Context
I led the Global Experience team at OLX Brasil. Our mission was to look after the experience across devices and categories, and create guidelines so every vertical, from cars to real estate, felt like one product.
The main navigation had never been designed for how much the product had grown:
- A hamburger menu on web and Android, a tab bar on iOS, and different shortcuts in each header.
- The menu kept growing, and nobody owned it.
- Features for professional sellers and for OLX Pay sat next to everything else, with no distinction.
- Searching meant going back to the Home, and moving between categories was hard.
- Every new header version had to be updated by hand in every instance.
Designing and testing
With a group of designers, we built the ideal sitemap. Then I ran ideation workshops, mobile first, since a small screen was the hardest constraint for the new structure. Two rounds of critique with every team the navigation would touch brought in risks, data and ideas before we went to users.
We tested words and structure with card sorting, interviews and cloze tests. Then ten private and professional sellers and buyers went through a usability test of the concept.
From test to code
After refining the visuals with the design system team, we delivered in waves, starting with a responsive navigation on the web. New components went into the design system: bottom navigation, header 2.0, nav bar, a new account page and the ad card.
For the A/B test to be fair, the new navigation had to go live on every page at the same time. That's where it got hard. A corporate restructure dissolved our team, key stakeholders were laid off, and some teams weren't ready to update their pages. The test ran with experience flaws we couldn't fix.
What the test showed
Even so, the numbers moved. In the apps, every tab went up: clicks on Explore grew 430%, Account 51.9%, posting an ad 23.3% and Chat 17.8%. On the web Home, the new nav bar raised use of the global navigation by 23%.
One number needed care: posting an ad from the web Home dropped 25%, with no impact on ad posting across the whole site. It was the main metric to stabilize before rolling out.