Counting Calories but Not Losing Weight? The Reasons, in Order of Likelihood

1. The log is short, and the research says by how much

Start here; it is where the evidence points. Lichtman and colleagues screened 224 consecutive patients presenting for obesity treatment and picked out ten who reported eating under 1,200 kcal a day and still could not lose weight. For 14 days their expenditure was measured by indirect calorimetry and their real intake worked out from body composition. Expenditure was within 5 percent of predicted — no slow metabolism. Intake was under-reported by 47 percent on average (± 16), activity over-reported by 51 percent (± 75). The diet was not failing; the diary was.

That was a small, selected group. The larger picture comes from doubly labelled water, a method that measures energy expenditure chemically and cannot be fooled by memory. In the OPEN study of 484 adults aged 40 to 69, energy intake on careful 24-hour recalls came in 12 to 14 percent below measured expenditure for men and 16 to 20 percent for women. On food-frequency questionnaires the gap was 31 to 38 percent. These were ordinary people making an honest effort. A 15 percent shortfall on a 1,600 kcal log is 240 kcal a day — a modest deficit, gone.

The packaging is allowed to be wrong too. In the United States, 21 CFR 101.9(g) permits up to 20 percent more calories, sugar, fat or sodium than the label states. The European Commission's tolerance guidance allows ±20 percent on fat, carbohydrate, sugars and protein in the 10–40 g per 100 g band, and the calorie figure on an EU label is computed from those.

None of this is an accusation. Under-reporting in these studies is measurement error, not dishonesty — memory, portion judgement and labels all drift the same way, for dietitians too. The point is only that when a log and a scale disagree, the log is the one to suspect first.

2. The bites that never reach the log

The gap is rarely a forgotten meal. It is the things that do not feel like food. A tablespoon of olive oil in the pan is roughly 120 kcal and invisible on the plate. Mayonnaise in a sandwich, dressing tossed through a salad, the spoonful tasted while cooking, the milk in four coffees, the second glass of wine — each is small, and together they are not.

Weekends do the rest. Five days at 500 kcal under and two days at 800 over is arithmetic: 2,500 down, 1,600 back up, a net weekly deficit of 900 kcal — barely a third of what the weekday log suggests. The weekday log is accurate. It just describes five sevenths of the week.

The fix is a wider net, not more discipline. For two weeks, log drinks, oils, sauces and tastes as their own entries. Most people find the missing few hundred kcal within days.

3. Portions were estimated, not weighed

“A bowl of rice” spans 150 g to 400 g depending on the bowl and the mood. Nuts, cheese, pasta, granola and peanut butter are the classic offenders: dense, small servings, so a guess off by half a handful is off by a hundred calories. You do not need to weigh forever; you need to weigh your five staples once, so that your eye is calibrated. There is a longer walk-through in our guide to counting calories.

A photo helps with half of this. It is good at naming what is on the plate and at catching items you would otherwise forget to type; it cannot see grams, or oil that has soaked in. Where photo logging is reliable and where it is not is written up separately.

4. The target itself was a guess

Your daily target came from a formula, and the formula has an error bar. A 2005 systematic review found Mifflin-St Jeor the most reliable of the common equations — within 10 percent of measured resting metabolic rate for more people than any other — while noting “noteworthy errors” when applied to an individual. Ten percent of a 2,200 kcal maintenance estimate is 220 kcal. If your real maintenance sits at the low end of that band, a planned 300 kcal deficit is 80 kcal in practice, and a bathroom scale cannot show you 80 kcal a day inside a month.

Then there is the activity multiplier, where most of the optimism lives. Three gym sessions a week is not “very active” if the other 165 hours are at a desk. Re-run the calorie calculator with the honest multiplier, then let the deficit calculator turn a rate of loss into a target — it rounds to 10 kcal and refuses to go below your BMR. The error bars are explained on the accuracy page.

If the number came from a calculator months ago and your weight has changed since, it is out of date twice over. How to count calories to lose weight covers how to size the deficit and when to recalculate it.

5. Fat is leaving; water is arriving

Body weight is fat, muscle, bone, food in transit and a great deal of water, and only the first changes slowly. A salty dinner, a new training programme (sore muscles hold fluid while they repair), a carbohydrate-heavy day refilling glycogen, the second half of a menstrual cycle — each can add a kilogram overnight and hide a week of fat loss beneath it.

This is why one weigh-in tells you almost nothing. Weigh every morning under the same conditions, average each week, and compare averages a month apart. A trend flat across four weekly averages is real. A number that jumped on Tuesday is water.

6. After months in a deficit, the body spends less

A smaller body needs fewer calories — expected, and a calculator accounts for it when you enter your new weight. What it does not see is adaptive thermogenesis: expenditure falls further than the weight change alone predicts. The most dramatic measurement is the six-year follow-up of “The Biggest Loser” contestants. At the end of the 30-week competition resting metabolic rate had dropped by 610 kcal a day; six years later, with 41 kg on average regained, it was still 704 kcal a day below baseline, and 499 kcal of that remained after adjusting for body composition.

That was an extreme protocol; the adaptation in an ordinary diet is smaller. But the direction is the same, and it has a quieter partner: non-exercise activity drops. People in a deficit fidget less, take the lift, sit down sooner, without deciding to. It shows up as a maintenance level lower than the one you calculated in month one — which is why step 5 below is to recalculate, not to assume.

7. When it is medical

Last because it is the least common, not because it is unimportant. An under-active thyroid, polycystic ovary syndrome, the menopause transition, and a range of medications — some antidepressants and antipsychotics, corticosteroids, insulin, certain beta-blockers — can lower expenditure or raise appetite enough to stall a deficit that would otherwise work. Worth knowing: in the Lichtman study, expenditure in the “diet-resistant” group was normal even though thyroid medication was common among them. A suspected diagnosis is a reason to see a clinician, not to skip the first six causes.

If you have worked through the protocol below and a genuine deficit is still not moving anything after six weeks, or the honest log shows you eating below your BMR, talk to a doctor or a registered dietitian before cutting further. The same goes if you are pregnant, breastfeeding, managing a medical condition, or notice the numbers starting to feel like the point.

A two-week diagnostic, in order

Change nothing about what you eat for two weeks. The point is to find out which cause is yours; cutting now would blur the answer.

1. Weigh yourself every morning, after the bathroom, before food, same scale, same floor. Write it down and do not react to it.
2. Average each week. Compare week two with week one. Ignore anything shorter than a week.
3. Log everything, by photo where you can: drinks, cooking oil, sauces, tastes, the crusts from someone else's plate. Photograph the nutrition label of anything packaged.
4. Weigh your five most common foods once — rice, oats, cheese, nuts, peanut butter — and compare with what you had been guessing.
5. Re-run the calorie calculator with your current weight and an honest activity level. Set that as your target.
6. Compare the honest log against the recalculated target. If the log is under the target and the weekly average has not moved, cut 100 to 150 kcal a day or add a daily walk. Not both, not more — larger cuts make cause 2 worse.
7. If the honest log is already at or below your BMR, do not cut. That is the clinician conversation.

Where a photo-and-label log helps, and where it does not

Disclosure: Olivka is our product, a calorie tracker that runs inside WhatsApp, Messenger, Telegram and Instagram. It is a free trial followed by a paid subscription, and it is subject to the same portion and hidden-fat limits as every other tool described here.

Causes 1 to 3 are about capture, and capture is where a chat log earns its keep. Photographing the plate takes two seconds — the difference between logging the sauce and “I'll add it later”. A photo of the nutrition label reads the manufacturer's own numbers rather than a database guess. Ingredients are matched to USDA FoodData Central or Open Food Facts, you get an itemised card with Confirm and Edit, and the daily total runs against the calorie and macro targets you set.

The correction flow matters more here than it sounds. When you remember the mayonnaise an hour later and say “there was mayo in it too”, the earlier entry is replaced with the corrected one — not joined by a second sandwich. A log that double-counts every fix over-reports instead, which is the opposite error and no more useful.

What it cannot do is see the oil. A vision model reads a plate about as well as a person looking at the same photo, so the tablespoon in the pan is still yours to mention. Say it, and it is logged. The tool removes the friction; the honesty is still yours.