BiteWise

1. Overview

BiteWise is a digital health solution designed for people juggling dietary restrictions, health conditions, and busy lifestyles. The brief was to design a health app focused on diet and nutrition. BiteWise directly solves a health problem and has room to meaningfully improve on existing apps, and allowed for several smaller ideas into one coherent product.

2. The Problem

People with dietary restrictions face a stacked set of barriers. Paper research made the shape of the problem clear before any design work started:

• People with allergies are often reluctant to eat out, unsure of ingredients or preparation methods (Clinical & Experimental Allergy Survey, 2010)
• 35% of people modify their diet due to adverse food reactions (European Journal of Allergy & Clinical Immunology, 2012)
• Dietary planning apps have been shown to help manage chronic disease through personalised advice (Journal of the Academy of Nutrition and Dietetics, 2017)
• 22% of college students live with a chronic condition that affects what they can eat (National Centre for Education, 2021)
• Relying solely on online sources for dietary choices carries real health risk for people with restrictions (our own survey of 80 participants, 2023)

Framed as a problem statement, five clusters emerged: diet restrictions (not knowing what's safe to eat), eating out (no way to tell if a place is diet-friendly), confusion around cooking (jargon, poor ability, lack of recipe variety), lifestyle habits (time pressure, family influence, food cravings), and budget (eating out is expensive, especially on a student income).

3. Research & Accessibility

To validate the problem, we surveyed 83 people across 27 questions, deliberately including people with access needs alongside general users. Key findings:

• 76 of 83 respondents were 18–40, the app's target range
• Most people ate out only 1–2 times a week but cooked at home 3–5 times a week, home cooking, not eating out, was the dominant behaviour to design for
• Vegetarian, low-carb, and gluten-free were the most common diets followed
• When looking for suitable meals, most people either searched online informally or didn't actively seek help at all. This was a sign the market wasn't serving them well
• The overwhelming preference for learning about a new diet or recipe was watching videos (48 of 83), far ahead of reading articles or consulting a professional

Accessibility (A11Y) was treated as a first-class constraint from the start rather than a retrofit, guided by Stephanie Walter's principle: "Why fix it, when you could design and build it right to start with?" Working from WCAG 2.0, three rules shaped every screen: touch targets at least 9×9mm regardless of device, no content flashing more than 3 times per second, and a font-size floor of 10pt to support low vision and dyslexia.

4. Personas

Five personas grounded every decision in a real constraint rather than an abstract user:

• Michelle, 32 — keto, wants to lose weight, struggles to cater for her own diet alongside her son's
• Thomas, 25 — vegan and lactose-free, wants to learn to cook, but struggles to read recipes due to dyslexia
• Stephen, 22 — low-fat due to Crohn's disease, wants meals near campus within a student budget, unsure what he can safely eat at university
• Anna, 38 — managing a household where her husband has diabetes, one child is obese, and two others are picky eaters; lacks confidence in what counts as "healthy"
• Olivia, 35 — gluten-free due to Celiac disease, wants reliable eat-out options she can find fast during a short lunch break

Scenarios were built around each persona to map how they'd actually use the app — Michelle using meal-prep planning and shopping reminders to save time and money, Thomas leaning on simplified cooking instructions to build confidence, Stephen using personalised recipe suggestions with an integrated shopping list, Anna reducing food waste through customised family meals, and Olivia using location-based, diet-filtered restaurant recommendations.

5. From Insight to Features

Brainstorming (informed by the survey, persona gaps, and a teardown of existing apps) produced a long feature list, which we grouped into primary, secondary, and "later" tiers:

Competitor analysis against Yummly, Simple, HelloFresh, Mealime, and EatThisMuch showed no single app covered meal planning, personalisation, specific health-diet targeting, and personalised eat-out recommendations together. BiteWise was positioned to be the first to combine all four.

Heuristic teardown of existing apps also directly shaped decisions: Yummly's tab navigation (Jakob's Law) validated a familiar recipe-detail layout, but its unlabelled icon buttons showed the cost of skipping clear affordance. Simple's meal-tracking app demonstrated good card layout and common-region grouping, but its "New Year" premium prompt disguised as a reward showed a dark pattern worth avoiding.

A second, feature-focused survey confirmed budget-friendly options and personalised meal plans as the two most-wanted features, ahead of ingredient swapping, goal tracking, and educational content.

Card sorting was used to structure the bottom navigation and confirm which features belonged under Kitchen, Eat Out, or a catch-all "Others" group, which fed directly into the final information architecture:

Kitchen → Discover → Recipe Generator → Eat Out → Profile/Settings,each with its own sub-flow.

A Hierarchical Task Analysis mapped the core "prep a meal for the week" task end-to-end from opening the app, hitting Prep Meal, selecting recipes per day/meal-time, and confirming, which became the backbone of the meal-planning flow.

6. Designing the Solution

Each core flow moved through sketches → low-fidelity → high-fidelity, refined against real usability data at each stage.

Onboarding — a lightweight sign-up (social or manual) followed by a short quiz on diets, allergies, and excluded ingredients, so meal suggestions are personalised from the first session.

The Nav Bar — informed directly by card sorting, settled into four anchors: Kitchen, Discover, Eat Out, Profile.

The Kitchen — the meal-planning hub. Multiple household members can be added to a shared kitchen, each with their own dietary profile, so a meal plan accounts for everyone at once (built for personas like Michelle and Anna). Includes achievement badges as a light gamification layer.

Adding a Member — designed around an iOS AppClip share flow, so inviting a family member or housemate takes one tap rather than a full account setup.

Recipes & Cooking — every recipe shows ingredients, swap options, and a step-by-step cooking mode.

Recipe Videos (accessibility feature) — built specifically for Thomas's persona. The British Dyslexia Association notes difficulties with language, mental calculation, motor skills, concentration, and organisation as common dyslexia barriers — so recipes offer a guided video mode with large, minimal on-screen text as an alternative to reading a full ingredient list.

Eat Out — a restaurant navigator filterable by diet, with both list and map views, addressing Olivia and Stephen's need to make fast, safe decisions when eating away from home.

Profile/Settings — general info, diet and allergy preferences, excluded ingredients, symptom logging, and household member management, all in one place.

7. Testing & Iteration

Guerilla testing — 17 people gave quick feedback on early low- and high-fidelity prototypes, focused on core features, screen consistency, and navigation flow. It surfaced three concrete fixes:

• Meal planning: users had to pick a date on one screen, then select meals on another — they'd forget which date they'd chosen. Fixed by merging date selection and meal selection into a single screen (aligned with Nielsen's recognition rather than recall heuristic).
• Eat Out navigator: users preferred browsing a list of restaurants over a map-first view. Fixed by defaulting to diet-filtered results with a map/list toggle.
• Settings: related settings were scattered and text-heavy, which was flagged as unfriendly for dyslexic users. Fixed by grouping related settings using the Gestalt Law of Proximity and adding icons to speed up scanning.

Moderated usability testing: 5 users, face-to-face, given 6 timed tasks with a 10-minute cap, followed by SUS scoring.

Task 1 (creating an account) took the longest on average (61.3s) and had the most clicks: a natural next candidate for a smoother onboarding flow.

Task 3 (adding a member via AppClip) was the most consistent across users (lowest coefficient of variation), validating the AppClip approach.

Task 5 (checking a diet tag) had the highest variation and one outright failure, tracing back to the same "tag isn't obviously tappable" issue found in Guerilla testing.

SUS score: 76.5 was above the 68 "average" baseline, placing BiteWise at the 73rd percentile of the SUS database (better than 73% of scored products), with minimal variation across the five testers (coefficient of variation 0.08).

A further heuristic pass caught two more issues worth noting: a "Done" button breaking top-to-bottom visual flow (fixed by repositioning it in line with the natural reading order), and a UX copy issue, the phrase "Go to Kitchen" reading as derogatory in some Middle Eastern cultural contexts, a reminder that microcopy needs checking against more than one cultural lens.

8. Final Prototype

The finished flow ties everything together: onboarding, weekly meal planning, adding household members via AppClip, guided recipe cooking, diet-filtered restaurant discovery, a Discover feed of community recipes and articles, and full profile/settings control, plus platform-native touches like Live Activity meal timers and lock-screen widgets that surface the app without opening it.

Link to prototype here ↗

9. Reflection

Key takeaways: Designing a health app isn't linear. New research kept surfacing well into later stages of the project, and continuous feedback shaped the design more than any single upfront research phase. Involving real users throughout (not just at final testing) consistently caught issues internal review missed. And task delegation across a five-person team mattered as much as any individual design decision.

Obstacles: landing on a clear target audience, cutting features to fit the timeline, structuring the nav bar, deciding the overall app flow, and applying UX heuristics and accessibility requirements from day one rather than retrofitting them.

Where this goes next: an onboarding tutorial, support for cooking at larger household scale, health-tracking features (BMI, weight), a budget tracker, gesture-based hands-free cooking mode, and maybe more importantly, a clear, dedicated statement that the app does not promote or endorse eating disorders, with a way for users to report harmful content