BestBouquet is a concept that I developed for a tool that provides bouquet suggestions for a variety of occasions.
Background
I took inspiration from a project discussed in UX Strategy (by Jaime Levy). She used https://itsthisforthat.com/ - a random startup value proposition generator - to come up with an idea for a value proposition (“Airbnb for Weddings”). When I tried out the site, I got “1-800 Flowers for Funeral Homes.” I decided to look up 1-800 Flowers to see if the value proposition had any merit. After visiting 1800flowers.com, I felt overwhelmed by the number of options available and imagined that if I were shopping for flowers for a funeral, I would want more direction. From there, I thought to create an app that would provide helpful suggestions for customers shopping for bouquets, and focused on a user flow for purchasing a bouquet for a funeral.
I also borrowed some inspiration from BuzzFeed quizzes because they are engaging, and if used in an ecommerce environment, they have a tremendous potential to drive traffic by directing shoppers to products that fit their needs and preferences.
The concept for BestBouquet grew from the understanding that different flowers and flower arrangements have different meanings in different contexts and cultures. From there, I imagined the following scenario:
Imagine a close friend passed away and you want to buy flowers for the funeral. You start by doing a Google search and find so many bouquets of different types, sizes, and prices. You dig a little deeper and learn that specific flowers have different meanings in different cultures. Your friend is of a different religious and cultural background than you, and you are worried about buying the wrong bouquet. You wonder if it is even appropriate for you to buy flowers for the service. You want to pay your heartfelt respects, and would like some quick advice on what bouquet to buy without needing to research the meanings of different flowers or go to the florist.
With this scenario in mind, I began sketching:
Competitive Analysis
After coming up with the basic concept for BestBouquet, I wanted to get a sense of what competition exists and to analyze how they solve for the same or similar problems. I discovered Gwyn - 1800flowers.com's current AI software that aims to help customers find the perfect product. Because I was focusing on the customer journey of finding flowers for a funeral, I tested Gwyn to see what she would suggest. After asking me several questions, she ended up providing me with a number to call to talk to an expert. As a pretend customer, this felt a lot like calling a service and being repeatedly transferred to different departments, each falsely promising to have the capacity to solve my problem. I hypothesized that exploring more of a quizlike model that takes a list of products and effectively narrows suggestions based on customers' responses to multiple-choice questions could be equally, if not more effective than an AI-based tool.
If I were to expand this portion of the project, I would focus on learning more about how and why people discover and buy bouquets and to design around those findings rather than moving quickly through concepting, sketching, and competitive analysis. However, because this project was part of my Prototyping and Testing class at SVC, I needed to quickly move into creating low-fidelity prototypes to test.
Low-Fidelity Prototyping & Testing
After sketching ideas, performing competitive analysis, and identifying competitor pain points, I created wireframes which I then converted into a clickable prototype in Marvel. During my first round of testing, my goal was to test to overall concept and flow of finding the best funeral flowers using BestBouquet.
After incorporating the feedback I received during testing and iterating on the funeral flowers flow, I brought my prototype into a higher fidelity. My priority was to create a clean, simple visual identity that supported BestBouquet's focus on floristry and felt trustworthy and reliable. Based on testing results, participants appreciated the simplicity of the interface, but at certain points in the flow they expressed that they wanted more information. Therefore, it was my task to balance providing enough information and maintaining a clean, streamlined experience.
High-Fidelity Prototype Testing
For my second round of testing, I created another clickable prototype in Marvel. This time, my goal was not only to test to that I had improved the flow between iterations, but also to discover whether the visual design of the UI supported my overall concept and reflected user expectations for this type of tool.
Before further designing and developing this product, it would be critical to validate the value proposition of BestBouquet. If evidence suggested that people would find it valuable, I would explore the pros and cons of creating it as a website or app and whether it could be connected to the specific bouquets in flower retailer's site, or if it should function as a standalone website or app that provides general flower suggestions. If this product were to be seen through development, I would need to conduct ethnographic research to ensure that the app offers appropriate bouquet suggestions. I would also delve deeper into the benefits and pitfalls of using and not using AI for this type of tool.
