A Food/Meal Recommendation System Built with Sahha
This brief guide aims at breaking down how health data from Sahha can be used to tailor food/meal recommendations for your users. Rather than offering static meal plans or generic food choices, this system adapts meal options based on trends in Sahha Archetypes, Scores and Biomarkers, ensuring maximum personalization to your users.
Example meal/food recommendations based on Sahha Scores & Biomarkers
- A user with low activity, poor sleep, and poor readiness → Prioritize easy-to-digest, energy-stabilizing meals with anti-inflammatory and gut-friendly ingredients.
- A user with high activity and high readiness → Performance-boosting meals with strategic carb-loading and protein-packed options to maximize output.
- A user with good wellness but fair sleep quality → Adjusted meal composition to support melatonin production, gut microbiome stability, and evening relaxation.
Example meal recommendations based on Sahha Scores and time-of-day
- Morning Meals → Adjusted based on readiness and sleep scores to either energize or stabilize blood sugar.
- Pre/Post-Workout Meals → Adapted based on activity scores and archetypes to optimize performance and recovery.
- Evening Meals → Tailored for recovery, sleep enhancement, and inflammation control, considering wellness and sleep quality scores.
In this tutorial, we are going to break down two approaches to building a meal/food recommendation system with Sahha.
- Broad Category Recommendations: Great for meal and food ordering applications.
- Granular Product Recommendations: Great for hyper-personalised nutrition applications
Let's dive in.
Broad Category Recommendations
Purpose: Designed for platforms with a fixed menu or limited ingredient tracking.
How it works: Matches meals to general Sahha Archetypes without needing precise nutrient tracking.
Implementation Strategy:
- Align meal categories with data from Sahha. For example:
- If Sleep Efficiency is poor, suggest “sleep-supportive” meal categories like whole-food, balanced dinners instead of specific magnesium-rich meals.
- If Activity Level is high, suggest “high-protein recovery meals” rather than tracking specific amino acids.
- Use predefined meal / food tags, for example:
- “anti-inflammatory,”
- “high-energy,”
- “light & digestible”
- Deliver recommendations for meals/food based on tag classification.
Archetype Used | Archetype State | Recommended Meal/Food Category |
|---|---|---|
activity_level | Sedentary | Low-calorie, high-fiber meals to support metabolic health |
activity_level | Highly Active | Performance-oriented meals with complex carbs, electrolytes, and lean protein |
overall_wellness | Poor Wellness | Anti-inflammatory meals rich in antioxidants, omega-3s, and gut-supporting ingredients |
overall_wellness | Optimal Wellness | Maintenance-focused diet with nutrient cycling for longevity |
sleep_quality | Poor Sleep Quality | Light, nutrient-dense meals rich in magnesium, glycine, and sleep-supportive amino acids |
sleep_quality | Optimal Sleep Quality | Meals supporting circadian rhythm alignment with nutrient timing |
readiness_score | Low Readiness | Easy-to-digest, high-energy meals with adaptogens and B vitamins |
readiness_score | High Readiness | Performance and endurance meals with strategic carb loading |
Granular Product Recommendations
Purpose: Used when the menu allows for detailed ingredient tracking and dynamic customization.
How it works: Matches specific meal items based on detailed nutrient needs from health scores.
✅ Implementation Strategy:
- Direct nutrient-focused recommendations based on health indicators.
- Ingredient-level tracking if supported by the meal provider.
- AI-powered dynamic meal suggestions adjusting in real time.
Example Logic for Granular Recommendations:
Meal/Food Category | Target Demographic | Associated Sahha Scores & Archetypes | Specific Meal Adjustments Implemented |
|---|---|---|---|
High-Performance Meals | Athletes, highly active users, individuals with high readiness | High Activity Score, High Readiness Score, Moderately Active to Highly Active Archetype | Increased complex carbohydrates for sustained energy, strategic protein intake, electrolyte balance, performance-enhancing micronutrients (e.g., magnesium, B vitamins, omega-3s) |
Metabolic Support Meals | Users with weight management goals, glucose regulation needs | Fair to Poor Readiness Score, Sedentary to Lightly Active Archetype | Balanced proteins, fiber, and healthy fats; reduced fast-absorbing carbs; incorporation of metabolic boosters like cinnamon, ginger, and chromium-rich foods |
Recovery & Sleep-Supportive Meals | Users with poor sleep, muscle fatigue, or high inflammation | Poor Sleep Score, Fair Readiness, Poor Sleep Quality Archetype | Magnesium-rich foods, glycine (collagen sources), tart cherries for melatonin support, light digestion-focused dinner meals (low inflammatory response) |
Gut-Optimized Meals | Users with digestive concerns, bloating, or fluctuating energy | Poor to Fair Wellness Score, Sedentary to Moderately Active Archetype | Probiotics (fermented foods), prebiotics (fiber-rich vegetables), anti-inflammatory ingredients (turmeric, ginger), and gut-sealing nutrients (zinc, glutamine) |
Energy-Sustaining Meals | Sedentary to moderately active users needing balanced energy | Fair to Good Activity Score, Sedentary to Lightly Active Archetype | Slow-digesting carbs, moderate protein, stable fat intake to prevent energy dips, focus on sustained satiety |
Anti-Inflammatory Meals | Users experiencing fatigue, poor wellness, or muscle soreness | Low Readiness Score, Poor Wellness Score, Poor to Fair Sleep Quality Archetype | Omega-3-rich meals, polyphenol-dense foods (berries, dark chocolate, green tea), whole food-based with minimal processed ingredients |
What else to consider?
Experimentation & A/B Testing
It's important to experiment and test different approaches so you can identify the best approaches to maximising engagement and retention of your implementation.
- Test different recommendation placements (homepage, checkout upsells, notifications).
- Control vs. Personalized Group: One group gets traditional recommendations, the other gets Sahha-driven health-based suggestions.
- Segment users based on engagement levels to see if certain groups (e.g., health-conscious consumers) respond better.
Key Metrics & ROI Measurement
Setting up some basic metric tracking to ensure you are increasing KPIs that matter to your application.
✅ Conversion Rate Lift → Do personalized health-based recommendations increase meal/food purchases?
✅ Average Order Value (AOV) Impact → Do people buy/consumer more when given lifestyle-based suggestions?
✅ Repeat Purchase Rate → Does health/lifestyle-based personalization improve long-term retention?
Next Steps for Implementation
- Integrate Sahha with a meal/food app for real-time user health & lifestyle analysis.
- Classify meals/food to user Sahha Data (Archetypes, Scores and Biomarkers) where applicable.
- Deliver timely recommendations as each user's Sahha Data changes.
- Gamify engagement by linking consistent meal/food choices to Sahha Data .