How Data‑Driven Design Boosts Click‑Through Rates in a Recruitment App

Data‑Driven Operational Design Overview

Operational designers must combine creativity with data thinking to maximize design value, accelerate commercial impact, and achieve high‑quality, high‑quantity, and high‑conversion outcomes.

Case Study: 58APP Homepage Optimization

To improve homepage traffic and recruitment conversion, the 58UXD design team collaborated with product and platform teams to run precise data‑driven experiments on the 58APP homepage. The first phase focused on feed‑flow cards and floating windows, producing 36 assets and 108 data groups.

Testing Process (Five Steps)

Identify Test Resources – Map all entry points and slots on the homepage, category pages, and third‑level pages, then prioritize slots based on traffic, design effort, and business needs.

Select Test Method – Choose A/B testing after evaluating AB, cross‑test, pre‑post, and gray‑scale methods, aligning the method with user attributes, goals, and cycle.

Decompose Test Layers – Build a mind‑map of six dimensions (state, layout, background, main, copy, guidance). Each dimension is further broken down, e.g., the main layer includes realism vs. 3D, scene type, number of people, dynamics, profession, and decorative elements.

Quantify Test Factors – Ensure each test material varies only one factor. Product‑level factors involve user segmentation; design‑level factors involve variations in copy, button, background, etc.

Analyze Experiment Data – Clean raw data, remove irrelevant or duplicate entries, and create new dimensions for refined analysis, focusing on business, product, and user goals.

Key Experiments and Findings

Test 1: Real‑person vs. 3D‑person images – 3D images achieved higher click‑through rates.

Test 2: 3D‑person static vs. dynamic – No significant difference; both performed similarly.

Test 3: Copy‑focused vs. image‑focused floating windows – Image‑focused designs with profession‑specific visuals yielded higher clicks.

Test 4: 3D‑person with generic vs. profession‑specific copy – Profession‑specific copy increased click‑through rates.

Test 5: Pure copy variants – Phrases like “Nearby Hot Jobs” and “High Salary Urgent Hire” performed best across professions.

Conclusions

Data‑driven design experiments revealed that 3D professional personas and profession‑specific copy significantly improve click‑through rates, guiding future UI material creation and supporting business decisions. The successful first phase has already informed a second, more refined testing round, further enhancing traffic distribution and conversion during peak periods.