Write down what you ate, not what was put in front of you. For each meal, note roughly how much was served or taken, then how much was left, or whether you went back for more. When the dish was shared, write your own share in plain terms, such as “about a third of the pizza”, and mark it as an estimate. A rough, honest amount is more useful than a precise-looking one you cannot stand behind.
Plates rarely empty exactly, and meals are often shared. This guide is about keeping the record close to what happened. It does not suggest that finishing or not finishing a meal means anything for your digestion, and a diary cannot tell you that.
Why “served” and “eaten” need separate notes
Nutrition surveys have dealt with this problem for a long time. The UK diary booklet used in a government survey of infants and young children gave parents separate columns for the quantity served and the quantity left over. Its example was mashed banana: 2 tablespoons served and 1 teaspoon left. If nothing was left, parents were asked to write “none”, so that a blank could not be mistaken for a forgotten entry. It was designed for parents recording a young child’s food for a survey, so a personal symptom diary needs far less. The idea is still useful: a blank and a “none” mean different things.
The MRC Epidemiology Unit’s measurement toolkit describes weighed food diaries in a similar way, with each entry including the food, the weight served and any leftovers. It also says that weighed records place the biggest burden on the person keeping them, and that people may change what they eat to make recording easier. That is a reason to keep your own method light.
A short leftovers note
For most meals, one extra phrase is enough:
- “Pasta bake, medium bowl, ate most of it, left a few forkfuls.”
- “Soup, large mug, finished.”
- “Takeaway curry, about half the container, rest saved for tomorrow.”
- “Sandwich, ate all but the crusts.”
Pick a small set of words and use them the same way each time, for example all, most, about half, a little, none. Words are easier to keep up than percentages, and they do not claim more accuracy than you have. If you want to be more specific for a meal you are curious about, a fraction (“about a third left”) works.
If a meal was eaten over more than one sitting, record each sitting with its time, as you would for snacks and nibbling. The time of eating matters for comparing meals with symptoms, and an unfinished plate eaten again an hour later is two separate eating times.
Seconds and extras
Going back for more is easy to forget because it feels like part of the same meal. A simple habit is to add a line when you do it: “Seconds: another half portion of rice, 19:25.” You do not need to total anything. A later reader, including you, can see that the meal was larger than the first helping suggested.
Small extras are also easy to miss. A 2022 systematic review in Advances in Nutrition compared self-reported intake with observed intake in 29 studies of 2,964 healthy adults across 15 countries. It found that omitted foods and misjudged portion sizes were among the most common sources of error. Vegetables (omitted in 2–85% of cases across studies) and condiments (1–80%) were left out most often, and portion sizes were both over- and under-estimated. The studies used different methods and measures, so the figures show a wide range of experience and not a rate you should expect yourself. They do suggest that sauces, sides and extras are worth a quick word.
Shared plates and family-style meals
When you take food from a common dish, you usually cannot see how much you ate in the way you can with a single plate. The MRC toolkit notes that a shared meal makes it challenging to complete weighed food records accurately,, and none of the sources I found sets a single rule for symptom diaries. The suggestions below are a practical approach and not an evidence-based standard:
- Name the dish and who shared it. “Shared pizza with two others” is enough. Names are not needed.
- Estimate your own share in plain terms. “About three slices of eight”, “roughly a quarter of the rice”, “two spoonfuls from the middle of the table”.
- Say it is an estimate. Write “approx.” or a question mark. The earlier guide on entries made at the time or added afterwards explains why marking a guess helps you read the entry fairly later.
- Note what you could not see. At a restaurant or someone else’s home you may not know the ingredients. Write what you do know, as described in the guide to recording a mixed meal, and mark the rest as unknown.
For a dish you cooked yourself, you can record the recipe once and then only the share you ate. The government survey diary above asked parents to record the whole recipe, the number of people it served, and then the amount served and left, which is more than most people need. Writing “stew, makes four, I had one portion” is a lighter version of the same idea.
Photographs: helpful, but not a measuring device
A photograph before and after eating can show what was served and what was left. Research on this comes from settings with trained raters and supervised meals. In a 2025 pilot study in the Journal of Human Nutrition and Dietetics, researchers compared before-and-after photographs with weighed plate waste for 123 lunch observations from 41 preschool children in Texas. Trained raters estimated the percentage eaten in 10% steps, and the results agreed well with weighing on average. However, the meals were pre-portioned, the sample was small, and the raters were trained, so the study does not show how well adults can judge their own plates at home or from a shared dish.
A laboratory study of 30 mock school salads, also from a research setting, found that photograph-based estimates of what was left on the plate agreed more closely with weighed amounts than estimates of the starting portion did. Raters there judged simulated salads, and the study did not involve real diners, so treat it as a hint and not as a rule. It suggests that if you take only one photograph, a “what is left” picture may be informative. The guide to meal photographs in a diary covers what a photograph can and cannot show.
Reading the entries later
When you look back over a few days, a few plain questions keep the amounts in perspective:
- Was this meal larger or smaller than my usual version, or did I just record it differently?
- Did I note leftovers and seconds as carefully on difficult days as on easy ones?
- Is a blank a meal I did not eat, or one I did not write down?
Amounts recorded as rough words can show whether meals were broadly similar, but they cannot establish that the amount eaten caused a symptom. The guide to telling which meal caused bloating explains why a meal that came before a symptom is an observation and not a cause, and the guide to missing symptom check-ins shows how gaps change what a diary can show.
If you are eating much less than usual because you are worried about symptoms, if you are losing weight without trying, or if symptoms are persistent, getting worse or worrying you, speak to a GP, pharmacist or NHS 111. A diary is not a substitute for care.
If you would like a lower-effort way to capture meals as they happen, you can get GutBloom on Google Play.
Sources
- UK government (MRC Human Nutrition Research) — Diet and Nutrition Survey of Infants and Young Children, Appendix E: Diary documents
- MRC Epidemiology Unit Measurement Toolkit — Weighed food diaries
- Advances in Nutrition (2022) — Contributors to misestimation of food and beverage intake in short-term self-report dietary assessment, adults
- Journal of Human Nutrition and Dietetics (2025) — Validity of weighed and digital photography diet estimation in a childcare setting
- Journal of the Academy of Nutrition and Dietetics (2018) — Digital imagery for starting portions and plate waste from school salad bars