Experiment Design

The experiment feature in this app lets you do something powerful: test whether a specific change — eating differently, timing a walk differently, trying a new food — actually makes a measurable difference to your glucose. This is called an A/B test or N-of-1 trial, and it is exactly how scientists test interventions, applied to your daily life.

How It Works

Every experiment has two conditions:

Control: Your normal behaviour. No changes. This establishes your baseline response under your usual conditions.

Test: The same situation, but with one variable changed. This is what you are testing.

You run both conditions, compare the results, and the difference tells you whether the change worked for you.

Why you need both: Your glucose response is not fixed. The same meal on Monday might produce a different reading than the identical meal on Wednesday, depending on your sleep, stress, and activity. If you only test the new food without a control measurement, you cannot tell whether any difference was caused by the change you made or by background variation in your day.

Example experiment:

• Control: Lunch with 1 cup of rice and ½ cup of dal, eating rice first.
• Test: Lunch with 1 cup of rice and ½ cup of dal, eating dal and sabzi first, rice last.
• Result: If the spike is consistently lower in the Test condition across 3 repetitions, the food order made a real difference for you.

Your Target

A difference of 10 mg/dL or more in the post-meal spike between conditions is considered meaningful (given normal day-to-day variation of ±5–8 mg/dL). A consistent difference across 2–3 repetitions is reliable.

Why This Matters

Population research tells you what works on average across thousands of people. Your body may respond differently. Experiment design lets you find out what is actually true for you — which specific foods spike your glucose the most, which interventions bring the biggest reduction, and what you can safely eat at what portions.

What You Can Do

• Keep everything else the same between Control and Test: same time of day, same approximate activity level beforehand, same sleep quality. The goal is to isolate the one variable you are testing.
• Change only one thing at a time. If you swap rice for millet AND add a walk after the meal, you will not know which factor drove any difference.
• Repeat each condition at least 2–3 times before drawing conclusions. One data point can be an outlier; three consistent ones are reliable.
• Use the app's experiment feature — it is designed to prompt you for readings at the right intervals and display the paired comparison automatically.

Based on: Guyatt G et al., NEJM 1986 (N-of-1 trials); Nikita J. Bhatt et al., personalized nutrition research; Zmora N et al., Cell 2015