImpactAccessibility became a required, leadership-backed step in shipping any component, and the organization's narrative around it shifted from cost to investment.

How I actually built that case, the specific data, the rooms, the objections I handled, is walked through live in an interview.

02Building a process teams could own

Fixing a set of components is finite work; keeping a living, constantly-growing system accessible is not. The opportunity was to make accessibility a routine the teams could run themselves, rather than a gate I personally manned. I built a repeatable loop (spec, review, annotate, hand off, dogfood) and embedded it into the design-system and product teams' existing workflows instead of bolting on a separate review. I documented the standard so a designer facing an accessibility decision had an answer without having to find me.

Visual — diagram Review-to-ship loop (spec → review → annotate → handoff → dogfood → compliance).

ImpactAccessibility moved from one-off quality fixes to a standing step in how components are designed, reviewed, and shipped, now repeatable across teams and no longer dependent on me. Cross-functional partners on the design system started to become accessibility advocates for the system and across Facebook.

03Accessibility as a growth lever

Accessibility's biggest enemy wasn't disagreement. It was the quiet assumption that it's a cost center: worth doing, but not the highest priority. The opportunity was to disprove that with product data. Several accessibility changes shipped to the production app increased core metrics rather than trading against them: content protection on video, more legible interactive text, and higher-contrast components. I packaged these into recurring "impact of accessibility" reports that showed, example by example, that designing for people with disabilities made the product measurably better for everyone.

Visual — chart + shipped components Directional chart of shipped accessibility changes vs. engagement direction (increases only, no values), plus imagery of the shipped production components.

ImpactAccessibility was reframed company-wide from a compliance program into a measurable growth lever, with senior leadership amplifying the case.

The specific metrics live in my interview deck.

04Scaling myself out

The opportunity here was also the risk: on a system far too large for one person, I could easily have become the single point of failure for accessibility. The senior move was to make myself unnecessary. I trained the design system's own designers to own accessibility, mentored designers across the company, and deliberately transitioned issue ownership from my team to the product and system teams, moving from owning the work to supporting it. To scale past a single system, I co-built an accessibility maturity framework in partnership with central teams that let cross-company design systems assess and improve on their own.

Visual — diagram Hub-and-spoke to distributed-network diagram. Me as the central node early, ownership distributing outward as I fade from the center.

ImpactThe system's own designers could deliver accessible components confidently without me in the loop, and the practice extended to design systems across the company.

Reflection

Five years on one problem, at this scale, seems like a lot. And it was. It changed how I think about building products.

Three lessons stuck. At this scale, quality is won or lost in the system, not the screen, so the only fixes that hold are the ones you push upstream, into the components and the definition of done. A quality investment only survives when it's framed in the language the business already uses: don't ask the company to value what you value; show how it advances what they already measure. And the most senior thing I did was make myself unnecessary, handing the work to the people and frameworks that could carry it without me.

If I had to compress it to one principle for building at this size: invest in what produces the outcomes, not the outcomes themselves. The component quality and the metrics were the visible results. The real work was leaving behind a design system, and a team to carry it, that treat accessibility as their own. And the part I'm proudest of is what our investment produced: an app used by billions, and the design system beneath it, that is meaningfully more accessible for the people of all abilities.