There is no shortage of calorie counting apps. Building another one would have been straightforward, and it would have found users. I did not build one, and I want to explain why — because the reasoning is also the most useful nutrition advice I can give you, whether or not you ever install anything of mine.
The experiment that should have ended calorie counting
In 1992 the New England Journal of Medicine published a study I have quoted to patients more often than almost any other.1 The researchers recruited people who reported being unable to lose weight despite eating, by their own account, under 1,200 calories a day. The obvious hypothesis was a metabolic problem.
The researchers measured actual energy intake with doubly labelled water — a technique that does not care what anyone writes in a food diary — and measured resting metabolic rate directly.
The results were not subtle. The group was under-reporting their intake by an average of 47%, and over-reporting their physical activity by around 51%. Their metabolic rates were normal. There was no metabolic defect. There was a measurement problem, and it was enormous.
These were not people lying to the researchers. They were people who believed their own food diaries — which is precisely what makes the finding so uncomfortable.
The effect has been reproduced many times since, in many populations. Self-reported dietary intake is one of the least reliable measurements in all of clinical research. Everyone under-reports; heavier people and dieters under-report more.
So why do calorie apps still work at all?
They do work, for some people, and I want to be fair about why.
Logging food makes you pay attention, and attention changes behaviour. That is the real mechanism — not arithmetic precision. Someone who logs consistently will eat somewhat less than someone who does not, largely because the act of writing something down interrupts the automatic hand-to-mouth loop.
But notice what that means. If the benefit comes from attention rather than accuracy, then an app whose entire purpose is producing a running total is spending all of its effort on the part that does not work, and none on giving your attention somewhere useful to land.
And there is a cost. A daily total that is 40% wrong, delivered with two-decimal confidence, teaches people to distrust their own bodies. I have sat with patients convinced something was medically wrong with them because the number said one thing and the scale said another. Usually nothing was wrong. The number was.
The question calories cannot answer
Here is the deeper problem, and it has nothing to do with accuracy.
Take two days that both total 1,800 calories. The first is vegetables, beans, oily fish, yoghurt, fruit and nuts. The second is white bread, crisps, a fizzy drink and a chocolate bar. Same number. A calorie counter reports both as an identical, successful day.
They are not remotely the same day. The first covers your protein, fills your potassium, calcium and magnesium, and delivers the B12 your nerves and red cells require. The second delivers energy and very little else — and leaves you hungrier sooner, which is its own well-documented problem.2
This matters more than it sounds, because falling short of nutrients is not a fringe scenario. Analyses of national dietary survey data find that large proportions of American adults take in less than the estimated average requirement of several nutrients — magnesium and vitamin D prominent among them — and that the overwhelming majority fall short on at least one.3 These are not exotic deficiencies. They are ordinary gaps in ordinary diets, and a calorie total is structurally incapable of revealing any of them.
What I actually wanted to hand my patients
In clinic the useful conversation is almost never "you ate 2,140 calories." It is:
- You are not getting enough protein to hold on to muscle while you lose weight.
- Your calcium has been low for months and you are past the age where that is harmless.
- You have cut out an entire food group and nothing has replaced its B12.
- You are pregnant, and your folate and iron requirements have changed substantially.
Every one of those is actionable. Every one of them is invisible to a calorie counter. And every one requires knowing two things at once: what you ate, and what you specifically need — which differs by age, by sex, and enormously in pregnancy and breastfeeding.
That is the app I built. You log food; VitalFill compares it against targets drawn from the NIH Dietary Reference Intakes for your own age, sex and life stage, and shows you which nutrients are filled and which fall short.
Why a filling silhouette rather than a table
An early version showed a table of forty nutrients with percentages. It was accurate, complete, and nobody looked at it twice.
Forty numbers is information, not insight. So the interface became a body that fills up: pick a nutrient and see how much of today's requirement you have covered. Two nutrients invert the logic — saturated fat and total sugars are limits rather than goals, so those change colour as you approach the ceiling instead of filling toward it.
It takes about two seconds to read, which is the entire design brief. A number you glance at and forget changes nothing. A picture you remember while standing in the supermarket changes what goes in the trolley.
What this does not do
I would rather say this plainly than have you discover it.
Tracking nutrients does not escape the under-reporting problem. If you forget the olive oil you cooked in, VitalFill is as blind to it as any other app. What changes is what the imperfect data is used for: a 20% error in a calorie total can be the entire deficit you were aiming at, whereas a 20% error against your B12 target does not alter the conclusion that you have been getting almost none for a month. Patterns tolerate noise far better than totals do.
Nor is any app a substitute for a blood test when a real deficiency is suspected, or for your own physician. What it does is answer the question that should come before any of that: is there any reason to think I am short of something?
That question is worth asking, because the alternative — which is what most people do — is to buy a supplement for a deficiency nobody has demonstrated. I have written about how badly that tends to go.
See what your own diet is missing
Free to download. Log a single day and you will know more about your nutrition than a year of calorie counting would have told you.
Not ready to install anything? The nutrient tracker on this site runs a single day's version of the same calculations in your browser — no account, nothing stored, nothing sent anywhere. There is also a fuller description of the app if you want the detail first.
The short version
Calorie counting asks how much you ate. It is the easier question, it is the one every app already answers, and the answer is usually wrong by a wide margin.
The more useful question is what you got. That one is harder to compute, needs a real food database and a real set of requirements behind it, and has to be personalised to actually mean anything. But it is answerable — and unlike a calorie total, the answer tends to change what somebody does on Tuesday.
That is why I built a nutrient tracker.
References
- Lichtman SW, Pestone K, Berman ER, et al. Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med. 1992;327(27):1893–1898. doi:10.1056/NEJM199212313272701 The doubly labelled water study behind the 47% under-reporting and 51% activity over-reporting figures.
- Hall KD, Ayuketah A, Brychta R, et al. Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial of ad libitum food intake. Cell Metab. 2019;30(1):67–77.e3. doi:10.1016/j.cmet.2019.05.008 The controlled feeding trial showing that matched diets differing in processing produce different intake and weight change — the basis for the claim that identical calorie totals are not equivalent days.
- Reider CA, Chung RY, Devarshi PP, Grant RW, Hazels Mitmesser S. Inadequacy of immune health nutrients: intakes in US adults, the 2005–2016 NHANES. Nutrients. 2020;12(6):1735. doi:10.3390/nu12061735 National dietary survey analysis behind the statement that large proportions of US adults fall below the estimated average requirement for several nutrients.