What is Zoe Health app?

Zoe Health is a personalized nutrition app that helps users understand how their body reacts to food by analyzing their blood sugar levels. Zoe uses a continuous glucose monitor (CGM) to track blood sugar response to food in real-time.

The program starts with an at-home test that analyzes the user’s gut microbiome, blood sugar, and blood fat. Based on this analysis and CGM data, the app provides personalized nutrition suggestions.

Rise of CGM trackers

CGM devices like Dexcom and Freestyle Libre offer continuous blood glucose readings, helping users to see how different foods impact their glucose levels in real time. By logging meals, users can see which foods spike their glucose and which keep levels stable.

For example, an app can recommend meals based on the user’s glucose response or suggest optimal timing for exercise based on glucose trends.

In the case of Zoe Health, the product is limited to food tracking and meal management and doesn’t include the impact of exercise or any other factors on glucose.

Efficacy

Designing products using continuous glucose monitoring is a trend that has been on the rise for a while. Multiple research studies are proving that using CGM along with activity tracking in health apps can lead to improvements: building new healthy eating habits, and increased intake of protein, fiber, and healthy fats.

At the same time, some claim that common sense advice is just as useful as personalized recommendations from the app:

The problem with personalised nutrition industry is that this is still a young research field, and there is not yet good enough evidence across the field to believe that we have yet found worthwhile novel interventions that are more helpful than standard advice. (source)

In this article, we will focus on the design and learnings that can be derived from such products.

Meal logging

Self-monitoring: psychology behind it 🔎

Studies have shown that simply tracking food intake increases awareness of eating habits. This awareness is a crucial first step for positive change.

Here’s how self-monitoring works:

• Increased awareness. By logging meals, users pay closer attention to what they eat and how much. This process disrupts automatic eating patterns and brings mindful attention to food choices.

• Identifying triggers. Through self-monitoring, users can identify situations that trigger unhealthy eating. This self-awareness helps them to choose healthier options anyway.

• Progress tracking. Tracking food intake allows users to set goals for portions, nutrients, or dietary changes. Seeing progress through data visualization in the app motivates them to stay on track.

This research found that self-monitoring weight led to greater weight loss compared to those who didn’t track their weight.

Mechanics in the app

The Fogg Behavior Model says that behavior change requires three elements:

1. Motivation. Willingness to change the behavior (e.g., wanting to improve blood sugar control).

2. Ability. The ease with which the behavior can be performed (e.g., a user-friendly meal logging).

3. Prompt. A trigger that initiates the behavior (e.g., a notification reminding users to log their meals).

Zoe offers several easy and quick ways to log the meal and see the result.

Similar to all popular applications, Zoe users can log their meals either using barcode scanning, selecting a meal from the database, or generating a recipe based on the name.

While all other methods are common, AI-based recipe generation is something that is still quite uncommon and is not yet widely used.

Let’s dive into the implementation of this feature released by Zoe:

• based only on the title of the meal, the app generates a list of ingredients of the meal,

• user can add or remove something from the list,

• user sets a portion that they consumed.

Out of dozens of meals I have tested, I was pleasantly surprised that the majority of the recipes were pretty accurate and adjusting the ingredients was not a burden. It’s definitely something that should be explored more by metabolic health products.

Meal logging using photos is not available, possibly due to the complexity of defining the meal ingredients correctly.

AI-powered meal generation

Meal analysis

The feedback loops: psychology behind it 🔎

Personalized feedback in the Zoe app acts as a continuous loop:

1. Action. Users log meals and make dietary changes.

2. Feedback. The app provides personalized feedback on meals and overall progress through scores and data visualization.

3. Evaluation. Users evaluate their progress.

4. Adjustment. Based on this feedback, users can adjust their diet.

Tracking meals is one of the core components of the user journey. How did Zoe make this flow easy and the results actionable?

Zoe Health assesses each meal you have consumed based on 3 pillars:

• blood sugar (based on the average and personal glucose spikes)

• blood fat (general knowledge, no personalization)

• gut health (general knowledge, no personalization).

They present the data with low detalization, resorting to “bad/good” and giving no data behind each metric. It may be a good way to simplify the process for the user and not overload them with a lot of information yet at the same time it reduces the trust in the system.

Zoe Health also uses a simple explanation to form relations with the specific food: enjoy with no limits/regularly/rarely.

Grading food is not a novel concept, yet Zoe aside from grading suggests alternative meals with “better” grades.

Zoe has an interesting approach to analyzing the meal based on a few parameters. To satisfy the needs of highly engaged users willing to dive into the details of each meal, they share a granular breakdown of each food including the amount of sugar, protein, and net carbs:

The status of each metric is color-coded for quick scanning, for example, “low sugar” is marked with a green dot, and “very low fiber” is red.

Daily review

Goal setting theory: psychology behind it 🔎

Goal setting theory, developed by Edwin Locke and Gary Latham, says that specific, challenging, and measurable goals lead to better chances of success. Here’s how Zoe uses these principles:

1. Specificity. The app doesn’t just promote generic healthy eating. Feedback on meals and specific scores allow users to see food's impact on blood sugar spikes.

2. Measurability: The app’s data visualization and progress tracking make it easy for users to track their progress toward goals, keeping them motivated and engaged.

On a day-to-day basis, users get an assessment of their current food intake in a score format as well as a separate assessment of each food they intake.

Weekly reporting

Aside from granular data on the specific meal or day food intake, Zoe dives into analysis of the parameters such as:

• low-quality fats;

• plant diversity;

• fiber;

• food intake window (hours);

• food intake.

Chatbot support

The app offers recipe generation based on user preferences.

Is a chatbot the best medium for this? A chatbot ticks many points: it can create a custom recipe based on the requirements, it has a human-friendly interface, and it allows exporting and saving the recipe in the app.

Typing out preferences in a chatbot is slower than clicking pre-defined options, also chatbots require back-and-forth conversation, it might be faster to choose options from a menu. However, the app doesn’t have its own advanced recipes section at the moment, and having at least a chatbot to help with recipe generation is useful. Previously, we discussed weight management apps and their recipe navigation patterns.

Meal swap recommendations

Scoring a meal is useful, but is it actionable? Knowing that some meal is not the most appropriate does not directly translate into the knowledge of how to swap it for something similar but healthier.

Here “Ingredient tweaks” come into play. Adding some ingredients can help lower the time of glucose absorption in the bloodstream positively affecting the health — and the app proactively suggests such tweaks.

Behavioral nudges: psychology behind it 🔎

Features like “ingredient tweaks” and “better” meal suggestions nudge users towards healthier choices. Here’s how nudges work in personalized nutrition apps:

1. Visual design. Nudges influence how information is presented, making healthier options more attractive. For instance, highlighting “better” meal suggestions or using green colors for healthy scores primes users to choose these options.

2. Framing. The way information is framed can significantly impact choices. “Ingredient tweaks” are small, manageable changes, not daunting to adopt.

3. Loss Aversion. People are more likely to avoid losses than pursue gains. Showing potential negative consequences of unhealthy choices (e.g., blood sugar spikes) can nudge users towards healthier options.

Aside from that, some products may be a bit healthier than others and the product suggests to check them out and make a change next time.

User education

User education regarding glucose spikes and management in the app is mostly contextual. As they log the meals, they receive valuable insights about them. Aside from that, the app introduced a section with useful information about foods.

Aside from more traditional formats, Zoe also used quick learning patterns such as short stories played automatically, and clips from the longer videos giving users to watch the whole episode if needed.

Takeaways

• Provide regular recommendations based on both individual meals and long-term trends (weekly/monthly).

• Consider offering various methods to log meals like barcode scanning, database item selection, and AI-powered recipe generation based on meal titles.

• While data analysis is important, prioritize actionable insights. Don’t just tell users a meal is “bad”, suggest ways to improve it (“ingredient tweaks”) or offer healthier alternatives.

• Present complex data clearly and concisely (meal scores). However, avoid oversimplification as it can erode user trust.

• Provide educational content: use short stories and video snippets to educate users in an engaging way.

Good luck! ☘️