TL;DR
The Problem
Four regions built four homepages with no shared rules, and the metric leadership watched said everything was fine.
The Turn
The first system we designed made the problem worse, and understanding why produced the one that worked.
The Solution
A fixed page spine with four content modes. Regions choose the moment and the content, not the layout.
My Role
Co-led the project.
Owned research programme across five markets, section-level analytics read that reframed the problem, the failure analysis on the first system, and the template and IA rules that shipped.
Timeline
Jan '25 - Oct '25
The Problem
The number everyone trusted was measuring the wrong thing.
6%
EU hero banner interactions
7%
NA hero banner interaction
65%
EU users interacting with search
What these conflicting numbers told me
The dashboards were healthy and the page was failing, and both of those things were true at once. That gap is what the whole project came out of.
The fragmentation was not subtle. Every market had drifted somewhere different, and each one had drifted for its own reason.
Engagement rate had been the only lens anyone was looking through. Underneath it, three separate failures were running at once, and each one needed a different fix.
My Read
What struck me was not that the regions looked different. It was that every team had a defensible reason for every decision they had made.
Brazil's page was aggressively transactional because Brazil converts on the homepage.
The UK's was restrained because the UK does not.
Fragmentation was not negligence, it was what happens when capable people work without a shared system, and that ruled out a visual refresh as the answer.
The research gave us problems.
Design principles gave us direction.
Three tensions kept surfacing across every market, every interview, every test. Instead of resolving them as trade-offs, we turned them into design principles.
Global Unity
Premium Presence
Regional Adaptability
My Read
The insight wasn't that regional teams needed fewer choices. Choice without capability isn't flexibility. It's just a more complicated way to go wrong.
The Solution
Regions now choose the moment,
not the layout.
The rules I wrote
In the final phase, the focus shifted from concepts to concrete rules. The decisions below are ones I directly shaped to make the homepage safer to use at scale, not just nicer to look at.
Learnings
What this project changed for me
I researched the wrong audience first.
The homepage has users and it has builders, and I spent four studies on the first before speaking properly to the second. The Builder Kit failed for reasons a week with regional owners would have caught.
The insight was in data we already had.
The section-level analytics had been sitting in the same dashboard as the engagement rate everyone was reading instead. I now ask what we are already ignoring before commissioning anything.
Still open
Whether four modes is the right number. Four is what regional teams could hold in their heads, not what came out of analysis, and I do not know where a fifth stops being useful and starts eroding the discipline that made the system work.
Where it stands today
The system is live and other teams build against it now.
The templates and the IA rules are in use across regional teams today, and BTIT is building the component variants against them. What started as a homepage redesign became the structural default that new market pages inherit, which is the part I am most pleased with and the part that took longest to argue for.