Enterprise Design · End-to-end Delivery

Big Data Consumer Profile · Huawei Cloud

From zero to launch — and back again with data.

Most projects end at handoff. This one didn't.

A big data analytics platform for Huawei's mobile operator clients — designed from scratch, shipped to production, and then validated with a full usability study including eye-tracking.

100+ hours of research. 17 users. 6 departments. 3 roles. One complete cycle.

Context

Version 2.0 had reached its limits.

Huawei's terminal mobile user base was growing fast — but the existing platform couldn't keep up. Limited user tags. No campaign workflow. No way to support the daily operational needs of the business teams depending on it.

The brief: rebuild the platform from scratch to serve data management, user insights, campaign execution, and performance analysis.

Three roles. Fundamentally different needs. One platform to serve all of them.

Three Roles, Three Realities

The platform had to work for users who didn't share the same goals — or even the same definition of "value."

CXO
Strategic overview

Needs a business overview and end-to-end flow of Huawei's global users. Expects richer predictive information to support strategy — not just reports.

Product Manager
User lifecycle management

Needs to manage the full user lifecycle through data. Expects the system to surface logical relationships and carry the user through end-to-end journey analysis.

Campaign Manager
Operations and execution

Needs campaign operations that are simple, efficient, and visible. Expects to understand churn, act on it, and track results — without friction.

Design

Three questions shaped the work:

  • How do you visualise multi-dimensional data without overwhelming the user?
  • How do you make self-serve analytics accessible to non-analysts?
  • How do you surface value the moment a user opens the platform?

The answers were different for each role. The interface had to hold all three.

Validation

After development and launch, the platform was opened to real users for structured usability testing.

The goal wasn't to confirm what worked. It was to evaluate in real usage scenarios — and identify opportunity points for the next iteration.

17 participants. 30-day heavy users. Quantitative and qualitative methods combined: questionnaires, interviews, observations, and eye-tracking.

Two findings stood out.

  • Visualised Overview — Eye-tracking showed clear, low-overlap scan paths, indicating the layout was successfully guiding users through information in the intended sequence. Users consistently fixated on colour-coded chart areas, confirming that visual hierarchy was influencing browsing order.
  • Multi-layer Filtering — The user group creation page showed clear F-pattern eye movement distribution, confirming the information layout was helping users query and filter in a natural, efficient way.

Impact

3
Roles Served

CXOs, product managers, and campaign managers — all served from one platform.

17
Users Tested

Real users. Real scenarios. 30-day heavy users recruited for the study.

100+
Research Hours

From discovery through post-launch validation. One complete end-to-end cycle.

✓
Information Hierarchy Validated

Eye-tracking confirmed scan paths followed intended layout logic.

"Most design decisions are validated by instinct.
This one was validated by data."