Case study 1: Advantage+ creative
Overview
Advantage+ creative is Meta's primary surface for creative automation, an AI-powered system that generates personalized ad creative at scale and one of the top annual run rate products in the Advantage+ brand. Between 2023 and 2024, the surface grew from 3 core features to nearly 30. My job was to make sure the content system grew with it.
The challenge
Rapid feature growth created two compounding problems. First, the sheer number of enhancements created high cognitive load for advertisers; Too many options, too little clarity about what each one did. Second, a lack of content consistency across the surface meant features were being misunderstood, which eroded trust in the automation doing work on advertisers' behalf.
The core question: how do we make non-creative advertisers comfortable with creative enhancements they don't fully control?
My role
I was the content design lead for Advantage+ Creative, responsible for both the craft and the infrastructure. That meant working directly in Figma with product designers, presenting to director-level stakeholders, and collaborating with UX research, product managers across numerous teams, and engineering to build solutions that would hold up at scale.
What I Did
Developed a feature taxonomy — from proposal to research
The existing feature organization had no principled structure. I drew on domain expertise to propose three candidate taxonomies:- Element-based — grouping by what part of the ad is affected (e.g. "Media," "Text")
- Job-to-be-done-based — grouping by advertiser intent (e.g. "Highlight products," "Tailor to placements")
- Risk-based — grouping by how much the feature changes the original creative (e.g. "Tune-ups," "Enhancements," "Transformations")
I worked with taxonomists to pressure-test each structure against standard hierarchical principles — right-sizing the number of features per group, ensuring groupings were mutually exclusive and collectively exhaustive — then iterated with product design to prepare concepts for research.
The winner: an element-based taxonomy with a dedicated breakout group for high-risk features ("For review"), which earned overwhelming advertiser preference in testing. The result was measurable: Fewer feedback instances, higher preview-aware adoption, and a decrease in both step dwell time and time to ad publish.
Built a content playbook to scale quality across teams
Research and consultation with our UX researcher surfaced three recurring comprehension gaps across the surface: features lacked descriptions, names were unclear, and placement language was absent entirely. I used those findings as the foundation for a content playbook built to fix the problems at the source.On the naming side, I partnered with engineering to set up agile content testing to evaluate naming hypotheses — specifically, noun-based names (e.g. "Music") versus verb-based names (e.g. "Add music") — and used the results to establish an evidence-backed standard.
On the governance side, I worked with product managers across teams to identify the friction points slowing down feature reviews: unclear rules around beta features, localization gaps, inconsistent conventions. The playbook addressed all of it directly.
The impact was operational as much as it was craft: the average service level agreement for governance dropped from 3 weeks to 1 week. Advertisers in research overwhelmingly preferred the new content patterns. And adoption rose across major locales including Hindi and Brazilian Portuguese, thanks to increased surface-specific localization guidance built into the playbook.
Established transparency and status standards across the surface
Beyond naming, I introduced a new transparency message pattern that clearly communicated to advertisers what enhancements had been turned on or updated — and in what state. I also established standards for glanceable feature status, giving advertisers a scannable, consistent way to understand the surface at a glance without needing to read into each feature individually.Impact landed
-
Established naming and terminology standards across a surface that grew 10x in feature scope
-
Delivered a content playbook that enabled cross-functional teams to write to bar without a content designer in every room
-
Introduced a feature taxonomy that improved consistency and reduced user confusion, validated through usability research and live experiments
- Helped maintain Advantage+ creative's position as one of Ads Manager's top revenue-driving surfaces by ensuring users could understand and trust the automation working for them
Enhancement begins at media selection, as we encourage a robust amount of creative to mitigate creative fatigue. Selection within the creative assignment modal required a crisp focus on task orientation and I led the development, testing and refinement of a task communication framework (”Choose up to 10 images and videos”) and sourcing paradigm (”Media from you” and “Media from AI”).