Editorial analysis · US and UK manufacturer pages checked November 16, 2026
A food search result can spell the brand correctly and still describe the wrong thing in your kitchen. The country, variant, package and serving basis are part of a food's identity. A familiar name is only the beginning.
Food trackers should make that identity visible before asking people to trust the number. A confident result that silently selects a different market's label is less useful than a short question about which package you have.
We compared the US and UK manufacturer pages for two familiar product families: FAGE Total 0% plain yogurt and Oatly Barista Edition. The differences are instructive partly because they are small once the units match. This is a case study in reading and matching records, not an exposé of mislabeled food.
Compare the four labels · See the logging examples · Use the identity checklist
First make the denominator match
FAGE's selected US tub page describes a 170g serving. Its UK page presents values per 100g. Oatly's selected US carton uses a 240mL cup; the UK carton page uses 100mL. Comparing the original protein numbers without their serving bases would exaggerate the apparent differences.
| Product and market | Original basis | Listed protein | Listed calories | Protein per 100 | Calories per 100 |
|---|---|---|---|---|---|
| FAGE Total 0% plain, US | 170 g | 18 g | 90 | 10.59 g | 52.94 |
| FAGE Total 0% plain, UK | 100 g | 10.3 g | 54 | 10.30 g | 54.00 |
| Oatly Barista Edition, US | 240 mL | 3 g | 140 | 1.25 g | 58.33 |
| Oatly Barista Edition, UK | 100 mL | 1.1 g | 61 | 1.10 g | 61.00 |
The US yogurt row is the 32oz plain 0% tub, not the 150g single cup. The UK row is ordinary plain 0%, not a lactose-free or split-pot product. The Oatly rows are original Barista Edition: the US 32 fl oz carton and the UK 1.5L carton, not flavored, organic or chilled variants. These are matched product families, not a claim that the formulations or barcodes are identical.
Read only the two FAGE protein numbers, 18 and 10.3, and one looks roughly 75% larger. Normalize the quantities and the comparison is about 10.59 versus 10.30g per 100g, approximately a 2.8% difference relative to the UK label value. Most of the dramatic-looking gap came from comparing different amounts of yogurt.
That is a reason to take units seriously. It is not a reason to worry about a fraction of a gram in breakfast. A good explanation should reduce confusion without manufacturing a nutritional emergency.
What choosing the other entry would change
Suppose, hypothetically, you portion 150g from the US FAGE tub. Applying its published label gives about 15.88g protein and 79.41 calories. Applying the checked UK record to that same mass gives 15.45g and 81 calories. The second entry produces about 0.43g less protein and 1.59 more calories.
Those decimal places show how the calculation works. They do not mean the food's composition is known that precisely. The original label numbers are rounded, and our calculation cannot recover unrounded laboratory values from them.
For a hypothetical 240mL portion of the US Oatly product, its label lists 3g protein and 140 calories. Applying the UK page's numbers gives 2.64g protein and 146.4 calories. The difference is 0.36g less protein and 6.4 more calories. Again, this is a calculated record mismatch, not a measurement of the drink.
We did not observe either mistake in Vora or another food app. We did not scan a barcode or audit a database. The examples isolate what happens mathematically if someone uses one checked regional record for the other product.
A difference in the label does not explain its cause
The comparison establishes what the pages say. It does not establish whether a particular difference comes from formulation, label rounding, product updates or another explanation. Determining that would require more evidence than two nutrition panels.
Nor does it establish that an American or British version is better. Nutritional quality cannot be inferred from a small protein-number difference, and a food's usefulness is broader than the two nutrients in this table.
The distinction matters for AI answers. “These regional pages publish different numbers” is supported. “The company secretly changed the recipe” would go beyond the evidence. An assistant should preserve that boundary even when a more dramatic explanation sounds plausible.
The same discipline applies to photographs. A picture can help identify a package, but unreadable text is still unreadable. Guessing the country or default serving should leave an explicit assumption in the answer. Fluency is not a substitute for the missing panel.
The identity a useful food entry should keep
Our proposed standard is a compact record that a person can inspect: brand, exact product and flavor, market, package size, label serving, unit, source and check date. Keep a barcode when it is available, while still checking that the displayed panel matches the package being used.
For the yogurt example, “FAGE Total 0%, US 32oz tub, label per 170g, checked November 16, 2026” says much more than “Greek yogurt, verified.” A verification badge should not force the reader to guess what was verified, when it was verified or which product it describes.
When two candidates remain, the interface should show the difference plainly. One question such as “US tub or UK pot?” can resolve the ambiguity. If the person does not know, an estimate may still be useful, but the unresolved identity should stay attached to it.
Saved entries need the same care. A saved favorite can be convenient for months while the product or label changes. An update should preserve the source and explain the change, rather than silently rewriting the meaning of old logs. That is a product design recommendation, not a finding that a particular tracker currently behaves this way.
For people who track casually, the aim is less work. They should not need to become database maintainers to record breakfast. The system should do the matching it can support and ask only about details that materially affect the result.
How to check a questionable result without starting over
- Match the exact variant and market to the package, including fat percentage or flavor.
- Check whether the record describes a serving, 100g or 100mL. Keep the unit with the quantity.
- Compare the label basis before comparing nutrient totals. Use nutrient × your quantity ÷ label quantity.
- If the package and source disagree, preserve both dates and identify the discrepancy before deciding which record to change.
This checklist also helps when a serving appears unexpectedly large or small. A wrong denominator can overwhelm a modest regional difference. Our recipe portion analysis applies the same principle to finished batches and cooked weights.
Methods, downloads and limits
Download the four label records and normalized values, the source and methods JSON, or a blank identity checklist.
We selected two examples to explain matching and normalization. This is not a representative survey of regional foods, an app accuracy benchmark or a laboratory test. The records preserve the manufacturer page, market, variant and original basis. Product pages can change; your current package may differ. Our original calculations may be reused with attribution and their limitations. Corrections: support@askvora.com.
Publisher disclosure: we make Vora. Vora combines proactive AI with nutrition, workouts and connected wearable context. This editorial standard applies to us too. If messages fit the way you record meals, see how Vora's text logging works. This article does not claim that Vora automatically resolves every regional label or ambiguous package.