Multiple Variants vs Multiple Products

Client:

Mercado Libre

Scope:

Product Design, Stakeholder Management, Research

Year:

2025

Visite Website

Summary

Iterating amidst an architecture shift at Mercado Libre by reframing user confusion as a mental model transition problem. I led research initiatives and designed conceptual design that stabilized adoption despite permanent system complexity.

Project

Context

Mercado Libre's legacy architecture for 10+ years was build under the "The parent product, child variations for one listing" paradigm. On 2022, the business pressure forced a shift: vendors needed independent pricing/shipping per variation to clear stranded stock. The solution at the time was to built a under the new model: "parent product, independent listings for its variations"

The problem inherited: Rollout was already underway, but adoption was collapsing. Sellers were confused, CX had an uptick of 47% on matters related to the project.

My contribution

As Product Design Lead, I was tasked with understanding why the new model was failing and provide solutions. I led the design and execution of research and user experience initiatives, which had these findings that I summarize in three core:

  • Challenge #1: Mental model collision: Sellers had both old and new listing types active simultaneously as a progressive rollout strategy.

  • Challenge #2: Cognitive mismatch:Some sellers couldn't grasp the new model because the new UI addressed the incumbent process of the two mental models by adding another layer of complexity.

  • Challenge #3: Findability breakdown Listings that were now "independent" were found elsewhere from sellers' familiar navigation, struggling to locate their own products.

My contribution

As Product Design Lead, I was tasked with understanding why the new model was failing and provide solutions. I led the design and execution of research and user experience initiatives, which had these findings that I summarize in three core:

  • Challenge #1: Mental model collision: Sellers had both old and new listing types active simultaneously as a progressive rollout strategy.

  • Challenge #2: Cognitive mismatch:Some sellers couldn't grasp the new model because the new UI addressed the incumbent process of the two mental models by adding another layer of complexity.

  • Challenge #3: Findability breakdown Listings that were now "independent" were found elsewhere from sellers' familiar navigation, struggling to locate their own products.

Design Approach

Rather than replacing the familiar parent-child model, I applied progressive enhancement strategy to preserve the visual structure sellers had used for a decade while adding per-variation controls. This approach addressed three critical failures of the previous rollout:

  • Mental model collision was solved through consistency of patterns of old workflows, continuing unchanged, while new capabilities layered on top.

  • Cognitive mismatch was addressed via progressive disclosure and clearer affordances. New features were hidden until needed, and visual cues dictated state of the variations.

  • Findability breakdown was addressed also by maintaining information architecture stability, listings remained in their expected locations, and controls appeared in-context and on demand.

The result: a solution that supported both old and new models simultaneously, designed for the transition state rather than an idealized end state."

Product Outcomes

Adoption of the solution approach reached 27% of eligible sellers by Summer 2025, up from 9% at the time of my intervention. Including Nike, on of the biggest sellers. The rollout stabilized enough to continue expansion to Markets and Categories

What remained broken

CX inquiries did not decrease. The two mental models coexist permanently, which creates ongoing confusion and cognitive load for sellers transitioning between them. Although big sellers benefited from the solution, Small and local shops where deeply affected