Playbook

How I lead

Vision without an operating model is a poster. This page is the model: how the organization was built, how it's shaped, what strategy it runs on, and how I know whether it's working.

The philosophy under the whole model is three moves: pave the paths, hold the taste, connect the dots. What follows is how each one is run.

1. Team from zero

I joined Clarivate in 2018 as the first UX designer for what is now the Life Sciences & Healthcare segment, one senior designer embedded in a Tech and Product organization. The years that followed brought acquisitions and reorganizations. UX grew considerably and was centralized under a dedicated company-wide leader for a period, then decentralized again when Clarivate segmented in 2024. Through every configuration one thing stayed constant: I carried the responsibility for Life Sciences & Healthcare UX, first informally, then as Principal, then as Director. And when the segmentation dissolved the central function, the segment needed a UX organization that could stand on its own. We built that one too.

Two of the org's three disciplines I founded outright:

Research. I brought the UX researcher role into the company. I made the case, built the awareness, created the demand, and personally hired every researcher on the team. The practice that now reaches 71% of the portfolio and gates our AI features simply did not exist before.
UX engineering. When Clarivate's 2024 segmentation gave LS&H the mandate to own its own tooling, I seized the window: we built Helix, our design system, from the ground up. It is now the standard for the segment's products, and the team that builds it is the third leg of the org.

Today the function is 16 people across three disciplines. It grew because the work created pull, and it stays because the work is worth staying for: 93% retention through a period when our scope roughly doubled.

→ The discipline's full story: Research as an operating system · the sequencing argument: Leverage before headcount

2. Three legs, one discipline

The org is deliberately multidisciplinary, with product design, UX engineering, and research under one roof:

Product design owns the experience of the portfolio and holds the craft bar as products modernize onto the design system.
UX engineering builds the infrastructure: the design system's tokens, components, and pipelines, and a Dev Guild with shared libraries, PR reviews, and a release cadence. For three consecutive quarters this group has run at 100% priority alignment. Deep work on one strategic thing at a time, not scattered support tickets.
Research runs discovery and validation across the portfolio, and operates as the quality gate for every AI feature we ship.

Why one roof? Because the problems that matter land between the legs. An AI pattern library is design plus engineering. Synthetic personas are research plus engineering. A validation gate is research plus design. Separate these disciplines and every real problem becomes a coordination problem.

Three legs, one discipline. The accent edges are the argument: the work that matters lives between the disciplines, so an org chart that separates them turns every real problem into a coordination problem.

The advocacy ladder. None of this structure was granted. It was earned in stages, and the stages form a ladder any UX leader will recognize:

Teach. I ran "UX 101" sessions in executive meetings and evangelized customer journey mapping product by product, back when UX still had to explain itself.
Institutionalize. UX practices were written into the organization's product-development handbook and its training matrix, so the method outlived the meeting.
Demand pull. Today, executives review UX metrics in monthly execution reviews, and product teams request research and design-system components unprompted.

The ladder's endpoint is the point: advocacy done right ends with the organization asking for UX, not being sold it. (→ The ladder as an essay: Stop selling UX)

3. Strategy: BUILD → SCALE → MEASURE

I run the org on an explicit multi-year maturity model, not an annual wish list.

The maturity model as an operating strategy, not a slogan: each phase has an exit condition, and the current year deliberately overlaps the next phase's start.

BUILD: stand up the infrastructure and create the demand. Outcome: a design system from zero to 59% designer adoption with five products live; research demand up 114% (7 → 15 products, 71% of the active portfolio); researchers repositioned to AI validation; strategic alignment up 21 points.
SCALE: expand what's proven. Design system toward 85% of the portfolio; research toward near-universal practice with a real research-ops layer (external panels, templates, SLAs); the AI pattern library shipped so proven patterns travel between products.
MEASURE: connect UX investment to business outcomes. Design-system ROI in time-to-market, research investment correlated with product adoption, an executive dashboard that speaks revenue and retention.

I will say the honest part out loud: most UX orgs, mine included, have historically run on leading indicators like adoption, penetration, and alignment. The MEASURE phase exists because "trust us" is not a durable position for a design org, and I would rather build the measurement machinery than inflate the claims.

4. Running the org on evidence

I instrument my organization the way a product manager instruments a product.

Priority alignment, the share of team effort landing on the organization's stated priorities, is measured quarterly from the ground truth of actual time data, not self-reporting. It moved from 87% to 93% over the year, with a 97% peak, and held through year-end pressure. Utilization, context-switching, allocation by business unit, and research coverage get the same treatment: roughly twenty analyses run against the org's own operational data, maintained as code.

The numbers change decisions, not just slides. Context-switching data set WIP limits (3 to 4 concurrent projects per person, one per UX engineer). Coverage data justified hiring a researcher before demand was proven, a bet the demand curve then validated. Quarterly dips get diagnosed and corrected, not narrated away. When multi-business-unit coverage sagged in one quarter, the data showed it, and the next quarter delivered the breakthrough.

This habit predates the tooling boom. I pushed the team onto product analytics and fake-door experiments years before "data-driven design" needed an AI angle. (→ The argument in full: Instrument your org like a product)

The evidence for all of it: five case studies →