Ultra-processed foods: the real problem is not the category, it is the frequency

In the past ten years the NOVA classification, developed at the University of São Paulo, has become the international reference for talking about ultra-processed foods (UPF). Four categories: fresh or minimally processed foods, culinary ingredients, processed foods, ultra-processed foods. The fourth category is the one that makes headlines, and often the most misunderstood one, both in the press and on the shelf.

The real problem is not that a single UPF is harmful in itself. An industrial snack eaten once a week does not significantly move any health marker in an otherwise balanced person. The problem is substitution: UPFs are, by design, hyper-palatable, calorie-dense, low in fiber and engineered to drive repeat consumption. When they become a stable part of a family's weekly rotation, they push out other categories, fruit, vegetables, legumes, whole grains, much more than people realize.

The most recent studies (Monteiro et al., 2024; BMJ 2023; Lancet Public Health 2024) show a robust association between high UPF frequency (more than 50% of daily calories) and increased cardiovascular risk, metabolic syndrome and all-cause mortality. But the same literature shows the risk almost disappears below the 20% threshold. That means the correct message is not to eliminate but to calibrate.

There is also a methodological issue worth naming: the NOVA 4 category is broad. It simultaneously includes sugary drinks, industrial salty snacks, shelf-stable packaged bread and flavored yogurts with additives. Putting everything together makes it hard to distinguish high-impact items from marginal ones. That is why Nutrilayer does not stop at NOVA classification: it crosses category with caloric density, lipid profile, added sugar content and fiber.

That is why in Nutrilayer we never label single products as 'good' or 'bad'. We instead measure the weekly UPF share of the basket and compare it with the share of fresh and minimally processed foods, plant variety and overall protein coverage. When the UPF share exceeds 30% of weekly calories, the system suggests concrete swaps, not renunciations, swaps, based on real in-store availability and customer preferences.

Concrete example. A family with two children that regularly buys packaged snacks for school does not receive a punitive alert. It receives three alternatives coherent with its usual basket: Greek yogurt with fresh fruit and nuts, whole-grain bread with ricotta and honey, homemade mini sandwiches with semi-whole flours. Comparable weekly cost, explicit preparation time in minutes, clear nutritional impact forecast. The choice stays with the customer.

For retailers it is an important perspective shift: instead of selling less UPF, it is about selling variety better. Pilot data show that customers with LifeScore™ above 70 generate average baskets 14% higher and return to the store 22% more often. Health, properly communicated, does not reduce revenue: it qualifies and stabilizes it, because it builds a relationship that goes beyond price.

There is finally a cultural dimension that deserves attention. In Italy 'industrial food' is often perceived in binary terms: either demonized wholesale, or defended wholesale. Reality is more nuanced: there is a quality food industry producing excellent minimally processed foods (tomato passata, canned legumes, frozen vegetables), and there is an industry producing low-nutrient hyper-palatables. Helping customers tell them apart is the task of a serious nutritional intelligence system.

The operational conclusion is simple: it is not the category that determines outcome, it is frequency inside a context. Monitoring frequency, showing it clearly to the customer and suggesting realistic alternatives is what produces change. The rest, including marketing crusades against single products, is noise.