How I Lead

Org design before headcount

When I joined the GoTo design organization in 2018, I had 5 designers reporting to me. Four years later, the team had grown to 14, across Europe and North America, spanning Junior to Principal, with managers underneath.

The headcount story is the one most leaders default to. It's the wrong story. The decisions I'm proudest of from that chapter weren't how many designers I hired, they were what kinds of seats I built, and when.

Three specific moves that mattered:

• Founding the Chameleon design system as an independent team. It started as a project inside one product team. I made the case to leadership that a design system serving a portfolio can't be owned by a single product, and built the org case for spinning it out. Chameleon now operates across the full GoTo portfolio, run by a team I helped found but no longer manage. Spinning a function out from under you is one of the harder org-design moves at this stage, and the right one when the function needs to outlast you.
• Creating the first Accessibility Lead role on the design team. Accessibility had been treated as a shared responsibility, which in practice meant no-one's. I argued for a dedicated role and grew Anna Mészáros into it from a mid-level IC seat. (Her LinkedIn recommendation tells the progression in her words.)
• Adding the first UX Engineer hire: a role that didn't exist on the design team and that the unified suite work demanded. We needed someone who could prototype in production code and own the design-engineering bridge. We hired one.

Headcount follows structure, not the other way around. If you find yourself adding designers and the team isn't getting more effective, the answer is almost never another designer.

Calibration is a craft, not an HR ritual

I run regular calibration with peer design managers and with research, content, and accessibility leads. I write growth maps in collaboration with each direct report, not from above, with them. The cadence is regular, the rubric is shared, the conversations are recorded so the bar holds across cycles.

The point of calibration is not to rank people. It's to agree on what good looks like before you have to make a promotion or hiring decision under pressure. Calibration done right means the answer is already in the room when the question arrives.

Under direct mentorship, people I've managed have moved from mid-level to Senior to discipline Lead, and from Senior IC to Manager. Those weren't accidents of opportunity; they were planned trajectories with structured feedback at each step. Anna Mészáros's recommendation on my LinkedIn (written from her vantage point as one of those reports, now in design leadership herself) speaks to that work in her own words. I'd encourage you to read what she wrote rather than what I'd write about her.

Decision rights, written down

I don't believe in design by consensus. I believe in clear decision rights, delegated as low as is responsible, and held accountable openly.

The way I structure it:

• Senior designers own visual-system calls and craft decisions within their product area: typography, color, component variants, interaction patterns.
• Design managers own staffing decisions for their team, and roadmap balance against Product and Engineering.
• I own the cross-product trade-offs, the executive partnership, and the org-design moves that affect more than one team.

When the room disagrees, the named owner decides; everyone else commits or escalates. Disagree-and-commit is the default, not the exception. Consensus is what teams do when no-one wants to take responsibility for the call; I'd rather make the call and be wrong sometimes than diffuse the responsibility and be slow always.

The corollary: when I'm the named owner, I have to actually decide. The fastest way to lose a team's trust at this level is to claim decision rights and then refuse to use them.

Operating rituals compound

Critique cadence. Design-review structure. 1:1 templates. Hiring-loop calibration. Onboarding course. Quarterly growth conversations. The handoff frame from design to engineering.

None of these are sexy. All of them compound.

At GoTo I built or rebuilt each of them over four years, sometimes more than once. I created the onboarding course for new UX hires (still in use after I left), co-created a learning-and-growth program for the broader UX organization, and stood up the design-leadership review rituals that kept the team coherent across two continents and the pandemic.

At Oodrive, I'm re-standing versions of each, adapted to a smaller team, a different stage, a French-language operating environment. The point isn't that the templates transfer. The point is that a design organization without rituals is held together by individual goodwill, which is the thinnest binding material in the world. The rituals are what let the team survive a key hire leaving, or a quarter going sideways, or a leadership transition.

If you want the templates, ask me. They're not trade secrets; they're commitments.

Cross-functional partnership at altitude

At Director level, design isn't about defending the discipline anymore. It's about being a peer to Product and Engineering leadership.

I co-author roadmaps with my Product counterparts; I don't react to them. I sit in the same strategic review forums as VPs of Product and Engineering, not the forums below those. When something is being decided that affects the design organization, I'd rather be in the room than be informed of the outcome.

The hardest part of the job at this level is earning that altitude without elbows. The version of design leadership that wins a seat at the table by being loud is also the version that loses it the moment a quieter, more trusted voice is available. I'd rather be useful than visible.

In practice, that looks like:

• Showing up to Product and Engineering review forums prepared: having read the docs, having a position, ready to be wrong publicly.
• Framing design problems as business problems with a design dimension, not as design problems that need business context wrapped around them.
• Taking the meetings I'd rather skip. The 1:1 with the new Head of Sales who doesn't know what UX research is, the partnership session with Procurement, these are where credibility is built or lost, not in the design-review forum where I'm comfortable.

People before product, evidenced

This is the principle I'd be most embarrassed to put in a beliefs list. It's a slogan on every design-leadership page on the internet, including mine until recently.

Here's the version I'll defend instead: healthy teams ship better products, and a team you've taxed to ship will give you exactly one quarter of velocity before quality drops. Building for the second year, not the first sprint, is the only operating choice that compounds.

Two examples of what that's looked like in practice, both from the GoTo years:

• During the early pandemic transition, when the team went fully remote overnight, I prioritized team continuity over delivery velocity for the quarter that mattered. Linda Kessler (a research partner I worked closely with through that period, peer to me under the same manager) describes it in her LinkedIn recommendation; I'd rather you read hers than mine.
• When a direct report needed support to relocate her family across borders, the company-side process for that wasn't simple. I made it work: followed it through with HR, immigration, and her manager-of-record, on a personal-stakes timeline rather than an org-policy timeline. Three years later she was running the accessibility function. The trust earned from showing up at the moments that aren't about work is what gives you the credit to ask for harder things when the work demands it.

People-first isn't a soft position. It's the longest-leverage operating choice I've ever made.

Measurement as a non-negotiable

If your design team isn't measuring, your design team isn't credible at the business table.

The tooling for design measurement is finally mature. SUS, CES, retention curves segmented by experience touchpoint, qualitative-quantitative pairings. Design teams can run their own dashboards now, without begging Analytics. I treat this as table stakes for any team I lead.

The caveats matter:

• Don't conflate satisfaction with success; they correlate, they don't equate.
• Don't optimize for a metric the team that owns the surface can game.
• Always pair quantitative with qualitative: a number without a story is a target, not a measurement.

At Oodrive right now, my year-one foundational work is exactly this: deploying SUS and CES first, dashboards next, a north-star metric for the design organization in flight. I'd rather have a small design team that measures than a large one that doesn't.

A note on AI

I keep my point of view on AI-in-the-design-pipeline updated separately, because the answer changes every six months and I want a single source-of-truth I can keep current. What I'm thinking about right now.

The short version: research synthesis (large productivity gain, used carefully); early-stage exploration (genuinely useful); final visual decisions (still human, and I'd argue should stay that way for the next 18 months). I'm running pilots at Oodrive in the research-synthesis layer first: lowest risk, highest gain. I got the IBM AI Fundamentals certification in late 2024 to make sure my mental model has ground under it, not just vibes.

What I won't do

• Manage by metrics alone. Metrics describe outcomes; they don't run teams. A leader who refers to a spreadsheet when asked how the team is doing is a leader who doesn't know how the team is doing.
• Promote into roles the company isn't willing to support. I'd rather hold a seat empty than over-promote and watch someone fail because the structural support wasn't there. The harm of an unsupported promotion lands on the person, not on the org chart.
• Ship a design system before the org is ready to operate it. Chameleon worked because it had a dedicated team, customers inside the company who wanted it, and the rituals that kept it from forking. A design system without those is shelfware in three quarters.
• Tolerate condescension toward Research, Content, or Accessibility. These are peer disciplines, not service functions. A design organization that treats them as downstream loses the strategic altitude its work depends on.
• Pretend I have all the answers. The next section is about what I'm still figuring out; that's not a soft close, it's load-bearing.

What I'm still working on

• Moving faster on AI tooling adoption. I'm careful by temperament and the field is moving faster than my caution. I'm working on shortening the gap between "I should look at this" and "we're running a pilot."
• Saying no earlier in cross-functional partnerships. I'm good at the long arc of trust-building and sometimes too tolerant of partners whose patterns aren't going to change. Saying no in week two saves the relationship; saying no in month nine just damages it.
• External presence. I've spent most of my career letting the work speak; I'm rebuilding the muscle of speaking about the work: writing, conference talks, mentoring formally. The case for being more public is one I've come around to slowly, and I'm still building velocity on it.

What you can ask my references

I keep a short list of people I'll happily connect you with: former direct reports, peers, and stakeholders across companies. Two are public on LinkedIn: Anna Mészáros (former direct report, now in design leadership) and Linda Kessler (research peer at GoTo, close to the team, reported to the same manager I did). Both endorsed me unprompted. Others I'll introduce on request, including senior IC designers I've grown into managers and Product / Engineering counterparts I've co-shipped with.

Specific things they can speak to:

• Whether the team felt steady through the pandemic remote transition (Linda's recommendation addresses this directly).
• Whether I can grow a mid-level IC into a discipline lead (Anna's progression to Accessibility Lead is the case in point).
• Whether I show up for people outside of the work itself.
• Whether my decision-making is consistent under pressure (former peers can speak to this, happy to introduce).