01 — My role
I joined Daki in July 2022, right after the company merged into a global group. The merger left two nearly identical apps evolving in parallel: the Brazilian one and the global one, with almost the same scope.
My role: lead the design of migrating Brazil onto the global structure without losing what held the local business together.
The context raised the bar fast.
In my second week, my manager left, and I took over the whole team.
Six people, with specialists in UX writing, service design, and design system, reporting directly to the CTO. Deadline stretched, pressure high.
02 — What was at stake
Keeping both platforms meant duplicating product, design, and engineering indefinitely.
Migrating meant risking Brazil’s conversion and user base, both built on a navigation model the global app simply didn’t have, starting with the categories section.
In an app-first operation, the app is the sales channel: degrading conversion meant degrading the revenue of the entire operation.
The decision to migrate came with the merger. My ownership was in the how.
03 — Product decision
I treated the migration as a product problem, not a reskin: conversion became the central KPI of the transition and, with the team, we mapped, area by area, what changed, what stayed, and which metrics could degrade.
I took that risk map to every business area, from operations to CX, marketing, and media, before they discovered the changes in practice.
The bridge between the global group’s product leadership and Brazil’s CX leadership was mine to hold: translating local business flows into roadmap priority. That bridge also unlocked revenue, because the global campaign tool arrived with more media spaces and assets adapted to Brazilian navigation.
Prioritizing also meant conceding. I accepted launching without filters and sorting in navigation (known gaps, with a workaround for the user) so I wouldn’t concede on the categories section, which held the Brazilian user’s mental model together.
There was one criterion: cut where there was a workaround, never where it broke the buying habit.
I didn’t have the product title yet, but I was already running the scope.
04 — Leadership and craft decision
I led in both modes: manager of a team of six, and hands-on in the most critical problem.
I organized the team by risk fronts, each specialist owning her own, with critique on a fixed cadence with me at the decision points: the bar was explicit without me being inside every deliverable.
The visual alignment to the brand guide and the design tokens, for instance, ran with the design system specialist, who prepared the system for multiple brands and themes while I held the most critical front.
The comparative study between the two apps and the heuristics analysis stayed with me, and they guided the navigation restructure: bringing the new app closer to the Brazilian user’s mental model and rebuilding the categories section that didn’t exist.
To validate, we prioritized the flows with the highest conversion risk and ran a moderated, in-person perception test with heavy users, on the beta app with real content, measuring perceived effort (CES) task by task. The test confirmed the new navigation, knocked down points we’d assumed were solved (adding promoted products, which almost no one completed) and turned into a backlog triaged into five fronts: ready to build, redesign, content configuration, monitoring, and prioritized.
And the period also grew the team: one of the product designers, central to the initiative from start to finish, was promoted from mid-level to senior in that cycle.
05 — Relative result
The migration happened with no meaningful CX noise: contact rate held flat through the entire rollout, which was exactly the guardrail we’d defined. The rollout advanced by percentage of the base, with the guardrails as the criterion to move at each step.
relative, in session → order conversion
two months after rollout to the full base
measured per session, by method choice
The gain came from the migration as a whole, not from one screen: the entire new structure converting better than the old one.
At the bottom of the funnel, cart-to-checkout nearly doubled.
And the operation gained new media-revenue capacity, with more campaign spaces adapted to the Brazilian context.
06 — What I’d do differently
I’d have pushed for direct, franker alignments with the global group’s product leadership from the start. Without those exchanges, deadline pressure turned into pressure to deprioritize features and interactions, while our bar was not to degrade the experience the Brazilian user already had. We worked it out, but managing noise that an expectations alignment would have avoided.