How to start.
-
01
Learn what counts
Ultra-processed means industrial formulations, NOVA group 4. Plain frozen vegetables, tinned beans and plain yogurt are fine. Packaged is not the same as ultra-processed.
-
02
Fix breakfast first
It's the most formula-heavy meal most people eat, and the most repetitive. Solve one breakfast and you've solved 365 of them.
-
03
Ignore the health aisle
Protein bars, meal shakes and diet snacks are still ultra-processed formulations. The branding changed. The manufacturing didn't.
-
04
Slow the meal down
Food that needs chewing buys you time. Fullness signals arrive late, and processed food is engineered to be finished before they land.
Why it works.
-
Calorie intake
In a 4-week inpatient crossover trial, 20 adults ate 508 calories a day more on an ultra-processed diet than on an unprocessed one matched for calories, sugar, fat, sodium and fiber.
-
Weight
The same 20 participants gained 0.9 kg during two weeks of ultra-processed eating and lost 0.9 kg during two weeks of unprocessed eating.
-
Eating speed
Meal eating rate was significantly faster on the ultra-processed diet, by about 17 calories a minute, a plausible mechanism for overeating before fullness signals register.
-
Mortality
Among 19,899 Spanish adults (335 deaths, 200,432 person-years), the highest quartile of ultra-processed food intake had 62% higher all-cause mortality (HR 1.62, 95% CI 1.13 to 2.33).
Who swears by it.
Chris van Tulleken
Ate 80% ultra-processed food for a month under UCL supervision; wrote Ultra-Processed People.
Jamie Oliver
Food Revolution and Ministry of Food campaigns push scratch cooking over processed and fast food.
Michael Mosley
Advocated steering clear of harmful processed foods in favor of vegetables, legumes, and fermented foods.
John's take.
I want to give you the strongest version of this argument and then tell you exactly how thin it is. The strong version: Kevin Hall locked 20 adults in a metabolic ward for four weeks and fed them ultra-processed or unprocessed food, matched for calories, sugar, fat, sodium and fiber, with instructions to eat as much or as little as they liked. On the processed food they ate 508 more calories a day and gained 0.9 kg. On the unprocessed food they lost 0.9 kg. Nobody was dieting. The food was doing the deciding. That is the cleanest experiment we have in nutrition and it points one direction.
Now the thin part. Twenty people. Four weeks. One trial. The processed meals were also softer and more energy dense, so we can’t say which part of processing drives the overeating. And the scary numbers, the 62% higher mortality, come from observational cohorts where people eating the most ultra-processed food also smoke more and earn less. Adjustment helps. It doesn’t fix that. When you see UPF blamed for everything on the internet, it’s usually that confounded data doing the talking, not the metabolic ward.
The definition is where most people go wrong, and it’s the thing I’d fix first. Ultra-processed doesn’t mean packaged. Plain frozen vegetables, tinned beans and plain yogurt are not the enemy, and telling people to fear all packaging just makes cooking feel impossible so they give up. The trap I see constantly is the health aisle: protein bars, meal replacement shakes, high-fiber snacks that are pure NOVA 4 with a green label. Meanwhile a bag of frozen peas is basically peas. Get that distinction right and this habit is workable. Get it wrong and you’ll be exhausted by week two.
Common questions.
What counts as ultra-processed food?
Ultra-processed means industrial formulations, NOVA group 4, not simply anything in a package. Plain frozen vegetables, tinned beans and plain yogurt do not qualify. The giveaway is a list of ingredients you'd never find in a kitchen: isolates, hydrolysates, emulsifiers, added flavors and colors built into a manufactured product.
Does ultra-processed food actually make you eat more?
In the one trial that tested it properly, yes. Twenty adults ate 508 calories a day more on an ultra-processed diet than an unprocessed one, and gained 0.9 kg in two weeks. Both diets were matched for calories, sugar, fat, sodium and fiber, so the effect came from processing itself rather than macronutrients.
How strong is the evidence against ultra-processed food?
Rigorous but tiny. The causal evidence is a single 4-week inpatient trial in 20 people. Its ultra-processed meals were also softer and more energy dense, so it can't isolate which aspect of processing drives overeating. The mortality links are observational and tangled up with smoking and income.
Are protein bars ultra-processed?
Almost always, yes. This is the most common mistake I see: swapping junk for health-branded bars and shakes, which are typically still ultra-processed formulations. A high-protein label doesn't change how the food was manufactured or how fast you'll eat it. Judge the process, not the marketing claim on the front.
Why does processed food make you overeat?
Eating rate looks like a key mechanism. Participants ate about 17 calories a minute faster on ultra-processed meals. Fullness signals take time to arrive, so food you can finish in four minutes gets past them. Soft, energy-dense food that requires almost no chewing is the design problem in one sentence.
Do I have to count calories if I quit ultra-processed food?
That's the appeal. Trial participants were never told to restrict anything, yet ate substantially less on unprocessed food and lost 0.9 kg over two weeks. Cutting ultra-processed food tends to reduce intake without counting. Just remember this is one small study, and your kitchen isn't a metabolic ward.