SmartIntellect

Redesigning a CDP for cross-border merchants

B2B

SaaS Product Design

Industry

Cross-Border E-commerce

Company

SHOPLINE in JOYY Inc

Timeline

2024.06 - 2024.09 (3 months)

My Role

Product Designer

Tools

Figma, Jira, Notion, Zoom

Business Impact

40% Paid Conversion Rate

SmartIntellect is a customer data platform for merchants on SHOPLINE. A merchant’s customer data lives in three or four places at once — their storefront, Google, social channels — and none of them talk to each other. SmartIntellect pulls those sources into one profile per customer, so merchants can see behavior, purchase history, and ad performance in a single view, then act on it with targeted campaigns.

Overview

Impact

Archetype Mapping

Scenario Analysis I

Problem I

Solution I

Scenario Analysis II

Problem II

Solution II

Design System

Overview

The Problem

The product launched its MVP, but people weren’t willing to pay for it

SmartIntellect was built to help merchants drive sales and conversions more efficiently. It launched its MVP and passed basic market validation. But even though most merchants saw its value, very few were actually willing to pay for it.

Why it happened

There are three factors that design can help improve

In research, most merchants felt the product was difficult to use, missing essential features, and visually inconsistent, leading to high learning costs and low operational efficiency.

Key factors merchants consider when subscribing to the CDP product

What I did

Streamlined workflows, added new features, and rebuilt the design system

After identifying these needs, I conducted competitive analysis and mapped design strategies.

I streamlined workflows and prompts to lower the learning curve, introduced new features such as a dashboard to expand product value, and rebuilt the design system with guidelines and reusable components.

After identifying these needs, I conducted competitive analysis and mapped design strategies.

Impact

The redesigned version lifted merchant paid conversion by 40%

Merchant satisfaction reached 82% for usability and 79% for completeness, while tasks such as creating segments and comparing data were completed in nearly half the time.

The redesign drove a 40% lift in merchant paid conversion, with 7 of 9 trial participants converting to paid users.

12%

Feature Satisfaction

(Design Value)

48%

Task Completion Time

(Design Value)

40%

Paid Conversion Rate

(Business Impact)

Archetype Mapping

Business Operators and Data Analysts

To better understand how users work with our product, I identified Business Operators and Data Analysts as the core target users, and focused my research on their key tasks, needs, and pain points.

Target user 01

Business Operator

Needs fast execution, clear steps, and immediate feedback.

User Task

Connection

Decision-Making

User Need

Customer data management

Customer segmentation

Target user 02

Data Analyst

Needs accuracy, flexible customization, and real-time data.

User Task

Analysis

Insights

User Need

Omni-channel analytics

Custom configuration

Core target users

By analyzing the tasks and needs of these two key roles across different scenarios and integrating previous research findings, I identified two key scenarios: Customer Management and View Analytics Reports.

Scenario Analysis I

In the Customer Management scenario

Users require smooth, efficient flows and clear feedback

In the Customer Management scenario, the business operator is the primary user. They focus on connecting customer data to SmartIntellect, as well as viewing, managing, and editing that data.

User journey map: Customer Management

Problem I

The current segmentation workflow is too time-consuming for operators

Although the current process allows them to fulfill the requirement of creating new customer segments, it is widely regarded as overly cumbersome. Operators often need to repeatedly make selections, leading to a prolonged process.

Current segmentation workflow

The current rule layout creates high cognitive load and slows down review for operators

All rules are listed in creation order, regardless of type. When editing a segment, operators must review each previously created rule to check for updates or missing logic. Over time, older rules become harder to locate and understand, increasing cognitive load and slowing down the workflow.

Current page display

As a result, users struggle to quickly locate, compare, or edit specific rules. The lack of visual hierarchy and grouping not only impacts readability, but also increases the time needed to complete segmentation tasks.

Solution I

Simplified segmentation flow for better clarity and usability

I redesigned the segmentation workflow to reduce friction and improve clarity—removing the rule selection step, introducing a vertical layout with grouped conditions, and applying conditional displays to streamline the interface. The result: a simpler, more readable process that helps operators complete tasks faster.

The redesigned rule builder, with the Segment Rule selection box removed and related conditions grouped under a single vertical connector.

SmartIntellect's original customer segmentation rule builder, with each condition in its own boxed panel and a separate Segment Rule selection box.

Before

After

/ 01

Fewer steps help operators complete tasks faster

Eliminated the Segment Rule selection box and introduced a vertical layout that displays rules directly on the page. Operators configure rules inline by clicking “Add this sort.”

/ 02

Logical grouping makes long rule sets readable

Grouped rules in the same category under a single “and,” using nested cards to keep the hierarchy readable.

/ 03

Progressive disclosure removes friction from event setup

Applied progressive disclosure so “Add property“ appears only when the selected event supports properties and has slots available, cutting unnecessary steps.

Disabled “Add property” and surfaced a tooltip on hover once a rule reaches its 10-property limit.

Scenario Analysis II

In the View Analytics Reports scenario

Users require flexibility, clarity, and efficient cross-view analysis

In the View Analytics Reports scenario, the data analyst takes the lead. They are responsible for creating various reports based on business needs and extracting key data from these reports to support business decisions.

User journey map: View Analytics Reports

Problem II

Data analysts have no centralized workspace for comparing reports

Although data analysts can conduct analysis using the current system, data is scattered across multiple reports and real-time feedback is lacking—making the process inefficient and delaying decision-making.

Every time I want to compare charts, I have to click back and forth across multiple pages.

Solution II

Designing a dashboard MVP for data analysts to address their most pressing pain points

Based on user needs and competitive insights, I led the design of the Dashboard MVP to help data analysts organize and analyze key charts more clearly and flexibly.

Under tight timelines and limited resources, I made strategic design trade-offs—removing complexity and prioritizing the most critical pain points. The result was a scalable and efficient workspace that addressed user needs while laying a foundation for future system growth.