How Eye‑Tracking and Heatmaps Reveal What Customer Service Agents Really Need

Introduction

“善借于物” means making good use of external tools to achieve goals. The article applies three tools—Hawkeye, Weheatmap, and iFeedback—to the design of a mini‑program customer‑service console, discussing what agents see, do, and receive.

Hawkeye and “What agents see”

Hawkeye is an iPhone/iPad app that records eye‑movement trajectories using the device’s depth camera, offering a low‑cost, mobile alternative to traditional eye‑trackers. In a study with nine agents, heat‑maps showed that content and timestamps attracted the most attention, leading to a redesign of the message list.

Because lightweight tools sacrifice some precision, results should be treated as indicative rather than definitive.

Weheatmap and “What agents do”

Weheatmap is a self‑built web‑behavior research platform that generates dynamic heatmaps from five event types: click, area, scroll, keyboard, and record. It differs from tracking scripts like Hotjar by visualizing supplied data only.

Analysis of click data revealed that agents frequently use right‑click menus for copying messages and that keyboard shortcuts improve efficiency. Based on these insights, a global right‑click menu and additional shortcuts were added to the service console.

Keyboard event data also guided the implementation of shortcuts for navigating to unanswered conversations and filtering messages.

iFeedback and “What agents receive”

iFeedback is an internal clustering tool that preprocesses text, performs word segmentation, clusters similar feedback, and conducts association analysis. Applied to three months of merchant inquiries, it surfaced high‑frequency terms such as “batch” and “refund”, identifying impulsive purchases in live‑stream sales and failed community group‑buying as key issues.

The clustering results helped prioritize problem‑solving and demonstrated how automated analysis can replace manual, inefficient filtering.

Conclusion

By borrowing the right tools, designers can better understand user behavior, optimize interactions, and enhance experience. Tools like Hawkeye, Weheatmap, and iFeedback illustrate how lightweight research instruments can inform product decisions without the overhead of full‑scale labs.