Case Study:
Melville Candy - Custom ERP Database App
Published Desktop App Design
Project Overview
The Product:
Melville Candy is a manufacturing company located just outside of Boston. They produce custom food products that are tailored to the requirements of each client. As the company grew, they needed a way to record and access all of their accumulated custom order data. To solve this problem, I developed a fully custom ERP database for Melville Candy, that allowed all departments to record, retrieve and recall data through a shared system. The ERP I designed gave all departments the ability to record pertinent data, as well as export PDFs and excel files of recipes, quotes, sales reports, price lists, and order specifications.
Project Duration: June 2014 - May 2021
The Problem: The manufacturer needs a way to organize hundreds of daily custom product orders so they can be easily referenced by all departments and quickly delivered to the production staff.
The Goal: Design and build an ERP (Enterprise Resource Planning ) app that will allow all of the departments in the company to use the same application to store and reference their data, communicate with one another, and create internal and external reports.
My Role: UX / UI designer conceiving of and actualizing the ERP app from concept to delivery. Conducting the user research to design the apps functionality. Building, updating, and maintaining the app daily over the span of 10 years.
My Responsibilities: Designing with empathy and without bias. Planning & conducting user research, interviews, & surveys. Synthesizing & diagramming interview & usability study data to form insights. Composing wireframes, building design systems, & creating prototypes via Figma & Adobe Xd. Iterating on designs & progressing user research.
Understanding the User
• User Research
• Pain Points
• Persona
• User Workflows
User Research
Summary:
I conducted interviews and created empathy maps to understand the users I’m designing for and their needs. A primary user group identified through research was: Production supervisors that need to be able to train their workers to all use the same process. This user group confirmed initial assumptions about what would be needed to manufacture Melville Candy orders, but research also revealed that supervisors were not the only users looking for an easy way to organize data. Other users included the production workers that were seeking out a way to have the order data prepped automatically so they did not have to spend so much time looking up product specifications.
Pain Points
Time: Hundreds of custom orders needed to be manually composited and delivered to the production staff, and it was taking up a lot of valuable time.
Accessibility: The production staff are not all tech savvy or fluent in English
IA: The orders that are delivered from management are confusing, long excel documents with no images, organizational structure, visual patterns, or relational functionality.
Missing Data: Staff does not always receive all of the data necessary to complete their work, and has to stop what they are doing to go back and find the missing information.
Persona
Persona: Margaret
Problem Statement:
The user, Margaret, is a Packaging Supervisor that works in a manufacturing company in a major city, who needs a quick and easy way to deliver custom order specs to her staff, because she needs them to pack hundreds of orders and meet a shipping schedule each day.
User Workflows
Department Flows
Initial Ideas:
I composed a few digital wireframes using Figma and Filemaker to begin experimenting with how the data tables should be connected. The relationships between the tables would be the crucial pathways between the departments. Through my user and company research, I learned that Melville Candy's business model relied heavily on suppliers and inventory so those tables became the starting point of the process chain. I determined that the ERP would utilize primary keys to connect the data between the tables, and those keys would be auto-generated by the database to replace previous man-made product ids. Past ids would be archived and used for reference only. The primary keys would be created when a product was connected with a client, and once the primary key was assigned, it would be used as a sku# for the client's custom instance of that product.
Main User Workflows
Custom Product Creation:
I designed the first user workflow targeting the Sales and R&D departments. Through my research, I determined that these 2 departments needed to work in unison to create custom products. Sales would receive requests directly from clients and would then deliver them to R&D. Whether the client was requesting a customization of a stock product, or a totally new product design, R&D would then turn the request into a new custom sku for that particular client, and record all of the product data in that new sku record. R&D would then create a list of the skus in a quote for the sales department. Sales would then access the same quote to add pricing and other client data. Once the quote was completed, it could be used to generate instructions and sample lists to give to the production staff. The Production staff would then use the instructions and lists to create the first sample run of the product, that would be reviewed by R&D, then shipped out to the client for their final approval. Colors were assigned to steps of the process to flag what stage each product was currently in. Red signified a new product that had not been approved yet. Purple signified that R&D had previously created a template of the product and it needed minimal customization.
Alternative User Workflow
Product Re-Orders and Final Release:
My research revealed that one of the most common scenarios that production would encounter was the "re-order". They needed a way to reference past orders and product data. I trained the production staff to look up products using their new sku# and review the recorded specification that was linked to that number. They were then instructed to follow the past order specifications unless the product had been flagged for changes by another department. Initial testing showed this method proved very successful in reducing the time production had to spend on collecting order data each day. It allowed the production manager, purchasing, and inventory managers to all plan for orders much more efficiently. It also made it possible for R&D to create product specifications and recipes in advance, instead of waiting for an order to trigger the process.
Starting the Design
• Sitemap
• Digital Wireframes
• Lo-fi Prototype
• Usability Studies
Sitemap
Organizing the Design:
With the initial design started, I began work on the information architecture. I used the App’s sitemap to guide the organizational structure of each screen’s design to ensure a cohesive and consistent experience. I made sure to consider the users journey when constructing the sitemap. My research suggested that most users would prefer to use a linear step-by-step process for each department, so I structured the site around that option.
Digital Wireframes
Planning the Design:
After ideating and drafting some paper wireframes, I created the initial digital designs for the product records. These designs focused on guiding users through the product creation steps. It was essential to have a field to enter the data for each part used in the creation of the product. Each user would record the product data. Creating the record would automatically create the sequential sku# and add the date modified and the account details for the user that created the record.
Low-fi Prototype
Connecting the Design:
Using the completed set of digital wireframes, I created a low-fidelity prototype. The primary user flow I connected was creating a product record then creating a quote and and adding it to an order, so the prototype could be employed in a usability study.
Usability Study: Parameters
Study Type:
Moderated
Location:
USA, remote
Participants:
10 participants
Length:
15-30 minutes per week
Usability Study: Findings & Insights
Testing the App:
I conducted two rounds of usability studies. Findings from the first study helped guide the designs from wireframes to mockups. The second study used a high-fidelity prototype and revealed what aspects of the mockups needed refining.
Useful
Users thought the app was very useful and looked forward to its implementation.
Step Numbers
Users wanted visual cues to guide them through all of their department's steps.
Customization
Users want a way to go back and review past records, orders and images.
Refining the Design
• Mockups
• Hi-fi Prototype
• Accessibility
• Product Launch
Mockups
Developing the Design:
I conducted user research to find icons that would make it easy for users to navigate to the appropriate departments from the home page of the database. I also included text labels to ensure there would be no confusion as to what button led where.
Mockups
Refining the Design:
My initial designs were based on user interviews and focused on collecting and recording as much product data as possible. As I conducted more research, I modified the prototype to break up the product record into several related data tables, to prevent the user from having to enter redundant data.
Mockups
Optimizing the Design:
I eventually built the product list to just be a superficial shell that housed only related data. Clicking on the arrow on the product list layout opens the product full build layout. The recipe, packaging, and inventory tables were all linked to the product build layout, and their subsequent data could be viewed from each related record.
Key Mockups
Hi-fi Prototype
Evolving the Design:
I built the prototype in Filemaker and specifically designed it to be in a constant state of testing. The initial goals were met, but as the systems began to be implemented throughout the departments, I received more requests for additional functionality. I made it my goal to increase the functionality by a certain percentage each week, and over the course of 10 years built the system into a robust ERP that included all departments in the company.
Related Tables
Design Framework:
I determined which table relationships would need to be created to deliver the desired functionality, and composed the relationship graph. As I added more and more data to the database, more related tables were also added as needed.
Accessibility Considerations
Included text and instructions in both English and Mandarin where possible.
Used icons to help make navigation easier, and included tooltips for enhanced guidance through the app
Used large images to help all users better understand the flow.
Development
I developed the final functioning app using Filemaker Pro Database software, and hosted the app on the manufacturing company's internal server.
The first production test was in 2014. The manufacturing company continues to use the software in 2021.
The app successfully improved a multitude of production issues, including order processing, sales presentations, and product management.
Before & After
Improving the Design:
One of the main goals of the ERP was to replace the outdated and inadequate system that was being used. Below on the left is an example of the production slip that was traditionally given to the production staff. There were no images, and not enough fields to properly deliver all of the necessary data. The example below on the right is the final refined production slip design that proved successful in making order processing much more efficient.
Product Launch
Using the App
Improving Production
After a number of training sessions to explain to the various departments how to use the ERP app, they began to create quotes and process orders in record time. The development processes had been expedited 100x over and it allowed for better planning and execution across all departments. Mistake rates and rework were down, and finally all departments were on the same page.
Going Forward
• Takeaways
• Next Steps
Takeaways
Impact:
The app shaved hours off of product development and made order processing much more efficient :
“The app makes it very easy to complete the orders much faster, we won't ever go back to using the old system.”
What I learned:
While designing the Custom ERP Database, I learned that the first ideas really are only the beginning of the process. Usability studies and peer feedback influenced each iteration of the app's designs. I plan to continue testing the app and adding features. I hope to eventually make it possible to access the database via the web.
Next Steps
Conduct another round of usability studies to validate whether the pain points users experienced have been effectively addressed.
Conduct more user research to determine other features that should be added to further improve the app.
Conduct more user research to determine if users would like to be able to access the database via web as well.
Let's Connect!
I really appreciate your time reviewing my work on the Melville Candy Custom ERP Database app! If you’d like to see more or get in touch, my contact information is below.
Thank You!