How Quantitative Design Verification Transformed SuNing’s “慧眼” Data Portal

1. Origin

Every project follows a cycle of requirement‑design‑development‑launch‑re‑requirement‑re‑design, relying heavily on the experience of product, design, development, and testing teams. Any flaw can cause serious issues, and adding design verification is an effective way for designers to optimize this process.

2. Huiyan Project UI Evolution

2.1 Project Overview and User Needs

SuNing had multiple data‑analysis products that required users to open many pages. Huiyan was created as a unified data portal to integrate all analysis tools and improve business users’ efficiency.

2.2 Interface Evolution

Initially, the data board occupied only 46.3% of the page. To enlarge the board, two redesign schemes were proposed.

Scheme 1 hides secondary navigation and collapses the report list, increasing board coverage to 58.4% and 69.4%.

Scheme 2 restructures navigation and prioritizes report areas, using a floating window for secondary items and raising board coverage to 90.9%.

2.3 Potential Issues

While Scheme 2 meets the main requirement, hiding the report list may cause users to wonder how to open new reports, potentially affecting task completion efficiency. These concerns can be measured through quantitative design verification.

3. Application of Design Verification

3.1 User‑Experience Goals

Huiyan aims to be SuNing’s sole internal data portal, enhancing daily data‑analysis efficiency. The explicit goal of the redesign is to enlarge the report display area, while the implicit goal is to preserve the experience of other sections.

3.2 Deriving Metrics with the GSM Model

The GSM model (Goal‑Signal‑Metric) translates abstract goals into concrete metrics. For Huiyan, goals are mapped to user behaviors (signals) such as task success, task time, errors, efficiency, and learnability, which become measurable metrics.

3.3 Using Metrics

For the task “open a new report,” signals like time spent on each step are recorded. Short dwell times may indicate an inaccurate default report, while long dwell times suggest difficulty locating the desired report.

3.4 Trustworthiness of Metric Data

Outliers caused by interruptions can skew results; applying confidence intervals helps mitigate bias.

4. Summary

The article demonstrates how design verification, supported by quantitative analysis, provides a rigorous framework for evaluating and improving UI redesigns, making design decisions more persuasive and data‑driven.