Methodology

How Nonna decides.

Ask Nonna runs one test on every product: would a 1955 Italian grandmother recognize these ingredients as food? That’s the whole heuristic. This page explains how that translates into ratings, what the rules deliberately don’t try to be, and why some of the most common critiques don’t land the way they look like they should.

The heuristic

A grandmother, not a scientist.

The 1955 Italian grandmother is a stand-in for recognizable real food. She bought olive oil, flour, tomatoes, eggs, salt, named spices, and meat from a local butcher. She didn’t buy modified wheat starch, calcium disodium EDTA, or “artificial flavors.” The test isn’t a scientific assessment of nutrition; it’s a simplicity test of recognizability. If a home cook could plausibly have it in their kitchen, it’s real food. If it requires industrial processing to exist, it isn’t.

This deliberately ignores a lot of nutritional nuance. A 1955 grandmother kept butter and lard in her kitchen, both of which modern dietitians have opinions about; Nonna says fine. Refined white sugar has been in kitchens for centuries; Nonna says fine. Whether those things are healthy is a different question and not the one Ask Nonna is built to answer.

Ingredient ratings

Three tiers, applied in order.

Every individual ingredient gets one of three ratings. The order of the checks matters: an ingredient is checked against the RED list first, then YELLOW, then defaults to GREEN if it looks like real food.

RED — Industrial chemistry

Things that exist only because of food-tech engineering: EU-banned or restricted additives (titanium dioxide, potassium bromate, BVO, BHA, BHT, partially hydrogenated oils, artificial dyes like Red 40 and Yellow 5), industrial sweeteners (high fructose corn syrup, corn syrup solids, glucose-fructose syrup), industrial additives (sodium nitrite, aspartame, sucralose, polysorbate 80, EDTA, hydrolyzed protein, textured vegetable protein), deceptive labels (“artificial flavors,” “enriched flour,” “caramel color,” bare “enzymes” with no source disclosure).

YELLOW — Modern but honest

Things that aren’t industrial chemistry but also aren’t kitchen-traditional. Industrial seed oils (canola, soybean, sunflower, etc.), undisclosed-but-not-synthetic blends (“natural flavors,” generic “spices”), processing aids and stabilizers (lecithin, xanthan gum, guar gum, mono/diglycerides), concentrated plant-source sweeteners (tapioca syrup, brown rice syrup, dextrose, maltodextrin), protein and starch isolates (whey protein isolate, pea protein isolate, wheat protein, resistant wheat starch). Common in modern packaged food; not catastrophic on their own.

GREEN — Real food

The default for any recognizable whole food or traditional ingredient. Flours, whole grains, sugar, honey, maple syrup, olive oil, butter, dairy in any form, eggs, meat, fish, produce, legumes, nuts, named spices and herbs, yeast, salt, vinegar, citric acid, cocoa, chocolate, vanilla extract. GREEN doesn’t require an explicit list match: if Nonna could buy it, grow it, or make it in her kitchen, it’s GREEN.

The product verdict

The aggregate test.

Three ingredient tiers, but the product verdict is binary — Approvato or Mamma Mia — because that’s the call you’re actually making in the grocery aisle. The roll-up follows what we call the aggregate test:

  • Any RED ingredient → product is Mamma Mia.
  • Zero REDs and a handful of YELLOWs → still Approvato. A few modern-but-honest additives don’t invalidate an otherwise-real-food product.
  • Zero REDs but the label is dominated by YELLOWs — five or more — and the product is Mamma Mia even with no single offending ingredient.

That last rule is what catches a product whose label is a wall of industrial isolates and stabilizers: nothing on it is industrial chemistry by itself, but the aggregate isn’t real food. The whole is more than the sum of its parts.

The threshold (that “handful”) is a calibration knob, not a constant. It gets tuned against real scan output. It’s intentionally a bit fuzzy — this is a heuristic, not a court ruling. A short label whose only extras are a single seed-oil blend still reads as real food; disclosure of that blend as several oils doesn’t make it worse.

The framing

“You know it when you see it.”

U.S. constitutional doctrine has a category for things that resist precise definition but are still real: Justice Potter Stewart’s famous I know it when I see it standard for obscenity. The point isn’t that judgment is arbitrary — it’s that some categorizations are legitimately gestalt rather than algorithmic. The community standard is what matters.

That’s the right frame for what Ask Nonna does. The verdict on a label isn’t a deterministic chemistry calculation; it’s a community-located judgment about whether a product reads as food. There’s a hard-science core (Red 40 is a synthetic dye whether you’re in Naples or Houston) surrounded by a much larger culturally-shaped boundary that genuinely varies.

A common critique

“US labels just list more.”

The most sophisticated critique of clean-label apps goes something like this: U.S. ingredient lists look longer and scarier than European ones because U.S. regulations require declaration of more incidental additives, processing aids, and minor components. So a US peanut butter with four ingredients and an EU peanut butter with two ingredients might be the same actual product; what differs is disclosure, not chemistry.

That’s partly true and worth taking seriously. Regulatory equivalence by category is real:

  • The EU lets manufacturers omit some carry-over additives and processing aids from labels that the U.S. requires to be declared.
  • The EU is stricter on flavor source disclosure (a “natural strawberry flavor” from a single dominant source has to be named), but generic “natural flavoring” from multiple sources is still legal there.
  • Some ingredients (like Red 40, BHA, titanium dioxide) are banned or restricted in the EU and allowed in the U.S. That’s a categorical asymmetry, not a labeling one.

But here’s why this critique doesn’t break the heuristic. Ask Nonna isn’t making a claim about which products are healthier based on label length. The test is: would a 1955 grandmother recognize these ingredients as food? When the label says “modified wheat starch” or “wheat protein isolate” or “natural flavor,” the answer is no — regardless of whether a European label would have hidden that under a more general category. The disclosure makes the product easier to evaluate, not worse.

And in the categorical-asymmetry direction, where the EU outright bans something the U.S. allows, the heuristic is pointing in the same direction the regulator is. That’s the core of Ask Nonna’s position.

What this isn’t

Not science. Not health advice.

Ask Nonna is a recognizability test, not a nutrition score. It doesn’t track calories, macros, vitamins, or any health outcome. It doesn’t claim that ingredients rated RED cause disease, or that ingredients rated GREEN are healthy. It doesn’t substitute for medical or clinical guidance.

What it offers is a fast read of what’s in the box, framed in a deliberately simple way: did this come from a kitchen, or did it come from a factory? People are welcome to disagree with individual calls. The disagreement itself, when aggregated across many users, is one of the most useful signals Ask Nonna captures.

If you need clinical nutrition advice, talk to a registered dietitian.

Disagree with a verdict?

Browse the public record, see how Nonna called things, and tell us when she got it wrong. The methodology improves with the community.