Ads teams were optimizing in silos — no shared model of the user journey.

Placement, creative, conversion — each team owned a piece, but no one had a shared model of the journey users actually take from "I see an ad" to "I bought something." Decisions felt local; tradeoffs across phases were invisible.

Synthesis across past research, paired with DS behavioral mapping.

Instead of a single new study, I led a synthesis effort across past Ads org research — and partnered with Data Science to make the journey instrumentable at runtime.

Five phases. Each one with a job, a need, a pain, and a metric.

Trigger → Consume → Comprehend → Consider → Purchase. The DS partnership is what made it usable: we didn't just describe the journey in slides — we made it instrumentable. Teams could now ask "which phase are we serving?" before debating creative or placement.

From a synthesis deliverable to org-wide grounding truth.

The journey was adopted as the canonical model for ads strategy across Instagram leadership — quoted in PRDs, used to scope new research, and embedded into the dashboards teams look at every day.

Product teams kept saying "native" — but couldn't define it.

Across PM, Creative, and Advertiser teams, "native" was the word used to describe what they wanted ads to feel like. No one could articulate it from the user's POV, and no one could prove it was worth investing in.

Two milestones. Qual built the framework. Quant validated it at scale.

A mixed-method program designed to first define nativity from the user's POV, then measure and prioritize the drivers globally.

Native ads sit at the intersection of four pillars.

Synthesized from the qual work: Signals · Timing · Content · Format. The quant pass validated and prioritized each pillar's drivers.

Example · Participant upload · IG Reels ad rated highly native

P0Relevance

25%

P1Real-person presence

20%

P1Casual language

18%

Other

13%

Looks like one of my friends could have made it. There's nothing professionally showy that makes it stick out. P17 · Power user · Adult cohort

Three teams, three concrete product bets.

The framework anchored Feed and Story ad optimizations — including Creator-Partnership ads — and three teams each owned a slice of the resulting roadmap.

Where does Instagram's shopping experience lead — and where does it lag?

Shoppers don't compare products in just one place. They cross between Amazon, Google, ChatGPT, TikTok, and Instagram — each with its own shopping superpower. Instagram leadership needed a clear-eyed view of where IG already leads peer platforms, where it lags, and which gaps were worth closing first.

An AI-native research workflow — UXR-defined criteria, AI-scaled analysis.

I defined the evaluation rubric and seeded the training set by hand; then built a Claude Code skill to scale content analysis across competitors. The final dashboard is what IG leads now use to track competitor feature availability over time.

Where the field converges — and where the gaps live.

Each row is a feature; each column is a platform. Filled dots mean fully present, half-dots mean partial. The grouped severity callouts surfaced which gaps were driving real shopper drop-off — anchoring the prioritization conversation in actual behavior, not opinion.

An AI-native workflow adopted by IG leadership.

Both the audit framework and the Claude Code skill became part of how IG leads scan the competitive AI landscape — not a one-off deck, but a living system the team now runs every quarter.

An ambiguous problem space — sales needs weren't well-defined.

Meta's global sales teams managed advertiser relationships through CRM tools, but the end-to-end workflow and day-to-day frictions weren't well-understood across product. With teams spanning NAM, EMEA, and APAC and roles ranging from Account Managers (mid-size advertisers) to Client Partners (large-scale strategic accounts), CRM decisions were being made without clear ground truth on how sales actually work — and the resulting needs were only partially addressed.

International in-depth interviews — across regions and roles.

Remote IDIs with sales across NAM, EMEA, and APAC — Account Managers (mid-size advertisers) and Client Partners (large-scale strategic accounts). Each session walked through the end-to-end quarterly workflow, tools used at each stage, and frictions hit along the way.

5 stages, each with its own frictions — and role-specific needs.

Sales work moves through five consistent stages. Frictions show up at every stage — shaped by role, by region, and by the tools sales reach for in the moment. CRM was supporting some stages well; others were left to manual workarounds.

From workflow gaps to product priorities.

The IDI findings turned an ambiguous problem space into concrete product priorities — anchoring the next round of CRM investment around the stages and roles where sales needs were most underserved.