ZAP+/ OLX Brasil

Listing Quality Score

Painel de anúncios do Canal Pro

Overview

Zap+ operated with 45,582 active advertisers and more than 7 million active listings on the platform. Within this ecosystem, the sponsored listings product (Super Destaque/Destaque, i.e. featured/highlighted listings) accounted for 19% of company revenue, the second largest revenue stream in the business.

The problem: the company started seeing a consistent increase in support tickets and complaints across three fronts:

  • Low visibility of sponsored listings
  • Lack of advertiser trust in the featured system
  • Poor lead quality

This combination led to a drop in lead conversion and began directly impacting revenue, putting the platform's second largest revenue line at risk.

Details

Role: Staff Product Designer
Focus: Discovery, Facilitation, UI/UX, Metrics

Lista de anúncios antes da mudança Formulário de criação de anúncio antes da mudança
Before

Outcome

Before proposing any solution, the goal was clear: increase the performance of featured listings, restore their perceived value, and improve the advertiser experience on the platform.

This meant solving not just a product problem, but a trust problem: advertisers needed to believe again that paying for a featured listing actually meant more visibility and higher-quality leads.

Discovery & Problem Understanding

The investigation started with an analysis of the quality score of sponsored listings, and revealed structural root causes:

Average score of 5.8 among sponsored listings, direct evidence that low-quality listings were being promoted.
The score only measured field completion, not the actual quality of the information (e.g., a listing could score high with poor photos or a generic description).
Featured status was being assigned without clear rules of eligibility.
Sponsored listings had been used internally as a sales and retention incentive, which reinforced low-quality listings entering the premium product.

This diagnosis made clear that the problem wasn't technical visibility — it was quality at the source. The system was promoting the wrong listings.

Ideation

To unlock hypotheses and solutions, I facilitated an ideation workshop with the teams involved, followed by technical feasibility alignment with engineering on how to raise information quality at the source. Two technical directions emerged as viable:

  • DataZAP for price comparison and address/building precision validation via geolocation.
  • Machine Learning for automated image quality analysis on listing photos.

From this process, three core hypotheses guided the solution design:

HypothesisExpected outcome
Qualitative score (not just completion)
→
Higher reliability and engagement
Minimum score required for featured eligibility
→
More relevant and effective featured listings
Data GranularitySingle, aggregated metric
→
Advertisers more actively engaged in improving their listings
Mapa mental do processo de ideação

Solution

The solution was designed across three touchpoints in the advertiser journey:

Listing creation page

  • New Listing Quality Card
  • Replaced the speedometer gauge with a progress bar, clearer and more actionable
  • Tags per information section (e.g., photos, address, description), which activate when there's improvement and disappear if the improvement isn't sustained
  • Targeted tips for improvement tied to each tag
Novo card de qualidade do anúncio em uso

Solution

Listings list

  • Added the quality score directly to the listing card
  • Disabled featured activation for listings scoring below 8.5 — the missing eligibility rule
Lista de anúncios com selos de qualidade

Solution

Home Canal Pro (Listings section)

  • Two new tracking charts: Deals in progress and Listing quality
  • A suggested improvements section with cards, giving advertisers an overview and clear next actions for their listings
Dashboard de desempenho do Canal Pro

Impact

Shipping the product was only half the work, the new rules (such as blocking featured status below 8.5) directly affected the commercial relationship with real estate agencies and brokers, so close collaboration with the sales team was needed to align messaging and bring the base along through the transition: training the sales team on the new quality logic, and direct communication with agencies and brokers about how and why the rules were changing.

With this change management effort running in parallel to the technical delivery, 3 months after implementation:

6.8
Average listing quality rose
from 5.8 to 6.8 (+17%)
60%
reduction in low-quality
featured listings
35%
reduction in support
tickets related to visibility
15%
Sponsored listing revenue grew
from R$184,823 to R$212,546
← Anterior: Caju Growth Design Próximo: Diin — Objetivos →