This was my first UX project - a concept project for my Springboard UX/UI certification. My personal interest in sustainability and inequality led me to choose food waste as a project topic.

Document groceries by capturing store receipts

Allow capturing of multiple sections for long receipts

Convert physical receipt quickly into digital list

Conversion of physical receipt into digital grocery list utilizes Optical Character Recognition (OCR). Conversion of item abbreviations on receipt to item names on grocery list utilizes compilation of major grocery retailers’ databases. Expiration dates and storage recommendations vetted by scientists form a backend database, corresponding to the correct items.

Narrowing the focus through
White Paper Research

It was impossible to consider all aspects of this massive global problem of food waste, with its many complexities, within a single project tied to a 6-month timeline. I gravitated toward the most severe parts of the problem that I found during white paper research.

Contributors to food waste along food chain

Contributors to household level food waste

As you can see in the charts above, private households and foods not being used before they go bad, contribute the most to food waste. This sharpened my focus toward perishable groceries and home cooks who make usage decisions.

I also found out that “youths tend to waste more than their elders”, and “people living in urban areas tend to waste more than people living in rural areas”. This made me narrow my target interviewees to young, urban dwellers.

Distilling behavioral insights from User Interviews

With a clearer focus, I was ready and eager to interview 5 young, urban home cooks to find out what is driving their behavior of not using foods before they go bad.

Through affinity grouping and coding, I synthesized and analyzed the data I collected from interviews, and arrived at 2 main insights:

1

Out of sight, out of mind

My interviewees struggled to keep track of foods that are not frequently or easily seen, thought of or used.

Uncertain and unconfident

My interviewees lacked certainty and confidence in the ways they stored foods and interpreted foods for spoilage.

After interviews, I constructed a user persona that encompassed the commonalities I noticed across the 5 home cooks. Meet Michelle, a home cook who wishes to reduce her food waste at home.

Next, I wanted to zoom in further by following Michelle along her weekly food waste journey, to uncover phases of the journey that serve as opportunities for behavioral improvements, which are also potential answers to our How Might We questions.

I saw that there’s an opportunity to improve her storage actions so she is confident that her foods will last longer, giving her a longer window to use them. There’s also an opportunity to improve her awareness of foods she has before the Cook phase, and help her remember to use those foods during the Cook phase. Finally, there’s an opportunity to improve her spoilage interpretation skills so she is confident that the food thrown away is truly spoiled versus still safe to eat.

User Stories as Design Objectives

Four main user stories, which serve as my design objectives, were derived from analyzing the user journey:

Due to limited time and the focus on a Minimum Viable Product (MVP), I had to let go of the spoilage interpretation user story. I thought it would be better to focus on the phases more upstream, so home cooks can preserve and use foods as much as they can, before they even get to that final Discard phase.

To uncover any competitor gaps, I studied 5 existing mobile apps that attempt to reduce household food waste. (Eidos is a student project not yet launched.) They all lacked the early focus of storage improvement, except “A Good Opportunity”.

But “A Good Opportunity” lacked convenience and efficiency. It presents storage tips as articles that require search. It would be better to tie storage information automatically to foods home cooks have. Busy urban dwellers like Michelle are more motivated to take action to reduce food waste if it takes less time and effort.

With Michelle and her user stories serving as a solid foundation, I began designing a solution for her - starting with rough ideas on paper sketches and then guerrilla testing the sketch prototype with 5 different people.

Three biggest issues discovered in testing helped me iterate from sketches to wireframes:

Before

No permission was asked for sending reminder notifications and no confirmation was given for which items have reminders set successfully.

The evolution of grocery list and storage recommendation:

Although they did not derive directly from guerrilla testing’s major issues, significant changes were made to the grocery list and storage screen during the process of turning sketches to wireframes.

Before diving into high fidelity mockup design, I first defined the brand platform and style guide as design guidelines. The name Kept was a deliberate past-tense word envisioning the future state of foods being kept for longer, used and not wasted. The main color brown was part of a harvest themed color palette, chosen because of its connection to food. Brown as a color also signifies reliability, a brand attribute I wanted this app to have.

My high fidelity design improved and evolved through two rounds of usability testing. Two major issues discovered in the first usability test were alleviated and one issue was completely resolved in the second usability test.

Impact of design change seen in second usability test:

4/5 participants found the first tap on their own for storage recommendations and setting reminders.

There is still room to improve because some participants said they considered a different first tap to what I designed. If there was more time, I would explore better ways of discovering these features.

Here are some ways to measure the success of this app from the perspective of home cooks:

Instead of overhauling the entire flow after testers struggled with where to tap first to find storage recommendations, I added tooltips, and testing showed that they alleviated the discoverability problem.