E-commerce · varejo de moda
Um app que o time conseguia tocar
Uma marca de moda alugava um app white-label que não podia mudar. Construímos o dela, migramos a base inteira e depois passamos um ano no funil.
O texto abaixo ainda está em inglês. O resto desta página está traduzido.
- Período
2024—2026
- Escopo
Zero a um
Teste A/B
App mobile
CRO
Dados
Resultados
conversão do app, de menos de 1%
da receita do ano anterior inteiro, só no primeiro semestre de 2026
lifetime value do cliente
retenção
O que eu fiz
Fui dona do escopo da v1 e da migração de todo usuário para fora do app antigo: ninguém ficou para trás.
Refiz as páginas de coleção e de produto, a linha de frente do trabalho de retenção.
Conduzi o programa de CRO até o checkout, que era o gargalo real.
Levantei a prática de dados. O time tinha tracking de eventos em que ninguém confiava e que ninguém usava.
The problem
The brand had decades of equity in its stores, and an app it was renting. Every change went through the white-label vendor's roadmap, which meant the product could not answer to the business. A retention problem you cannot ship against is not a retention problem, it is a dependency.
What I found
The work
The migration was the part with no second chance: every existing user had to land in the new app without losing an account, an order history or a saved card. I owned that scope, and the definition of done was that nobody was left behind.
With v1 live, the work moved to the funnel. Collection and product pages first, because that is where retention is actually won or lost. Then the CRO program, which led to the checkout as the real bottleneck.
Underneath all of it, the data practice had to come back. The team had event tracking nobody trusted, which meant every argument was an opinion.
What changed
In the first half alone, the app did 107% of the previous full year's revenue. Customer lifetime value doubled, retention reached 42%, and new customer acquisition rose 50%.