Why calorie tracking apps fail on Indian food

Roughly half the people who start a calorie-tracking app stop within a month. The usual explanation is motivation. Having watched a lot of Indian users try and fail, the more honest explanation is that the tool was designed for a different kind of dinner.

Four assumptions sit underneath almost every popular tracker. All four hold reasonably well for a grilled chicken breast with rice and broccoli. All four break on a thali.

Assumption 1: food arrives in separable components

Western tracking apps are built around the idea that a meal is a list. Protein, starch, vegetable — three entries, three portions, done in thirty seconds.

An Indian dinner is not a list. It is three rotis, a katori of dal, some sabzi, a spoon of achar, a bit of curd and some rice, arranged so that everything is eaten with everything else. Logging it as separate entries is technically possible and practically exhausting.

Three minutes does not sound like much. Multiply by three meals a day for thirty days and it is four and a half hours of typing about food. That is where the drop-off comes from — not from a lack of willpower, but from an ongoing cost that quietly exceeds the perceived benefit.

Assumption 2: portions are measured

Tracking apps ask for grams, cups or ounces. Indian home cooking uses none of those. It uses the katori, the karchhi, the size of roti your family happens to make, and “thoda sa”.

These are not vague — a household's katori is remarkably consistent — but they are not in the app's vocabulary. So the user is asked to convert an unmeasured serving into grams, in their head, several times a meal, and the resulting numbers carry an error the user knows is there. Data you do not trust is data you stop entering.

Assumption 3: the database is trustworthy

Most large trackers rely on crowd-sourced food entries. For packaged Western food this works acceptably — a branded cereal has one barcode and thousands of people scanning it converge on the right answer.

Search “dal” in one of those databases and you get dozens of entries uploaded by strangers, ranging from 90 to 400 calories per serving, with no indication of which preparation, which lentil, or how much oil. Search “sabzi” and the results are close to meaningless, because sabzi is a category, not a dish.

A user picking the first result is not tracking their food. They are tracking a stranger's guess about a dish that may not resemble theirs. And after a fortnight of that, the numbers stop feeling like information.

Assumption 4: you cooked it, so you know what is in it

Western tracking assumes the person eating is the person who portioned. In most Indian households, food is cooked for the family in one pot and served by someone else.

You did not measure the oil in the sabzi because you did not make the sabzi. You cannot weigh your share of a dish that was cooked for five. Unlogged cooking oil alone is routinely 200 to 300 calories a day — enough to explain most of the “I am eating in a deficit and nothing is happening” cases.

What actually works

Photograph the plate

The composite-meal problem is a decomposition problem, and decomposition is something a vision model does well and a human doing data entry does badly. One photograph replaces six searches, and the portion estimate — while imperfect — is anchored to something real rather than to a guess about grams.

Calibrate once, then estimate

Measure your household's standard servings a single time, as described above. After that, “one katori of dal” is a real number rather than a guess.

Accept a 10% error and move on

Perfect accuracy is not available and is not required. What produces results is consistent directional tracking over months. A log that is 10% wrong every day still tells you accurately whether Tuesday was heavier than Monday, and that is the information that changes behaviour.

Chasing precision is, ironically, one of the main reasons people quit. The precision is not achievable, the failure to achieve it feels like failure at the task, and they stop.

Log for four weeks, not forever

The purpose of tracking is education, not permanent administration. After a month you will know roughly what your usual meals cost and where your calories actually come from — which for most people is a genuine surprise. At that point you can stop logging everything and keep the knowledge.

What this means in practice

If you have tried calorie tracking and given up, the useful thing to know is that the failure was structural rather than personal. The tool asked you to do six times as much data entry as the person it was designed for, using units your kitchen does not have, against a database that did not contain your food.

Fix those and the habit is straightforward. Start with the basics of a calorie deficit, and get the numbers for the things you eat most — roti, dal and breakfast will cover most of a typical week.

Figures on this page are typical values compiled from public nutrition data and are approximations — real dishes vary with recipe, oil, and portion size. This article is general information, not medical or dietary advice. Talk to a doctor or a registered dietitian about any medical condition or before making significant changes to your diet.