Inven
Tory
AI
Analytics
Business Management
SaaS
1. Overview
2. Research
3. Ideation
4. Designing
5. Delivery

Product Development

1.1

Develop a tool that streamlines and optimizes business management operations.

My Role
Overview
InvenTory is a service-as-a-solution (SaaS) designed to support small and large businesses with accurately tracking and managing their operations and making informed business decisions. Users can integrate and consolidate data, indicate and track operations' metrics and key performance indicators (KPIs), visualize and analyze trends, and streamline processes through automating workflows.
The Goal 🎯
We noticed that businesses lose revenue due to inaccuracies with inventory management and sales forecasting. The goal of this project is to discover and develop a solution based on user research to help businesses manage assets and make thoroughly-informed business decisions.
DURATION
March 2024 - April 2024
(3 weeks)
TOOLS
Figma, Zoom, Slack, Google Suite
TEAM
Suzanne Nguyen
Payam Olfat (UXD) 
Austin Aguilar (UXR)

Background Research

2.1
Demand Forecast and Data Errors
The team wanted to identify and understand the most prominent and troublesome problems faced by entrepreneurs, small and large businesses, e-commerce businesses, etc. After some research, we discovered that demand variability is one of the biggest challenges for businesses.

A recent evaluation shows that businesses have a 20-50% forecasting error rate, ultimately resulting in revenue loss. Another study showed that 69% of retailers and 66% of consumer product companies struggled with accurate supply forecasting due to lack of accurate and updated information on fluctuating customer demand. Errors in data, data management, and/or modeling all lead to increased forecast errors, further causing under-stock and loss in clientele and revenue, or over-stock and money getting locked in excess inventory.

A survey conducted on 200 businesses showed that 72% of the businesses' supply chain were negatively impacted by COVID-19. Despite being a unique global challenge, the pandemic highlighted issues that have existed prior in data analysis and demand forecasting.  

Reducing business prediction error and revenue losses requires accurate data management and modeling and a comprehensive evaluation of human needs and behavior, social/cultural changes, and local and/or global conditions. However, these variables don't always accurately or immediately reflect in quantitative data.

Our goal was to discover how to develop a data management and analysis product that would best support small businesses, and would be a strong foundation for supporting large businesses at greater scale. We needed to develop a product that accounts for variability in human, social, and market behavior to efficiently and accurately manage data and generate forecasts.

Generative Research

2.2
We conducted five remote, moderated, semi-structured interviews via Zoom. We sent out a survey, which interviewees were required to participate in prior to their interview.

The user group of interest were 18-60 year old individuals who have educational exposure to or working experience with data management systems.

I organized the survey questions and interview script around three main objectives: 

   ⊹ Strengths & Weaknesses: Understanding user needs and pain points with data management and accuracy.
   ⊹ Opportunities: Understanding users' decision-making process when managing and analyzing data.
   ⊹ Adaptability: Discovering what can provide users a more smooth and valuable transition into data management systems.

The survey also included multi-select questions asking users how they learned to use data management systems and their preferred method of data management and analysis. Users rated the difficulty of learning and using data management systems on a Likert scale, providing us with some quantitative data.
Data management systems are used differently across many scenarios, career roles, and industries. It was vital to prepare for the whole spectrum of backgrounds and experiences participants may have to ensure interviews were efficient and productive. I looked into the roles and responsibilities of several roles that popularly have general experience with data management systems. This information helped us redirect the interview on-the-fly, strategically pivoting when necessary to discover unique, valuable insights from users.

Interviewers were advised to take a semi-structured approach to discover varying and nuanced complexities that users experience with data management systems. I hosted a two-hour Interview Workshop with the team to discuss what we should expect, and to support and prepare the team for conducting interviews. I reviewed and expanded on all research objectives and questions to communicate their core purpose. I also shared several examples for how each question could be tailored to various scenarios, careers, and industries while still maintaining its core objective purpose.  

After each conducting interviews, I followed up with the team regarding their interview sessions. As the Lead UX Researcher, I enjoy being able to hear about others' experiences especially if I couldn't quietly accompany every interview. I appreciate this process as it opens a space for learning and constructive feedback. I was able to learn about unique ways teammates found an opportunity to maneuver around the interview script, and ask valuable, insightful questions. It also helps me reflect on how I could improve the research objectives, interview questions, or interview workshop. I also enjoy being as involved as I can, where and when I can. By having follow-up interview debriefs, I supported my teammates in flushing out their interview notes, and clarified questions I had about their interview notes. This ensured that I had and understood all the notes and information correctly before synthesizing data and extracting insights.

Once all the data was collected and cleaned, I organized the data in affinity diagrams to identify common themes and categorize user needs and pain points. I used empathy maps as a deep-diving tool to uncover and organize valuable nuances that can be turned into actionable insights.
We discovered several common themes among interview insights. The affinity diagram (Figure 1) on the right is a comprehensive chart on all the themes discovered. The empathy map (Figure 2) reveals deeper insights around user needs and pain points.

The largest takeaways are that there are various tedious tasks preceding data analysis and modeling. Having educational or professional exposure to data management and modeling is relatively advantageous in the early stages of a career path, but the advantages diminish overtime as new products and processes inevitably need to be learned. Users are challenged with constantly being creative problem-solvers, updated with computer skills, and resourceful to tackle the steps preceding data analysis and business forecasting. The later half of their battle is data management and making sure their models will be accurate.

Below are the general qualitative and quantitative findings that contributed to the conclusive research findings.

Interviews Insights
   ⊹ Business owners struggle with disrupted workflows due to errors in data accuracy.
   ⊹ Businesses struggle with demanding manual labor of inputting and tracking data.
   ⊹ Users feel many data management systems lack streamlined processes for business-specific data analysis.
   ⊹ Despite learning how to use their company's data management system, many individuals still fall back on spreadsheets out of familiarity and reliability or to compare and check systems' accuracy.

Survey Insights
   ⊹ 63% of participants felt that learning about a specific industry was the most challenging part of business decisions.
   ⊹ 56% of participants found learning about their industry challenging.
   ⊹ 44% of participants preferred data management programs.
   ⊹ 44% of participants preferred spreadsheets.

Competitor Research

2.3
We conducted market research of data management products to understand what solutions other have explored. We documented and organized their strengths and weaknesses, and used those findings to understand opportunities within our product.

There are a lot of business management products and services that already exist for data management and analysis. These products and services focus on providing users with a singular platform to manage and execute business operations such as data management, data analysis, data visualization, streamline development, etc. Some products offer content management services (CMS) enriched by customer relations management (CRM) tools and provide users with abilities to organize customer data, discover insights, and create business content.

However, most of these products don't provide users with accurate and efficient system for streamlining processes. Users still face the tedious task of manual data cleaning, reformatting, and/or re-entry before analysis.

The table below lay outs competitor's features and strengths.  

User Persona

2.4
Manager Michael
Michael represented 18-60 year old individuals that own or manage a business.

   ⊹ Work responsibilities include data entry, data auditing, data analysis, analytical reporting, and making business decisions.
   ⊹ Time consumed by conducting market research and manually generating models to compare business-specific operational efficiencies.
   ⊹ Often occupied with correcting data inaccuracies.
   ⊹ Stressed about losses due to model errors and unaccounted for market conditions.

Journey Map

2.5
The journey map represents the stages of individuals' experiences, and the opportunities we could provide.

The large issues for businesses were:

   ⊹ Investing time and money into adopting and learning new
business systems that support data management and analysis.
   ⊹ High priority tasks are delayed because of manual data input and tracking.
   ⊹ Report errors because of missing data or poor analytical models.
   ⊹ Spending additional time or money to correct errors caused by poor reports.

Michael spends a lot of time auditing and processing business data and generating reports, but his informed business decisions aren't aligned with the outcome. He ends up losing time from the initial process, revenue, clientele, and additional time to correct the damage.

How Might We...

2.6
Based on our research insights, there are several core user pain points. To target these challenges, we asked:
Research Insight: Data entry, auditing, and analysis takes a lot of time, and takes away from tending to main business objectives.
How might we support users in the process of data input and analysis so users can focus on making informed business decisions?
Minimize manual data management labor
How might we decrease the occurrences of human error creating data inaccuracies and holes?
Research Insight: Data auditing and correction creates disruptions in workflow and delays progress of task-at-hand.
Reduce data errors and holes
Improve reporting experiences
How might we provide users with an accurate and streamlined process for reports?
Research Insight: There is a lot of work associated with generating reports, and gathering information for comprehensive reporting is tedious.
How might we help users quickly and accurately establish a systematic process suiting their team's/business' needs?
Research Insight: Streamlined processes created in one team or scenario doesn't apply for every other scenario, and often requires manual side-calculations to compare for accuracy.
Improve accuracy of streamlined work processes
How might we help users efficiently and effectively manage their data and generate accurate reports?

Ideation Workshop

3.1
We used our research findings and design goals to brainstorm ideas and organized them based on complexity and user impact. Our strategy was to prioritize the most impactful developments that were relatively least complex at this scope.

The highlighted notes in the figure below are our finalized ideas for development. They best complemented and supported each other to deliver three core features catering to user needs and pain points:
   ⊹ AI Business Companion
   ⊹ Dynamic Data Storage
   ⊹ Versatile Data Management and Analysis

Value Proposition

3.2
Streamline business processes through a single data management system. Reduce the amount of time spent gathering, organizing, and analyzing data, and focus on making quick and informed data-driven decisions. Consult with an AI companion to receive comprehensive business-specific reports and actionable insights to make data-informed business decisions.

Storyboard

3.3
We made a storyboard illustrating our product ideation and a use-case. The scenario is based on user insights, research, and additional potential cases.

For example, our research shows that data entry, data analysis, and market research are tedious processes that are prone to errors and data gaps. Having a generative AI companion is a solution to reduce errors and time spent creating and organizing data libraries and generating data analysis and visuals. However, it can also be used for generating comprehensive reports using qualitative and quantitative data.

Sketching Ideas

4.1
We sketched a rough depiction of our product ideation and defined design goals. We roughly imagined a user flow based on the design goals to layout a cohesive design. The user flow-sketches incorporated and detailed all of the design goals in every screen. We were strategic about real estate, hierarchy, and accessibility due to the amount of features in this first scope of development. This laid the foundation for us to develop an intuitive, seamless, and functional product.
4.2

Task Flows

We established task flows to dive into the nuances and design details we must consider before prototyping. Our four task flows help sequence and identify product details we must design. Each flow embodies one or a combination of our design goals.

We recognized that developing our design goals as a cohesive product was a complex task. For users, each feature would be more valuable in partnership with the other features.

Because of our product's complex nature, we prioritized developing designs to provide users with affordances and to support new users in learning to use the product to its fullest capacity. We developed an adaptive user interface, and focused on optimizing real estate, hierarchy, and accessibility.

Users may also be hesitant about sharing business data. We give users the opportunity to fast-track registration and explore the product's capabilities and framework before sharing any business information and data.

The task flows embody one or a combination of our design goals with the additional development considerations. The four task flows are:

Flow 1: Onboard and integrate external databases/accounts.
Flow 2: Customize dashboard module.
Flow 3: Read Tory's report, and check your inventory database.
Flow 4: Check stock levels in inventory database, and interact with Tory to make supply orders.

Low-Fi Prototypes

4.3
I was involved in the UX writing to support the creation of a prototype that told a story. Throughout the whole design process, I extensively collaborated with and supported our Lead UX Designer to advocate for and communicate user needs and design goals.
Users can easily integrate insights and data sets from various platforms to manage data and create reports through a singular platform.
Users can input, view, and manage data using a user-friendly, scannable, and interactive data grid. Data consistency helps users and AI with learning data definitions and business formats, and supporting efficient and effective data processing and analysis.  
Generative AI, Tory, can help streamline tedious tasks such as developing comprehensive reports and data visualizations. It can provide users with personalized qualitative and quantitative reports on matters related to their business.

Usability Testing & Feedback

4.4
We wanted to understand how participants would manage and analyze data to make reports and business decisions using our solution. Four tasks evaluated usability of our design goals' development.

We performed four remote, moderated usability test, and collected qualitative and quantitative data. We used the concurrent think-aloud (CTA), concurrent probing, and retrospective probing methods to collect qualitative data. Participants were also asked to screen share as they performed the usability test, and we asked permission to record for our review purposes.  We recorded tasks success rates for quantitative data.

Scenario
   ⊹ Small business owner with a brick-and-mortar and e-commerce store
   ⊹ Interested in an adaptive, efficient, and accurate data analysis process
   ⊹ Needs quick way to define and track key performance indicators (KPIs)  


Task 1: Create an account and sync all Shopify data. Skip 2-factor-authentication.
Task 2: Customize the name of "Module 1" to be "Sales Reports" and save your dashboard changes.
Task 3: Read Tory's report and check up on your inventory
Task 4: Check current stock levels for your "Java the Hutt" Coffee Beans, and have Tory help you find and email another supplier suitable to your business needs.
Usability Test Results
All participants completed all four tasks with a 100% success rate, but there were several challenges observed by and communicated to us throughout the usability tests.
Before
After
1. Lack of Affordances
Participants felt overloaded with tasks and information throughout the registration.
Changes made:
   ⊹ Added “required” indicators for certain input fields, and added option to skip certain steps to decrease registration difficulty and length.
   ⊹ Tory, the AI business companion, converses with the users to explain what is needed through each step.
Before
After
2. Unclear CTAs
Participants took long to discover icons and action buttons because they weren’t distinguished enough. They also shared that button labels were misleading.
Changes made:
   ⊹ Reduced the usage of unnecessary buttons and icons.
   ⊹ Thoughtfully designed main function icons and rearranged layout to improve icon representation, improve icon visibility, and decrease confusion.
Before
After
3. Unclear Notifications
Participants took long to identify AI, Tory’s, module and urgent messages.
Changes made:
   ⊹ Expanded Tory’s real estate, and gave the AI module a consistent position on all tabs for easy accessibility.
Before
After
4. Crowded Interface
Participants said it was hard to identify and digest information on the dashboard.
Changes made:
   ⊹ Compartmentalized all module details with module title.
   ⊹ Inverted module and page background colors to raise modules.
   ⊹ Added outlines and shadows around modules to clearly distinguish sections and improve visual layout.

High-Fi Prototype

5.1
This is a walk through of our high-fidelity prototype after reiteration targeting usability tests findings.

Reflecting

Our goal was to discover how we could build a product that helps entrepreneurs and small businesses. We brainstormed problem areas to look into, and after much research, we discovered that businesses face large losses in revenue and time when dealing with logistics. This is something the team was especially passionate and eager to learn more about and design a solution for because we've all dealt with some form of tedious retail or sales inventory tasks.

Our generative research showed us that entrepreneurs, small businesses, large businesses, analysts, and many other roles and titles, have common responsibilities and challenges in dealing with business management and data analysis. These roles share a common challenge: being required to perform various responsibilities and have a wide set of skills. Customer relations, market research, manual inventory labor, and manual data input and cleaning were among some of the general responsibilities and skills needed. Our team saw an opportunity in reducing the tedious labor that goes into accomplishing these tasks.

To provide a solution for individuals like Manager Michael, we decided that streamlining data management and analysis while increasing modeling accuracy was our largest priority. However, we were only given three weeks for product development, so we had to develop a development strategy that prioritized features most impactful for users. We focused on developing a dynamic dashboard, modeling features, streamlining capabilities, and inventory management.

The biggest challenge we faced was creating a scannable and simple interface that provides users with key points and necessary tools for effective and efficient data management and analysis. We had to experiment with white space, different layouts, and different iconography. These changes were successful improvements, but we foresee further changes to maintain an accessible and scannable interface while integrating more developments.

In future developments, we hope to conduct another round of usability tests on our latest prototype iteration to discover areas of improvement on our product in its first stage. To take the product to the next stage, we hope to make improvements on the current product, flush out features like dashboard templates and customizations, and start development for features that weren't prioritized in our first three-week cycle.