Youjin's Portfolio

Youjin Lee

Shaping pricing decisions

Applying information architecture to reduce cognitive load and support confident decisions

Image of a multi-step game that draws cards that users can choose. Named Machines and Margins.

Overview

My client was a national retail chain seeking an application to support their pricing and sales-forecasting workflows. I was one of two designers on the product team, where I led design strategy and partnered closely with the product manager and tech lead.

We collaborated directly with internal employees who relied on Excel spreadsheets and manual processes to complete their work. Our goal was to streamline these workflows, reducing time spent on repetitive tasks so users could focus on higher-value strategic work.

INDUSTRY

Grocery retail

ROLE

Product designer

SKILLS

UX Research
Information Architecture

YEAR

2023

Long-haired person looking at their laptop thoughtfully and different types of datagraphs displayed next to them

Initial research

To kickoff our research efforts, we conducted high-level walkthroughs with pricing analysts across multiple departments to better understand their current experience. Despite being in different departments, the users all had the same goal:

USER GOAL

Confidently make data-driven pricing decisions to support their department's financial objectives

Problem framing

The main challenge was that the legacy application gave the analyst a wealth of data, but consuming it was overwhelming. We also realized that there were small but significant process differences across departments.

We identified the following problem statement:

PROBLEM STATEMENT

How might we create a flexible experience for users across multiple departments that minimizes cognitive load and speeds up their current data review process?

Based on the insights from the research, we were able to outline our design goals grounded in the problem statement.

Flexible view of data

Intuitively organized data

Reduced time spent

Information architecture

I was confident that focusing on information architecture would shed valuable light into how we solution for the problems identified and meet our project goals. The biggest question we wanted to answer were as follows:

RESEARCH QUESTION

How do each of the datapoints inform and influence the decisions made by analysts?

We conducted two additional rounds of research to collect insights and gain a deep understanding of the initial question we set out to answer. Our learnings were as follows:

ROUND 1 of RESearch

Card Sort

OUTCOMES

Identified data relationships and users' mental construct of data

Identified data relationships and users' mental construct of data

Identified data relationships and users' mental construct of data

Gained context around why and how data was reviewed

Gained context around why and how data was reviewed

ROUND 2 of RESearch

Focus Group

OUTCOMES

Prioritized data by influence on decisions and frequency of use

Prioritized data by influence on decisions and frequency of use

Prioritized data by influence on decisions and frequency of use

Identified differences across departments as well as mutual needs

Identified differences across departments as well as mutual needs

We used our learnings to create an information hierarchy that would inform the ease of access to the information or how much visual weight it would have.

HIERARCHY PRINCIPLES

Surface critical information at a glance for fast, confident decisions

De-emphasize secondary data by increasing access effort to preserve focus

Using these principles, we further broke down the datapoints we had gathered and mapped them to better understand how we could structure the pages efficiently.

Product identifiers and success metrics were crucial and a typical starting point for an analyst's task so they had the highest priority in terms of access and visual weight.

Even data points that were "never" referenced had to be accounted for in case it was needed in the future or by a user with unique needs. However, it would require multiple clicks to access.

The pricing data was a bit more complex. The decisions made about the organization of this data were informed by a mix of how often it was needed, its need across user types, and how crucial it was to decisions.

Wireframes

Using the outputs of our research, we ran multiple rounds of user testing. From our lo-fidelity mock ups, we learned that seeing data in tables was the most familiar and comfortable view, allowing for quick scanning. With our clickable prototype, we received generally positive feedback and reception to how we grouped and displayed data across all user types.

With our findings in mind, we put together final wireframes - each piece of the screen carefully informed by each of our insights.

A laptop screen that shows a header with information about the product, a section for high-level financial metrics, and lastly, a table with tabs that is filled with data. There are arrows pointing to each of the areas that represent the functional organization, flexible view of data, hierarchy of information, and glanceable review

Identifiers

The first step to any data analysis was correctly identifying the product and reviewing context.

Projections and objectives

Since the analysts' main goal was to ensure they were meeting department goals, we made sure these were easy to reference with dynamic feedback on progress.

Flexible views

To solve for the different departments and their unique needs, one of the interactions we added was a way to easily select how the data was organized, with potential to add additional views in the future.

Column groups

Since the product pages had massive amounts of data, we added column groups to make it even easier to scan and sort through data. Users could expand and collapse groupings to decrease their cognitive load and focus on only the data they wanted.

References