Content that holds
complex things
together.

The challenge

Events had to work across three new and very different surfaces simultaneously, each owned by a separate team with its own roadmap, design system, and constraints. A user might encounter the same event across multiple surfaces in a single session.

Without a shared content framework, decisions made for one surface risked creating confusion for the people actually trying to find and join events. Delivering this work meant managing communications across all three teams, syncing the framework to their individual roadmaps, and ensuring that what I was building fitted into designs that were evolving in parallel.

My approach

I developed a 'metadata first' approach: harnessing user research to identify the minimum metadata someone needs to make a decision on whether to join, subscribe or invite someone to an event, and then which additional metadata could persuade them to act. I created a single source of truth for event metadata across all surfaces, covering every state (e.g. upcoming, on now, friend-invited, and more).

I worked with XFN (PM, PD, engineering) to align decisions. Throughout, I worked as the connective tissue between three separate product teams running syncs, managing dependencies, and making sure the framework remained coherent as each surface's design evolved.

The framework also unblocked downstream work on the events store shelves, where I went on to shape the copy and the product decisions on what each shelf displayed across states and surfaces.

The outcome

Delivered a coherent events metadata framework across all three surfaces in time for soft launch, ahead of Meta Connect. The metadata-first approach was adopted as the shared design principle across the XFN team, and the source of truth became the reference point for engineering and QA.

The downstream shelves work, using the framework, produced measurable lifts across key engagement metrics.

The challenge

Workplace Live was used by a wide range of people, from solo creators to large enterprise customers running company-wide all hands and Q&As. Across all of them, there were too many names for too few real roles. Some terms (moderator, co-host) described jobs that didn't technically exist as backend roles. Others (presenter, host, speaker) were used interchangeably depending on who was writing the copy.

The result was a product that contradicted itself across surfaces, and a team that couldn't have clear conversations about who did what.

My approach

Following a user roles workshop, I defined two roles that mapped cleanly to backend permissions: Host (the person whose name is on the broadcast) and Organizer (anyone with permission to create and manage broadcasts on behalf of a host).

I wrote definitions, usage rules, best-in-class user statements, and jobs-to-be-done for each role, produced cheat sheets for XFN reference, and identified and fixed specific strings using the old language.

The outcome

The two-role model was adopted across the Live product, socialised to 92 members of the XFN group, and cheat sheets became the go-to reference for design, engineering, and PM. The completion rate improvement reflected a clearer, more confident experience for both hosts and organizers from the moment they set up a broadcast. Users could understand their permissions, know what was expected of them, and get to the moment of going live without friction or confusion about who was supposed to do what.

The simplification had benefits beyond the user-facing experience: by aligning the 6 notional UI roles to the two that actually existed, it also reduced backend complexity, and removed a layer of confusion between what engineers were building and what users were seeing. The new experience has been used since then for all company-wide Meta Q&As.

The challenge

Events 2.0 was a large, complex product that lived across multiple surfaces and contexts. There was no shared framework for identifying which content decisions needed to be kept consistent across all of them, and which could vary. This made it easy to make locally sensible decisions that created inconsistency at a product level.

For users, this meant encountering inconsistent language, labels, and interactions across the same product: the kind of small, hard-to-name friction that makes an experience feel unreliable without users being able to say exactly why.

My approach

I mapped the critical user journeys across Events, noting every element or feature we encountered. I stack-ranked these against a list of surfaces and crossover points, then grouped the results by theme. This produced an initial list of 7 primitives: Event details, User roles, Audience engagement, Invites and sharing, Entry points, Live status, and Technical options. I documented each with its scope, key decisions, and open questions, then socialised the framework across the XFN group as a living resource.

The outcome

The 7 primitives gave the wider team a clear mental model for understanding where content decisions had cross-surface implications. The living document meant XFN could reference it directly to unblock themselves without pulling in a content designer every time.

By resolving these decisions at a systemic level, the work ensured that users experienced Events 2.0 as a coherent product regardless of which surface they entered from, removing the low-level cognitive load that comes from a product that doesn't quite feel like itself.

The challenge

Terminology underpins everything content designers do at Meta. It keeps language consistent across products, prevents legal and compliance risk, and lets teams have clear conversations about what they're building. But despite a formal process for submitting and reviewing terms, submissions across the org were low and CDs weren't engaging. As Term Lead, I needed to understand why before any intervention could work.

My approach

I led a team of three junior content designers to investigate, over six months. The research surfaced two consistent reasons CDs weren't submitting: they lacked confidence in how the process worked, and many felt it was "too late to ask." They'd been in the org long enough that asking basic questions felt exposing.

We designed and ran Term amnesty sessions: no-judgement training where any CD could come and learn how, when, and why to submit terms, regardless of how long they'd been in role. We paired the sessions with digital handouts CDs could refer back to.

The amnesty work surfaced something else: there was no shared resource for CDs joining a new product area to get on top of terminology from scratch. Every CD was reinventing the process. I responded by recruiting one of the junior CDs to co-author a four-part Term Hygiene guide with me: how to audit existing terms, how to set up and share a term tracker, how to run a taxonomy exercise to separate product-defined terms from internal noise, and how to record and maintain terms over time.

The outcome

The amnesty sessions and handouts produced significant shifts in terminology engagement across the org (see stats below).

The Term Hygiene guide was incorporated into three permanent channels: new-hire onboarding for all content designers at Meta, term-lead training, and the translation toolkit. The co-author received a strong end-of-year review for their work on the guide.

By embedding the resources into the discipline's infrastructure, the work raised the floor for how the whole CD function approached terminology, making it easier to onboard and develop new CDs as teams scaled, and reducing the risk of inconsistent or legally ambiguous language reaching users at scale.