Trust Your Gut

Project Overview

For my Master's thesis project, I chose to tackle the wicked problem of food recalls and food safety.

It is not in a company's best interest to initiate a food recall, meaning often consumers unknowingly ingest contaminants and fall ill. User research uncovered that once a recall is publicized, it is too late to prevent damage to the health of a community. My solution is a mobile app whereby simply logging groceries and tracking symptoms, users enable the app to identify potentially harmful food products. These products are then professionally tested for bacteria and undeclared allergens.

Role

Entire product design from research to conception, design, and testing

Topics

Food Safety, Healthcare, Personal Health

Timeline

9 months

Methods

User interviews, problem tree analysis, user personas, user journey maps, storyboard, MSCW method, user flow, wireframes, moodboards, style guide, prototyping

Tools

Figma, Miro, Procreate, Adobe Photoshop, Adobe Illustrator, Adobe InDesign, Google Suite

Research

I started this journey by exploring the problem space in detail. My first step was to understand exactly what a food recall is and what causes them to happen. Before primary research, the goal here is not only to gain a thorough understanding of the space I plan to work in, but also to see where the current system may leave room for improvement.

What is a Food Recall?

"Voluntary action by a manufacturer or distributor to protect the public from products that may cause health problems or possible death."

How Do They Happen?

Undeclared Allergens
44.1%

Bacterial Contamination
31.9%

Foreign Material
8.9%

Size of the Problem

In 2021, there were 270 food and beverage recalls issued by the USDA and the FDA. Recalls are becoming more frequent in recent years due to the increasing complexity of products, more complex global supply chains, greater consumer awareness and more stringent product safety legislation.

In that same year, there were over 1,000 people confirmed who got sick by food that was later recalled. However, a major component of the problem is that these illnesses often go unreported as consumers are unable to connect their symptoms with any particular food item.

Effects on Consumer

  • 1 in 6 Americans per year contract a foodborne illness
  • 128,000 hospitalizations
  • 3,000 deaths
In addition to these numbers, an important note is that while most people who are affected by foodborne illness recover without lasting effects, some suffer long-term consequences such as kidney failure, nerve damage, or chronic arthritis.

Effects on Company

  • Average recalls costs company $10 million
  • Litigation costs
  • Drop in market value
  • Lost sales
  • Negative brand reputation
As a result of these risks, it is not usually within a company’s best interest to notify consumers about a recall. The current system places reliance on the business to put out a recall alert, but because of this financial detriment, an informational divide between companies and their consumers emerges.

Case Study: Jif Peanut Butter (2022)

To exemplify the assertion that companies might not always be transparent about safety practices, here’s a well-known case highlighted from 2022.
Graphic of a peanut butter jar detailing how many cases recalled, customer complaints, and interview subjects affected
This was certainly one of the larger recalls of the ear with over 9.5 million cases of peanut butter recalled, and it affected half of my later interview subjects. However, I’m highlighting this case for a different reason: while the recall took place in May 2022, a report reveals that plant management knew products were contaminated with salmonella as far back as December 2021. When companies are entrusted with the responsibility of practicing food safety, there’s a lot that tends to fall through the cracks.

Case Study: Daily Harvest (2022)

Another major recall from 2022 happened with the Lentil + Leek Crumbles from Daily Harvest, a vegan lifestyle brand popular among younger generations.
This case was especially shocking because not only was it widespread, but the symptoms were so unusual and severe. Most assume that food poisoning is an unpleasant couple days in the bathroom, but in certain cases it can lead to lifelong damage. With the Daily Harvest product, at least 30 people had to have their gallbladders surgically removed.

The contaminating ingredient turned out to be tara flour from Peru. While the investigation was ongoing, Daily Harvest was actively deleting comments and complaints on their social media to try and hide their mistake.

Problem Tree Analysis

I condensed all my secondary research into this problem tree analysis. It was helpful to visualize what causes a food recall to happen as well as the potential outcomes for both companies and consumers.
On top of the serious impacts for both companies and consumers, my major takeaway was that all of the cause of food recalls come from the manufacturing side. As one individual working on a thesis project, I don’t intend to tackle capitalism in the United States.

I have to hone in on a more approachable task. On the cause side of the tree, what I’m particularly interested in is how the company fails to alert consumers.

Interviews

Research Objectives

Coming into my primary research, I wanted to answer the question “how might we reduce health problems caused by recalled food items?” In order to do this, I needed to develop a firm understanding about the current state from the perspective of the affected consumers. My high-level research objectives were as follows:
I used these objectives to inform my interview script. For user interviews, I sought to recruit grocery store shoppers who have been personally affected by a food recall.

High Value Questions

I aimed to elicit storytelling from my participants to understand their attitudes and experiences surrounding food recalls. I separated my questions into three distinct categories: shopping attitudes, food recall experience, and moving forward.
After finding out that you may possess a contaminated food item, what was your next step?
How did this experience affect your relationship with the brand?
Please tell me about any preventative action you have taken to avoid a future recall.

Recruitment

Recruiting interview participants proved to be far more difficult than i thought. I initially went the social media route and made hundreds of posts throughout Facebook, Reddit, Instagram, Twitter, LinkedIn and even TikTok. I also was able to join a Facebook group specific to the aforementioned Daily Harvest recall.
This group had over 100 members, all of whom I assume were directly affected by the recall. However, no one seemed able or willing to speak with me to further discuss their experience.

This was frustrating, but this difficulty recruiting later proved to be an insight on its own. I realized that people are generally unwilling to waste the mental bandwidth worrying about something over which they had no control. This insight becomes relevant later on.

At this point, I was reaching out to everyone I know about recalls and truthfully becoming a little desperate. I had to take drastic measures.

I adopted a new strategy of standing outside grocery stores and asking shoppers coming in and out if they’ve ever experienced a recall. Beyond those unwilling to interact with a stranger, I was surprised to discover how many people have never even heard of a recall.

I was now confronted with the challenge that not only did I have to attempt to solve this wicked problem, but I had to make people understand it was a problem in the first place.

Ultimately, I was able to find 12 willing participants in time, 5 of which came from talking to people outside the grocery store and only one from social media, despite my hundreds of posts.

Interview Design

Once I was finally able to get all my participants, the interviews themselves were quite straightforward. The majority took place over Google Meet besides a few in person. Interviews averaged around 25-30 minutes, but for people who were especially brief I improvised additional questions to keep them talking.

I recorded all of the interviews with a transcriber service to consult later on so I wouldn’t be distracted taking notes. After completing each interview, I compiled all of the major takeaways into a concise debrief in Figma. Below is an example of one of the debriefs.

Interview Insights

0/12 participants take any preventative action for recalls
1/12 participants actively look for information about recalls
"How many recalls are out there right now that I'm unaware of?"
"There are so many things that could make you sick that you just can't see."
"Everyone’s just out to make a buck so there’s not too much trust in the morality of it."
As with my interview recruitment, I was somewhat shocked to discover my participants’ lack of proactivity in terms of food recalls. However, as I touched on before, looking over these interviews for a second time helped me realize that the major challenge is consumers feeling powerless to maintain their own food safety. There is certainly anxiety at play, but it transforms into cognitive dissonance when they don’t know what they can possibly do about it.

Research Synthesis

Affinity Mapping

After completing my 12 interview debriefs, I sorted all of the findings into an affinity map using Figma.
Here I highlight one particular section of the affinity map which speaks volumes. For my interview subjects’ attitudes surrounding recalls, you can tell with a glance how huge the category “little to no concern” is compared to the others.

In addition to my difficulty recruiting and the previous note that zero subjects take preventative action for recalls, this reaffirmed to me what I was considering at this time to be a lack of concern surrounding recalls.

In the affinity map, I also take a deeper look at how consumers currently find out about the recalls that may affect them.
The primary means of notification seemed to be word of mouth: people sharing recall news with their friends and family. This worked but was not a reliable or consistent solution. Social media users and news watchers were often overwhelmed with irrelevant information. The most effective solution seems to be direct notification, where stores such as BJ’s and Costco email the affected consumers about potential recalls. However, even in this case, notifications often came weeks or months later when it was too late to prevent damage.

User Persona

I compiled my primary user research into this persona called Busy Bradley. His goals and needs are to know about the food recalls that may affect him and also know what do if faced with one.

Later on in the process, I realized this persona was not entirely exhaustive and I create another to encompass more of my intended audience.

User Journey Map

The journey take the hypothetical user Bradley outlined above and predicts his grocery journey, informed by my user interviews. In this scenario, Bradley is affected by a Jif recall and unsure on what to do next.
As we can see, Bradley's journey is extremely frustrating and unintuitive. He is faced with a sickness and unable to connect it with a specific incident until he finds out about the peanut butter recall weeks later. He is upset that he has to waste food and money and miss work for a few days. Even after notifying his family and friends, he is still uncertain on if he is making a difference.

His main takeaway from this experience is to avoid dealing with another food contamination in the future, but he doesn't know what safety measures he can take. This is the area of the journey that I plan to focus on. My ambitious goal is to transfer the power from profit-driven companies to individual consumer responsibility.

Synthesis Conclusion

Thinking back to my problem tree analysis, I knew I couldn’t single-handedly tackle capitalism in the United States for my graduate thesis. In the current system, big businesses hold all of the power

What I sought to achieve was to reverse the power dynamic so that consumers could feel more in control of their own health and food safety instead of being at the whim of whatever big business decides to tell them. I wanted to design a product or service that could allow consumers to put forward recalls and food safety concerns themselves, through a community-driven approach. This brought me to my final how might we statement.

How might we mitigate the health risks of food recalls by redistributing power to consumers?

Proposed Solution

Through diving back into research and synthesis, I came up with the following solution to address the power imbalance between business and consumers

Trust Your Gut is a mobile application centered around users sharing and receiving information related to food recalls. The visual design invokes a lighthearted and compassionate style to lessen the gravity of a serious and sometimes shameful topic.

After going through onboarding, users are asked to input their dietary restrictions in order to receive the most relevant information. They scan their grocery receipts, which the backend analyzes and transforms to populate a standardized list of grocery items.

The principal feature of the application is symptom report. When feeling sick, the user will start this flow. First, they select the symptoms they are experiencing through an interactive body module. Next, they select the foods they have consumed in the last 24 hours, as well as any suspicions they may have and additional notes.

Employing state-of-the-art data analysis, the system uses all this information to determine if critical mass is reached for an alert. If so, the user will be presented with a note that there is cause for concern for a particular product and asked if they still possess it. If they do, they are urged to select an FDA-approved testing center through the application and drop off a sample to be officially tested. Staff will test the product and push a notification to all app users who have the product if they discover any bacterial contamination or undeclared allergens.

Storyboard

One of the major challenges in UX/product design involves taking a nebulous problem and turning it into an actionable solution. To help visualize this solution, I created the following storyboard which shows how Trust Your Gut helps enable transparency in food safety.

Scoping

With such a complex problem space, scoping becomes especially valuable to make sure I'm not taking too much on my plate. I started to write out all the of the features that could be included in the app. One thought that came up was allergy support.

Pivot Point

At this point in the journey, I had an important revelation. I realized that I have a huge target audience I glossed over in my initial discovery phase: people with allergies and GI diseases such as IBS and Celiac. What was especially exciting about this user group is that they already know very well the importance of caution in what they put in their bodies. As opposed to the previously discussed challenge of consumers not fully understanding the scope or even existence of the problem, I wouldn’t have to convince these users of anything. I created an additional user persona to encompass this new target audience.

MSCW Method

In my program we often use the analogy of a wedding cake versus a cupcake. The desired cupcake is comparable to an MVP, in that it is a finished yet not overly elaborate product. I used the MSCW (Must, Should, Could, Would) method to scope down my solution.
This scoping went through a large amount of iterations as I worked on nailing down my solution. In the end, I determined the hero features to include a recall news feed, a community board to connect and share concerns about food safety, receipt scanning, and symptom report which will be further demonstrated in my user flow.

Other ideas such as integrating with existing grocery apps to fill in receipts and providing services for restaurants could be valuable additions for a future iteration, but for now I intend to focus on the core solution.

User Flow

The user flow displays all the potential paths through the app that the user may take. Most importantly, it outlines how the symptom report will work, the main feature of Trust Your Gut.

 Let's look at the scenario through our persona Bradley's eyes. Bradley, experiencing a stomach ache, taps the CTA button to indicate he's feeling sick. Next, he is prompted to select his symptoms from an interactive body module. He then selects the food he's consumed in the last 24 hours which is populated from his grocery list. He will fill out if he suspects anything to be the cause and put in additional notes.

After inputting all his information, the app will run data analysis to determine if a statistically significant amount of other users have experienced similar symptoms with the same food items. This time, the system flags some chips suspected to be causing issues. Bradley is asked if he still has those chips, which he does, and finally he is urged to select a testing center to bring his sample. Professional staff will then test those chips to discover if they are in fact contaminated with any bacteria or undeclared allergens.

If the testing staff receive positive results, every user who has recently logged those chips is notified. All the inputted grocery receipts are stored in a database, which means the system will know who recently bought the chips and let them know about the possibility of contamination.

This way, people can find out if they have anything to worry about in their pantry without relying on big businesses or government agencies to make the first move.

Let's start building our app!

Wireframes

Before diving into high-fidelity designs, I find it essential to draw wireframes of the key experiential screens to gain a good understanding of the layout and content. I sketched out the majority of my screens below, which helped me better understand the flow and how I want my app to look without worrying about visual identity.

Visual Style

Mood Boards

I sought out inspiration in developing my visual identity by creating mood boards based on five relevant words: minimal, health-conscious, lighthearted, professional, and compassionate. Here are my two favorite mood boards from this exercise, which helped me nail down my color scheme.

Style Guide

I ultimately decided to move forward with the compassionate mood board. The idea of compassion is important to convey in my app because our digestive health can be extremely personal and often embarrassing to share. I aimed to capture this feeling using pinks and purples commonly associated with love (and conveniently similar to that of the internal organs) and greens associated with growth and good health.
I chose the font Poppins which I feel conveys a lighthearted yet professional tone I aim to embody, and in addition has a wide variety of text styles from which to chose. I also created vector illustrations of stylized stomach, brain and intestines. These cute characters are meant to pop up frequently throughout the user flow to demonstrate empathy for the user and lighten the serious or frustrated mood they may be experiencing in moments of medical distress.

The name Trust Your Gut came to me on a walk around the neighborhood after an intensive brainstorm session. Trust Your Gut captures the core spirit of my solution: to empower users to trust the signals from their own bodies, rather than be forced to believe what companies and corporations tell them to be true.

Prototype

After planning out the strategy, scope, structure, skeleton, and surface of my app, it became time to build my prototype. Included below is a fully functioning prototype I created in Figma, as well as a selection of core screens from the app.

Usability Testing

The principle of user-centered design hinges on the fact that designing is not the end. Potential end users must be heavily involved throughout the entire process. To confirm that I have in fact designed an appropriate and effective solution, my first step was to conduct usability testing with several real users.

Study Design

To validate my prototype design, I conducted usability testing with 5 users. 2 of these users I recruited from my primary interviews, but I also felt determined to test on my recently discovered user group of those with severe allergies or dietary restrictions. I managed to recruit 4 participants with Celiac Disease.

For the testing process, I asked users to share their screen going through my Figma prototype over Zoom or Google Meet. I asked them to carry out 6 discrete tasks and share as they went along. I recorded the interviews for later reference and compiled notes into Figma as the testing session occurred.

I kept track of the success, confidence level, and ease participants shared for each task and wrtoe down any additional notes as well.

Results

I averaged out the scores of all 5 participants across the 6 tasks, which yielded the following results.
It was validating to find that my users moved through the majority of tasks smoothly and efficiently. One user struggled to identify the community board icon but was able to make a post after figuring it out.
“Easy, and user-friendly, even a child could figure it out”
“Visually pleasing and cute; makes light of something difficult and unpleasant”
However, there was one flow that users had a lot of difficulty with. The main flow of selecting symptoms, an essential part of the core experience, scored the lowest in testing. Users failed in two places: firstly, 2 users could not identify where to go to report symptoms.

This icon underwent a lot of iterations. I designed the main CTA button to look like an upset stomach, which I attempted to familiarize users with through frequent repetition of the stomach logo in onboarding.

This strategy in fact backfired when one user assumed the button was simply a logo on the navigation bar as opposed to a clickable button. 3 users were able to figure out the icon’s meaning, but this deserved further consideration.

The other source of friction users experienced during this flow had to do with the taxonomy of symptoms. I initially grouped gastrointestinal symptoms based on the body part associated with them; for instance, users could locate headache in brain and bloating in stomach.

However, some of the symptoms were more nebulous: does gas happen in the stomach or intestines? Is fatigue a symptom of the brain or the whole body? As a result, several of my user testers struggled to figure out how to navigate to the specific symptom they were looking for. Furthermore, there was a challenge in the clickable area of the body parts being too small. One user made a valuable point that a user of this app should not have to have a thorough understanding of anatomy to select their symptoms.

Recommendations

In addition to the notes above, I received some specific feedback on both minor and major changes to improve my prototype.

Design Revisions

The first major revision I made was adding a brief walkthrough after the initial onboarding flow. The intention is to help clear up any confusion surrounding the iconography of the navigation bar, and to introduce new users to all of the main features of the app which they may not be familiar with before installing.
I also made sure to include a clear button to skip through the walkthrough in case any users don’t want to read and would rather learn by doing, as one of my user testers indicated.
I indicated the alert popup to be clickable by adding a slight drop shadow and arrow indicator.
Additionally, I edited the onboarding survey by merging dietary restrictions in with allergies so that, as an example, vegetarian users wouldn’t have to be notified about recalls involving chicken or ground beef.

Reimagining Symptom Report

As discussed earlier, a major problem surrounding the symptom report flow was exposed during usability testing. Users didn’t understand at first that they needed to click on the body parts, and even after they had trouble figuring out which symptoms belonged in what category.

I altered these screens to include category headings which drop down with symptoms. Instead of requiring users to tap on the body parts, they could tap on the category header, which is more in line with convention. I also spent time reworking the taxonomy in order to make it more in line with common sense rather than requiring knowledge of anatomy.

In an ideal world, I would like to conduct another round of usability testing  to evaluate the effectiveness of these changes.

Validation Testing

Conducting usability testing helped me clarify the effectiveness of my prototype and identify room for improvement. However, a well-designed app is far from the only ingredient for a digital product to succeed.

According to IDEO, there are three criteria that are likely to predict a product’s success in the market: viability, desirability, and feasibility. A successful product sits at the intersection of these criteria.
In order to test if my product meets these criteria, I wrote out a list of risky assumptions. These are the foundational beliefs of my app, meaning if they were not met the product could be in serious trouble when it reached market. I organized these assumptions into the following matrix.
The top three risky assumptions I identified are as follows:
When I began planning out steps to test these assumptions, the first seemed easy enough but the second two presented a major obstacle. Through my own brainstorming and talking with my advisors and colleagues, it seemed like there would be no way to test these two suppositions without willingly making someone sick or getting “lucky” (I use this word facetiously) with a participant coincidentally contracting food poisoning during a two-three week experiment. Therefore, I decided to shift to include the next two riskiest assumptions.

Pretotyping Design

I took the final three risky assumptions and turned them into actionable hypotheses which I designed experiments for utilizing Alberto Savoia’s pretotyping framework.
Risky Assumption
XYZ Statement
Pretotype Experiment
The app will be able to attract a large enough user base to make the alert system effective.
At least 3% of website visitors will sign up to express interest in the app.
Fake Door: Create a landing page outlining the key features of the app and a sign-up form where users can express further interest. Advertise this page on social media and track engagements.
Users will find the lighthearted visual design appealing, rather than trivializing the seriousness of the issues.
The lighthearted landing page will draw at least 15% more clicks than a more serious design.
Fake Door + A/B Testing: Create two alternate landing pages outlining the key features of the app, one with the original (lighthearted) visual design, and another with a more serious visual design. Advertise both pages on social media and track which receives more engagements.
Users will remember to scan their receipts every time they go grocery shopping.
At least 2 out of 6 testers will remember to log all of their grocery receipts.
Mechanical Turk: Recruit potential users and ask them to send a picture of their grocery receipts over text during a three week period. In exchange, I will send them issues or concerns about the product(s) they have purchased.

Fake Door Experiment

The core feature of my app is the alert system. In order to make the alert system effective, there need to be a significant number of actively engaged users. To test this assumption, I used Wix to set up a Trust Your Gut landing page which outlines the key features of the app and encourages users to plug in their emails to be notified when the app is available.

This could effectively test feasibility of the system and had the added benefit of testing desirability. I set a goal of reaching 3% sign ups out of all website visitors.
I created an ad using Photoshop and ran low-cost campaigns on Facebook and Reddit to draw traffic to this landing page over the course of one week.

Results

I received 308 unique visitors in total, with a maximum of 123 in one day. 17 visitors signed up with their email, which evaluates to 5.6% of visitors, well above the 3% margin. In Alberto Savoia’s language I was able to get significant skin in the game.
It was validating to find that my users moved through the majority of tasks smoothly and efficiently. One user struggled to identify the community board icon but was able to make a post after figuring it out.

A/B Testing

The next assumption, I had been curious about since I first finalized my visual style. I wanted to test if users find the lighthearted visual design appealing, rather than trivializing the seriousness of the issues. As explained earlier, I strove to make my app’s style cute and approachable in an attempt to lighten the serious or frustrated mood users may be experiencing. However, I felt perhaps there was a fine line to ride in this sense.

To test this assumption, I created an additional Wix landing page utilizing exactly the same language and layouts, but with a more serious visual style evocative of a traditional healthcare website.
I switched to a serif font for headers and a deep blue common among other healthcare services I looked at. I also made an alternate ad using the serious style and ran a week-long A/B test, a feature built in to Meta’s ad platform.

Half the audience would be shown an ad in the original lighthearted style and the other half would see the updated serious ad. My hypothesis was that the lighthearted style would receive 15% more clicks or greater.

Results

I was somewhat surprised to find that not only were the results far closer than I thought, but the serious ad in fact did slightly better. The serious ad had a $0.33 cost per click vs. $0.41 of the lighthearted.

In terms of clickthrough rate (CTR), the serious ad drew 2.54% vs 1.98% for the lighthearted.

While these results did not meet my personal benchmark, I discovered that both ads scored well above the industry standard. As of 2022, the average Facebook CTR is 1.04% for technology and 0.83% for healthcare. My results were promising in terms of desirability.

Mechanical Turk

The final assumption I sought to test in lieu of logging symptoms was if users would remember to scan their receipts every time they go grocery shopping. This is an important component of my app to make sure that everyone who is at potential risk of a contamination can be properly reached and notified. I recruited six users (five of which also completed usability testing) and set up a simple experiment to validate this assumption.

Over the course of two and a half weeks, I asked users to send me their receipts after they went grocery shopping. They would send photos and in return I would act as proxy version of my app and send them any concerning information about their groceries. I was testing the hypothesis that at least 2 out of 6 testers would remember to log all of their grocery receipts.

I maintained a table in Figma to remain organized on all the receipts I was getting and where they were coming from, and frequently browsed the FDA’s web page for active recalls to help keep my testers up to date.

Results

Notably, I received all of zero receipts within the first week of my experiment. I could have written this off to nobody having gone shopping in that week, but I found that unlikely.

An important note is that I was sending no reminders or updates in the first week beyond the initial ask. This was an intentional decision because my app design did not incorporate push notifications reminding users to upload their receipts. I hoped it would be a self-motivated act, but again, this is exactly why I was conducting testing.

After the first week of no engagement, my advisor suggested that I change my strategy. For the remaining time of the experiment, I decided to send each user a weekly text reminding them to send their grocery receipt with an additional note about ongoing recalls.

As demonstrated in the table, once I started sending these reminders, the results improved dramatically. I received the total receipts of the time period from 4 out of 6 of my users, surpassing the 2 out of 6 mark. One user sent me 3 receipts over the remaining one and a half weeks. Despite the 2 users who did not engage, I felt confident about the success of this experiment.

Recommendations

Reviewing my initial planning sheet, my experiments in attracting a large enough user base and scanning receipts succeeded, while the evaluation of style preferences failed to reach my hypothesis by a small margin.
Risky Assumption
XYZ Statement
Pretotype Results
The app will be able to attract a large enough user base to make the alert system effective.
At least 3% of website visitors will sign up to express interest in the app.
PASS - 5.6% of website visitors signed up to be notified when the app is available.
Users will find the lighthearted visual design appealing, rather than trivializing the seriousness of the issues.
The lighthearted landing page will draw at least 15% more clicks than a more serious design.
FAIL - 0.5% less users clicked on a lighthearted ad over a more serious design.
Users will remember to scan their receipts every time they go grocery shopping.
At least 2 out of 6 testers will remember to log all of their grocery receipts.
PASS - 4/6 testers remembered to log all of their grocery receipts*.

*with regular reminders
I feel confident in the results of the Fake Door experiment, but the other tests warrant further exploration and possible changes in future iterations.

For the grocery receipts, the Mechanical Turk experiment revealed to me the importance of building notifications into my app design reminding users to regularly upload their receipts. I could send periodic notifications, but I had the idea to take it a step further by enabling location tracking to notify users at the opportune time: right as they are leaving the grocery store.

The visual style assessment also suggests changes may be necessary. I was happy with the fact that both ad campaigns came in above the industry benchmark, but the lighthearted style did not beat out the serious style as I had anticipated. My suspicion for this is the power of convention: people are used to seeing a digital healthcare product in a certain way, and therefore have more trust in a product which meets those expectations.

I believe the small difference between the two campaigns does not imply the necessity of a complete design overhaul, or at least not before further testing. In a future iteration, I would consider dialing back the lighthearted illustrations and perhaps finding more of a middle ground between the cute and serious styles.

Business Strategy

At this point in the process, I’ve designed a fully-functioning prototype, tested the flow with users, and validated some key assumptions that go into my product.

However, if I want to take this app to market, there’s more planning to do. It’s important to envision a business strategy to further clarify my next steps. By clarifying my goals, securing funding, guiding my decision-making, and staying focused, a business plan will help me build a successful and sustainable app that makes a positive impact on people's health and wellness. My first step was to write out a mission statement, and then to figure out how to achieve it.

To empower consumers to take charge of their digestive health by enabling transparency in food safety.

Short-Term Goals

Long-Term Goals

SWOT Analysis

Strengths
Weaknesses
  • Connect consumers with relevant food safety information without a dependency on companies to take the first step
  • Present a platform for like-minded users to share thoughts and concerns about food safety
  • Caters to users’ specific needs, such as allergies and eating habits, to produce a streamlined, personalized experience
  • Compiles all recall information in one platform
  • Reliant on an active user base to populate app with critical information
  • Lack of alerts at a given time could provide users with a false sense of security
  • Not exhaustive; app does not provide an ailment match for symptoms or general guidelines for food safety
  • Limited scope focused solely on food recalls and contaminations
Opportunities
Threats
  • Can continue iterating and expanding on features and functionality to provide more for food safety
  • Can grow to collaborate with or hire on medical professionals to provide professional information for users experiencing medical distress
  • Can collaborate with existing social networks to greatly expand audience and reach
  • Reliant on collaboration from FDA or comparable government agency to reach full potential
  • Could be subject to lawsuits from food production companies for spreading negative information about their products
  • May face security threats and/or data privacy concerns related to user data and information

Competitive Analysis

Earlier I looked at current solutions for food recalls, including manufacturing-side management softwares and notification systems at big-box grocery stores. Now that I have a better understanding of my product, I will take a deeper look at market competitors.

From my research, there is no product on the market which uses a community-driven approach for food recalls like Trust Your Gut does. However, there are a few apps which offer a similar value proposition.
All of the apps provide unique and valuable features which could inform Trust Your Gut’s model. Trust Your Gut’s main competitive advantage, however, comes from the user-driven feedback loop.

In addition to the platforms which offer a similar value proposition, I also looked at two successful apps in adjacent sectors to inspire my business model.
NIMA Partners
Innovative approach incorporates an IoT gluten detector: portable, easy-to-use machine can analyze a crumb for gluten and return results within minutes.

Features:
  • Scan barcodes to collect product information
  • Track results of packaged products and restaurants
  • Crowd sourcing compiles test results from all users
Yuka
Independent app which scans barcodes to analyzes the health impact of food products and cosmetics, providing a score from 0 to 100.

Features:
  • Offers detailed information about nutrition and additives
  • Provides recommendations for healthier products
  • Independently run and funded means no conflict of interest
Both of these apps offer innovative solutions to the often debilitating challenge of food safety. If the opportunity arose I would like to pursue relationships with these businesses. Trust Your Gut could partner with NIMA to offer instant gluten detection, as well as Yuka to provide a well-rounded summary about the nutrition and safety of food products.

Growth Strategy

Conclusion

If you've managed to read all the way to here, I commend you. This was an exhaustive approach to a wicked problem which I approached with my full UX toolkit. This full case study along with more information regarding technical requirements (possible development practices, database diagrams, etc.) will shortly be available in the form of an ePublication posted here. Thank you for reading.
Overview
Research
Interviews
Synthesis
Solution
Scoping
User Flow
Wireframes
Style
Prototype
Testing
Revisions
Validation
Business
<
return to top
Next: VR Nursing Simulation >
© Elon Skolnik 2023
elonskolnik@gmail.com