Food Consistency Testing

Your glucose response to the same meal is not identical every time you eat it. Your body's reaction varies from day to day based on how well you slept, how much stress you are under, how active you were, and even your gut microbiome. Understanding this variability helps you interpret your data correctly and know how many repetitions you need before drawing reliable conclusions.

How It Works

Glucose response variability is measured using the Coefficient of Variation (CV) across repeated tests of the same meal:

CV = (Standard Deviation of spikes ÷ Average spike) × 100%

For example: if you test the same breakfast three times and get spikes of 35, 42, and 38 mg/dL:

• Average spike = (35 + 42 + 38) ÷ 3 = 38.3 mg/dL
• Standard deviation = approximately 3.5 mg/dL
• CV = (3.5 ÷ 38.3) × 100 = 9.1% — very consistent

Research shows that even in people without diabetes, glucose response to the same food varies by approximately ±10–20% from day to day. This is normal biological variation, not measurement error.

Consistency categories:

• CV below 15%: Consistent response. Your body reacts predictably to this meal.
• CV 15–25%: Somewhat variable. Day-to-day factors are meaningfully affecting your response.
• CV above 25%: Highly variable. Background factors (sleep, stress, activity) are dominating the response more than the food itself. More repetitions and consistency in conditions are needed before comparing meals.

Your Target

For reliable comparisons between meals (such as rice dosa vs. millet dosa), aim for at least 3 repetitions of each. If your CV for a particular meal is very high, you need more repetitions and tighter control of background conditions (same sleep, same activity) to get a meaningful comparison.

Why This Matters

A single reading after a new food tells you almost nothing reliable. Two readings give you a rough estimate. Three or more readings with similar conditions give you something you can trust. Knowing this prevents you from drawing the wrong conclusion — for example, deciding that a certain food "spikes me badly" based on one unusually high reading on a night when you also slept poorly.

What You Can Do

• When testing any food or meal change in the app, plan to repeat the same test at least 3 times before comparing it to a control.
• Take notes on the context of each test: did you sleep well? Were you particularly active or sedentary? This helps you understand why readings vary.
• If you get one very different reading in a series — say, 68 mg/dL when the others were around 40 — look at the notes for that day. There is usually a clear reason.
• The app's experiment feature shows you the average and range across multiple trials, not just individual readings, which gives you a more reliable comparison.

Based on: Zeevi D et al., Cell 2015 (personalized nutrition); Rodbard D., Diabetes Technology 2011; standard N-of-1 methodology