Foodnoms
Stay on track, even when dining out.
Industry
Health
Enviroment
Mobile
An ever so familiar problem
Have you ever sat in a restaurant, staring at a menu, wondering how it fits into your macros? I have, many times. Tracking your nutrition is easy when you’re cooking at home. You can easily measure ingredients, log meals, stay in control.
Eating out is different, and gets complicated when menus don’t show macros, leaving you guessing or giving up because tracking feels impossible. I see this struggle firsthand, dedicated people wanting to stay on track but forced to make blind decisions without clear nutrition info.
Research backs this up. In a study on dining-out habits and nutritional awareness, over 68% of respondents admitted they struggled to estimate calorie counts accurately, often underestimating by as much as 30%.
% of people who struggle with macros dining out
68%
Source: The Times
Another insight that stuck with me? Decision fatigue. The more choices a person has, the harder it is to make a confident decision, especially when nutritional uncertainty is in play.
A smarter way to track nutrition
FoodNoms has always stood out to me because it does nutrition tracking differently. Where other apps feel cluttered, slow, and packed with ads, FoodNoms is clean, privacy-conscious, deeply integrated with Apple’s ecosystem, and automate food tracking through small design choices that make the task feel less like a chore.
But even the best nutrition trackers hit a wall when it comes to restaurants. Unless a restaurant provides nutritional information (which most don’t), users are left estimating. And if they’re serious about their fitness goals, “estimating” isn’t good enough.
I wanted to create a feature that feels as intuitive as the rest of FoodNoms, and the first design decision was input. Should users manually type in menu items? Maybe not, because that might feel like too much work, especially when dining out is meant to be social and enjoyable. Instead, I leaned on computer vision.
AI processes the menu, using OCR (Optical Character Recognition) to extract text, then applies natural language processing to recognise dish names and ingredients. It cross-references this data with a nutrition database (like USDA and FoodNoms' own dataset) to estimate macros, using machine learning to refine accuracy over time.
5 parties collaborate to analyse and present nutritional information
A key finding from my research into AI-assisted food tracking was that users are more likely to trust an AI-generated macro breakdown if they can verify and edit it. This shaped the next part of the experience: instead of assuming the AI would always get it right, I designed confirmation functions where users can adjust portion sizes or correct any mismatches. Providing a balance between automation and control.
Next came accuracy. Not all menus describe their food in detail, which meant AI needed to do more than just read text. It had to interpret it. For example, if a menu lists “Grilled Chicken Salad,” it should recognise that as a high-protein, low-carb option, even if the menu doesn’t explicitly mention macros.
This requires a smart database matching system, where FoodNoms cross-references menu items with existing nutritional data, adjusting estimates based on portion sizes and cooking methods.
Research into food tracking adherence also influenced my approach. Studies show that users are more likely to consistently track meals if they can do so within 10-15 seconds. This meant designing a minimal but powerful interface. A simple camera overlay for capturing menus, a clear and digestible (pun intended) macro breakdown, and an effortless way to log meals.
Designed to guide, not dictate
Another challenge I had to consider was decision-making in real time. Beyond tracking their meals, people want guidance on what to order. This led to the creation of “Suggested Picks”, a subtle feature that highlights meals aligning with a user’s preset dietary preferences, without forcing a decision. So the app helps you make the best choices clear at a glance.
Feedback and learnings
Most homes are dynamic spaces with a collection of artefacts that require care. The app seamlessly integrates maintenance tracking, with timely reminders for servicing appliances, replacing filters, or checking on valuable electronics. These reminders offer helpful insights, some with links to verified DIY practices.
This is another area where automation makes a difference. Instead of users having to remember when their washing machine needs servicing or when to change the air filters, the app predicts maintenance schedules based on manufacturer recommendations and usage patterns.
Back to top
All content, designs, and case studies on this website are my intellectual property. Unauthorized use, reproduction, or distribution of any material is strictly prohibited. For permissions, contact me at ali.adebolafuad@gmail.com
Last updated December 2024.