There is a category of food that reliably disappears without registering. A share packet of crisps. A sleeve of sweet biscuits. A pastry with a coffee, a muesli bar in a lunchbox, a tub of cinema popcorn, a block of chocolate on a weeknight, a glass of juice, a bowl of sweetened cereal, a handful of rice crackers reached for because they seemed like the sensible option. The energy goes in, the fullness never quite arrives, and the usual explanation offered is a failure of self-control.
Those foods look nothing like each other. They sit in different aisles, at different prices, with different marketing. Some are openly indulgent and some are positioned as the healthy choice. Sweet and savoury, premium and cheap, the muesli bar with the wholegrain claim and the puffed corn snack with none. And they behave in the body almost identically.
At Anova we think that is worth pausing on, because it is the tell. When products this different produce the same result, the result is not coming from the sugar, or the fat, or the salt, or any single ingredient you could read off a panel. It is coming from something they have in common that the panel does not show.
They have four things in common, and every one of them was a decision.
How fullness is actually built
Satiety is not a switch. It is a chain of signals assembled over roughly twenty to thirty minutes.
Food arrives in the stomach and stretches it. That mechanical distension is the earliest signal. As the stomach empties, nutrients reach the small intestine and trigger the release of gut hormones, principally cholecystokinin and peptide YY. These slow gastric emptying further and signal fullness to the brain through the vagus nerve. Steinert and colleagues set this negative-feedback loop out in detail in Physiological Reviews in 2017. It is the body’s built-in portion control, and it is the same machinery weight management medications act on.
The chain has four inputs that a food can supply or withhold: how much the food weighs and how much it stretches the stomach, how long it takes to chew, how fast the energy arrives, and what nutrients reach the intestine to trigger the hormonal response.
Remove all four and you have a product that can be eaten indefinitely.
The trial that isolated it
The cleanest evidence comes from a controlled inpatient trial run by Kevin Hall’s group at the US National Institutes of Health, published in Cell Metabolism in 2019. Twenty weight-stable adults spent two weeks eating an ultra-processed diet and two weeks eating an unprocessed one, in random order, with everything provided and intake unrestricted.
The design is what makes it useful. The two diets were matched for presented energy, energy density, macronutrients, sugar, sodium and fibre. On paper they were the same diet.
Participants ate 508 kilocalories a day more on the ultra-processed diet, roughly 2,100 kilojoules. They gained 0.9 kilograms across the processed fortnight and lost 0.9 kilograms across the unprocessed one. A follow-up analysis by Forde and colleagues found the mechanism sitting in plain sight: people ate the ultra-processed food at 48 kilocalories per minute against 31 on the unprocessed diet, more than fifty per cent faster.
Same potato, opposite result
The Australian Food Composition Database makes the point without needing a trial. Plain salted potato crisps carry 2,140 kilojoules per 100 grams, with 33.3 grams of fat, 3.2 grams of fibre, and 1.8 grams of water. Boiled potato, peeled and drained, carries 298 kilojoules per 100 grams and is roughly three-quarters water.
Seven times the energy, same weight, same vegetable. The water was driven off, fat was added, and the chew was removed. Each of those steps subtracts one of the four inputs.
This is a specification, not an accident
Here is the part that is rarely said plainly. The properties that make these foods fail to satisfy are not byproducts of manufacturing. They are the product requirements, and the industry has names for them.
The food that disappears. In 2013 the investigative journalist Michael Moss, reporting for the New York Times Magazine and in his book Salt Sugar Fat, interviewed food scientist Steven Witherly about extruded cheese puffs. Witherly’s term for their central quality was “vanishing caloric density”: a food engineered to dissolve on contact with saliva, so that the mouth registers something substantial and then finds nothing there. The crunch arrives, the flavour arrives, and the food is gone before chewing has properly begun. The energy stays. The signal never fires. Witherly described the puff, admiringly, as among the most marvellously constructed foods on the planet. He was not being ironic. By the specification he was assessing it against, it is superb.
The flavour engineered not to say stop. Satiety is partly flavour-specific. Barbara Rolls demonstrated in 1981 that the pleasantness of a food declines as you eat it while other foods stay appealing, which is why a strong distinctive taste eventually tells you to stop. Moss reports that this is understood inside product development and designed around: a big, distinct flavour overwhelms the brain, which responds by suppressing the desire for more, so the commercially successful formulation is the one that is intensely moreish without ever resolving into a single dominant note. The stopping mechanism was identified, and then engineered out. One soft drink flavour, Moss reports, went through 61 formulations and nearly 4,000 consumer tastings with regression analysis applied to find the optimum. That is not cooking. It is optimisation against a target, and the target is the point at which you stop.
Variety inside the pack. The same finding runs the other way. Rolls found that people offered sandwiches with four fillings ate a third more than those offered one filling, and that three distinct yoghurt flavours produced higher intake than the single favourite flavour. A 2022 meta-analysis in the American Journal of Clinical Nutrition confirmed the variety effect across the literature. Multipacks, flavour-assortment boxes and share platters are that finding, packaged.
The packet with no end in it. The strongest evidence of all sits here. A 2015 Cochrane review led by Gareth Hollands at Cambridge pooled 72 randomised trials covering 6,711 participants and found that people consistently consume more when given larger portions, packages or tableware. The authors estimated that removing large sizes from the diet entirely could cut daily energy intake by 12 to 16 per cent in UK adults and 22 to 29 per cent in US adults.
The detail that matters most for anyone who has been told to try harder: the effect did not vary by body mass index, by susceptibility to hunger, or by how much people consciously restrained their eating. Self-control did not protect the people who had it. The packet won regardless.
A 50 gram packet contains a stopping point. A 175 gram share bag does not. That is the entire difference, and it is a decision someone made in a meeting.
The category, named
The pattern is consistent enough to be a rule rather than a list. A food fails to satisfy when it is dense in energy, low in water, low in fibre and protein, and requires almost no chewing. The more of those boxes it ticks, the further you will get through it before anything registers.
Sharing the crisp profile closely: extruded cheese puffs and corn snacks, corn chips (2,053 kilojoules per 100 grams), prawn crackers, rice crackers, pretzels, and most savoury biscuits.
Sweet and equally undetectable: sweet biscuits, croissants and pastries, doughnuts, chocolate, ice cream, and sweetened breakfast cereals.
Liquid, which removes chewing entirely: soft drink, fruit juice, flavoured milk, sweetened yoghurt drinks, and most smoothies.
Wearing a health halo: muesli bars, dried fruit, veggie straws, and rice cakes, which is a marketing position rather than a satiety property.
Popcorn deserves a separate paragraph, because it is the cleanest demonstration in the supermarket. The kernel does not change. Air-popped and plain, popcorn is a whole grain, high in fibre, extremely light, and one of the better volume-per-kilojoule snacks available: three cups weighs around 24 grams and carries roughly 390 kilojoules. Cooked in oil and coated in butter for a cinema tub, the same food reaches around 2,300 kilojoules per 100 grams with 35 grams of fat. Nothing about the corn was altered. Fat was added, and the serving vessel was made enormous. That is the whole trick performed twice, on a food that started out on the right side of the line.
Food form: why the label misleads
The physical structure of a food changes how it behaves even when its nutrients are identical. This is the same principle behind the sugar that hides in food that doesn’t taste sweet.
Haber and colleagues demonstrated it in The Lancet in 1977, feeding people whole apples, apple puree and apple juice matched for carbohydrate. The juice was consumed about eleven times faster than the whole fruit and the puree about four times faster. Satiety fell in exact step with how disrupted the fruit was.
A 2014 meta-analysis in the American Journal of Clinical Nutrition confirms the general rule: faster eating produces higher energy intake at a meal, with no compensating drop in later hunger.
The satiety index, and what it cannot tell you
Australian researchers put numbers on this thirty years ago. Susanna Holt and colleagues at the University of Sydney fed people fixed 1,000 kilojoule portions of 38 common foods and tracked fullness over two hours, indexed against white bread at 100. Boiled potato scored 323. A croissant scored 47.
The limits deserve saying plainly. The index measured short-term fullness after single fixed portions in groups of eleven to thirteen people per food. It did not measure what anyone ate the next day. It is a compass, not a ranking to eat by.
One thing this article is not saying
Energy density on its own does not make a food a problem. Whole nuts are energy dense and reasonably satisfying, because chewing only partly ruptures their cell walls and some of the fat is never absorbed. Olive oil, full-fat yoghurt, avocado, eggs and cheese are all energy dense and none of them belong in the category above.
The distinguishing feature is not richness. It is engineering: whether the properties that trigger fullness were removed on purpose. A food that grew that way and a food that was designed that way are different objects.
What this means in practice
The four levers follow from the mechanism, and none is about willpower.
Serve it out of the packet. This is the single highest-yield change, because it restores the stopping cue the packaging was designed to delete. This is a change to the environment rather than to the person, which is exactly why it works when resolve does not.
Buy the small pack even when the large one is better value. The pricing that makes the share bag cheaper per gram is part of the same design.
Change the form rather than the food. Whole fruit over juice. Intact grains over milled. Air-popped over cinema tub. Boiled potato over fried and dried.
Add water and volume. Vegetables, legumes or soup alongside lowers the energy density of the whole eating occasion. A 2014 meta-analysis found about 31 per cent greater satiety after a pulse-containing meal, though it did not reduce intake at the next meal.
Protein sits underneath all of this, and the case for protecting it when the plate gets smaller is well established.
The question most people are asked, by themselves and by others, is why they lacked the discipline to stop. It is the wrong question, and it was always going to produce the wrong answer. The better one is why an entire product category exists in which stopping was identified as a commercial problem and solved.
This article discusses general findings from nutrition and appetite research, and reporting on food industry practice. It is general information, not personalised dietary advice, and is not a substitute for care from a GP, dietitian or other qualified health professional.