What seven days of meals can reveal

Have you ever looked back at a day of eating and thought, “I should have done better”? Food tracking can easily feel like a report card, especially when blood sugar numbers join the marking.

But a seven-day food record is most useful when it is not used to judge you. Its real value is helping you notice one repeatable connection: which ordinary meals, portions, timings, and circumstances tend to leave your blood sugar steadier—or less steady. A week is long enough to reveal clues that one meal cannot, yet short enough to try without turning your life into a permanent logging project. Think of it as a small investigation, not a test you can pass or fail.

For seven days, write down what you eat and drink, roughly when, and—if you already check glucose—any readings you would normally take. You do not need calorie calculations, perfect measurements, photographs, or a paid app. If detailed tracking makes you anxious or preoccupied, keep it minimal or stop; the record is meant to support you, not take over your day.

Descriptions such as “one bowl of rice with beans and vegetables” or “two slices of bread with eggs” are useful. Add one short note when life affected the meal: “rushed,” “very hungry,” “walked afterwards,” or “poor sleep.” These details stop food from taking all the blame. Stress, activity, illness, sleep, and medication timing can also affect readings.

Do not start by labelling foods good or bad. At the end of the week, scan for the same situation appearing twice or more. Perhaps bread alone at breakfast leaves you hungry by mid-morning, while bread with eggs or yoghurt carries you longer. Maybe a large serving of rice is followed by a higher reading than a smaller serving eaten with beans, chicken, fish, or vegetables.

You might instead notice that the meal itself seems fine, but eating very late after skipping lunch leads to a larger portion than you intended. The lesson is not “never eat bread” or “rice is bad.” It is that a repeated situation gives you somewhere practical to experiment.

One surprising reading proves very little. Even similar meals can produce different numbers on different days. If Tuesday’s potato meal was followed by a high reading but Sunday’s was not, look at portion, timing, activity, sleep, stress, and whether the readings were taken at comparable times. A single result is a clue, not a verdict. A pattern across several similar situations is more useful—and still does not mean you failed.

Then choose only one gentle experiment for the following week. Keep the food familiar and affordable: slightly reduce the starch portion and add more beans or vegetables, pair fruit with plain yoghurt or nuts if available, or avoid arriving at dinner extremely hungry by having a simple earlier meal. Change one thing, not five, so you can tell whether it helped. The best experiment is one you can repeat in your actual life, not one designed for a perfect week.

Your record may show no clear food pattern. That is useful information, not wasted effort. Blood sugar can shift because of illness, hormones, medication, or other factors. If readings are repeatedly outside the range your care team has set, or you are having frequent lows or concerning symptoms, share the record with them. It can make the conversation more specific. Do not change medication based on a food diary alone.

Seven days of tracking cannot tell you everything about your diabetes, but it can replace a vague sense of “I’m doing it wrong” with one useful clue. This week, record your meals simply and circle just one pattern that appears more than once. You are not collecting evidence against yourself. You are learning what your everyday life is already trying to tell you.

Until next week,
Michael

PS: This is exactly the kind of problem we’re building Cocoa to help with: noticing patterns around familiar meals without turning food into a score. If you’d like to test our diabetes food companion app, reply to this email. Testing spots are limited while our small team learns closely from each person.