How 58租房 Revamped Its Roommate Matching: A Full‑Cycle Design Case Study

Overview

To help co‑rental users find satisfactory listings, 58租房 offers multiple search methods and rich property information. In co‑rental scenarios, the roommate significantly impacts the living experience; thorough pre‑move‑in roommate understanding reduces conflict and raises satisfaction. As user numbers grow, the roommate‑finding process must become more efficient and satisfying. This article reviews the latest roommate‑finding redesign, from identifying focus areas to establishing evaluation metrics, detailing design decisions, and validating outcomes.

Finding Redesign Breakthrough Points

Through user interviews and questionnaires, dozens of roommate‑related pain points were collected and grouped into three main categories (see Figure 1). Users fall into two groups: those without a house who need both a house and a roommate, and those with a house seeking a roommate. The old roommate information model lacked sufficient detail for both groups, prompting a richer model.

The new model displays basic personal information (gender, age, occupation, hobbies) and roommate requirements (gender, occupation, schedule, pet and smoking restrictions). Additionally, users without a house specify housing preferences (location, budget, move‑in date, amenities), while house owners provide property details (photos, location, rent, layout, current occupancy, amenities). Centralized data processing enables precise matching and recommendation, improving decision efficiency.

Design Empowering Business

Rich user data is essential for accurate content matching. Design must balance data collection, operational load reduction, and search efficiency.

Data Collection

To reduce user anxiety during form filling, the long form was split into short, prioritized steps: personal information, roommate requirements, and housing information (or house‑search requirements for users without a house). Each step presents a short form with a progress indicator, easing user expectations.

Guiding Personal‑Info Completion

Since few users proactively publish roommate posts, prompts were added throughout the browsing and usage flow: informational cards in the list feed, content‑view limits, filter‑panel entry points, and personal‑completion cards on management pages (see Figure 3).

Improving Find‑Person/House Efficiency

Four dimensions were addressed:

Filter Upgrade: Separate filters for users with and without a house, covering more criteria.

Information Structure: List pages now prioritize personal details, show three images for quick first impressions, and expose a chat button for fast connection (see Figure 4).

Common Albums: Group users into albums such as colleagues, hometown peers, alumni to foster resonance (see Figure 5).

Efficient Connection: Highlight online posters, bubble prompts for important messages, and toast notifications for new matching posts.

Operation Reduction

Basic experience enhancements include:

Collapsible sections in forms (e.g., employment details appear only when “employed” is selected).

Visual metaphors that replace dense information with intuitive, ritualistic designs.

Widget optimization: a unified option panel opens once to fill all selections, reducing repetitive clicks (see Figure 6).

Validation Metrics and Assessment

Design evolved from User Experience Design (UED) to User Growth Design (UGD), requiring designers to align with business growth early. Designers participated in setting both conversion and usability metrics, ensuring that design decisions directly support business objectives.

Continuous Optimization

The redesign yielded improvements but also introduced new challenges. A cross‑functional user‑research group conducts eye‑tracking tests, industry research, competitive analysis, brainstorming, and closed‑loop user interviews to iteratively refine the experience.