Helping merchants handle exception orders without leaving their primary order-processing workflow
This project was completed during my internship at ByteDance, within the Douyin E-commerce team (TikTok Shop China), focusing on the merchant-facing platform: Doudian
I integrated Order Tools into Order Management, enabling merchants to handle exception orders without leaving their primary workflow. To accommodate the integration, I also restructured the filtering area to improve screen efficiency and optimize the use of vertical space
Where merchants get their work done
Douyin E-commerce merchants process thousands of orders every day in Order Management. But when exceptions arise, such as address changes or shipping negotiations, they rely on a separate module, Order Tools, to handle these cases
Out of context, out of use
Order Tools lived outside merchants' primary workflow, forcing them to constantly switch context whenever they needed to take action. Instead of supporting high-frequency operational tasks, the fragmented experience interrupted merchants' workflow and made critical tools difficult to discover and use, ultimately resulting in low adoption of Order Tools
A system–merchant mismatch
Feature-driven
System FlowThe system is structured around features and modules. Order Management and Order Tools are treated as two separate destinations, each with its own entry point
Action-driven
Merchant LogicMerchants, however, don't think in features. They focus on what needs attention right now and what action to take next
How did I integrate order tools into the workflow?
How did I make room for integration?
Why didn't we move forward with them?
V1: A button that opens Order Tools with a side panel
Improved access and order management context, but did not change how users decide when and why to take action
V2: An overview showing all pending actions
Data points aggregation caused information overload, making it harder to guide action as the system scaled
What I've learned
- Products constantly evolve through the addition, removal, and adjustment of features, and being able to anticipate and accommodate this change is central to building scalable products
- In a mature ecosystem, users rely on muscle memory. Design interventions should focus on seamless integration rather than radical change
