Calculated vs lab-tested nutrition facts: which does the law accept?

Every founder putting a packaged food on a shelf hits the same fork: send the product to a laboratory for nutritional analysis, or calculate the values from the recipe. The lab route is sold as the "safe" one — but that framing gets the law backwards. In every market we work in, the regulation does not ask how you produced the numbers. It asks whether the numbers on the label match the food in the package. Understanding that distinction is the difference between spending your compliance budget where it protects you and spending it out of anxiety.

The rule you are actually being measured against

Nutrition-labeling regimes are, almost universally, outcome-based and method-neutral. They define compliance by comparing a sample of the finished product against the declared values, within a set tolerance — and they are silent on whether you reached those declared values by analysis, by calculation from a food-composition database, or by a mix.

• In the United States, 21 CFR 101.9(g) defines compliance by market sampling against tolerances, and FDA's own 1998 database guidance treats calculation from a nutrient database as a manufacturer's legitimate choice.
• In the United Kingdom and EU, assimilated Regulation (EU) No 1169/2011 states outright that declared values may be based on the manufacturer's calculation from the known or actual average values of the ingredients used — calculation is named in the text, not tolerated as a loophole.
• Across Mexico, Chile, and Colombia, the warning-seal systems (NOM-051, Ley 20.606, Resolución 810) turn on nutrient thresholds per 100 g or 100 ml. Whether a product crosses a threshold is a question you can answer from the formulation before you ever pay for a test.

The common thread: these regulations recognize calculation as a basis for nutrition labeling. What none of them offer is a government office that pre-approves your label. Be wary of any tool or lab implying its numbers are "certified" or "approved" — for packaged foods that status generally does not exist, and the final labeling decision rests with the food business operator no matter which method produced the numbers.

Where calculation is not just cheaper — it is the correct method

For some line items, a lab physically cannot give you the answer:

• Added Sugars (US) and analogous distinctions. A lab measures total sugars in a sample; it cannot tell which grams were added versus intrinsic to fruit or dairy, because they are the same molecule. That is a fact about your formulation, so it must be derived from the recipe.
• Per-serving and per-container math, dual columns, rounding. These are deterministic transformations of the base values, governed by the regulation's own tables. A lab hands you raw analytical numbers; the statutory panel still has to be computed from them.
• Reformulation. Change a supplier or nudge a ratio and analytical results are stale. A recipe model recalculates in minutes; a re-test restarts the clock and the invoice.

Where a lab test still earns its cost

Calculation and analysis are complements, not rivals. Budget a confirmatory analysis when:

1. Your process shifts composition unpredictably — deep-frying (oil uptake), fermentation, heavy browning, reductions. Database math approximates these less well than clean formulations.
2. A value sits on a threshold or a claim boundary — a warning seal, a "low fat," or a tolerance edge where a few analytical grams flip the outcome.
3. A buyer, retailer, or co-packer contractually requires one. Sometimes the test is a commercial term, not a legal one — and that is a fine reason to run it.

The pattern that works for most small manufacturers: calculate first to design the recipe and reach the market, then verify analytically where the risk concentrates, and recalculate rather than re-test on every change in between.

The thing that actually defends a label

Whichever method you use, enforcement — from a regulator's market sample or, in the US, from class-action litigation — rewards the same thing: documentation. "Our software said so" is not a position. "Here is the database and its version, the match chosen for each ingredient, the yield and retention adjustments, the rounding rule applied to each line, and the regulatory provision each step relied on" — that is a position.

So the real question is not calculated or tested. It is: whatever produces my numbers, does it also hand me the trail? Insist on that trail from any tool or consultant, not just the finished artwork.

---

FoodFact calculates nutrition from authoritative food-composition databases, applies each destination country's format and threshold rules, and issues a report that cites the provision behind every value — the documentation trail this post is about. See what one order produces for your market on the United States, United Kingdom, or any country page, and read the acceptance and data-source questions in detail on the FAQ.

This article is general information about nutrition-labeling rules, not regulatory advice. The final labeling decision, and responsibility for the finished product, rests with the food business operator.