Appearance
24.3 — How to Read a Food Label
A cereal box says "30% less sugar".
Less than what? Less than the same manufacturer's other cereal, which was extremely high in sugar. The reduced version can still contain more sugar per hundred grams than a biscuit.
"Less" is a comparison, and the label never has to tell you what it is comparing to in a way you would find useful.
Almost every claim on the front of a package is marketing that has been engineered to be technically true.
Almost every fact is on the back.
This page is how to read the back in about ten seconds.
The two panels that matter
The ingredients list, and the nutrition table. Everything else on the packaging is design.
The ingredients list
Ingredients are listed by weight, in descending order.
Which means the first three ingredients are most of what you are eating.
If sugar, or any of its aliases, is in the first three, it is a sugary product regardless of what the front says.
And here is the trick worth knowing: manufacturers split sugars.
A product can contain glucose syrup, dextrose, fruit juice concentrate and invert sugar as four separate entries, each appearing far down the list — while their combined weight would have put sugar at the top.
The names sugar appears under: sucrose, glucose, fructose, dextrose, maltose, lactose, glucose syrup, high-fructose corn syrup, corn syrup, invert sugar, molasses, treacle, honey, agave nectar, maple syrup, rice syrup, malt extract, fruit juice concentrate, and anything ending in -ose.
And "no added sugar" does not mean low in sugar. A fruit juice product with no added sugar can contain a great deal of sugar, because the sugar was already in the fruit, and juicing removes the fibre that made the difference (Chapter 24.1).
Other things worth spotting:
"Vegetable oil" without specification — usually palm oil.
A very long list of things you do not recognise — a marker of ultra-processing (below).
Salt appears as salt, sodium, monosodium glutamate, sodium bicarbonate, sodium nitrite and others.
The nutrition table
Always read the per 100 g column, not the per serving column.
Per 100 g is the only number that lets you compare two products.
Serving sizes are chosen by the manufacturer and are frequently unrealistic — a bag of crisps stated as two servings, a tub of ice cream as four, a pizza as three. Some countries require the serving size to reflect what people actually eat; many do not.
The traffic light thresholds — per 100 g of food:
| Low (green) | High (red) | |
|---|---|---|
| Fat | ≤3 g | >17.5 g |
| Saturated fat | ≤1.5 g | >5 g |
| Sugars | ≤5 g | >22.5 g |
| Salt | ≤1.5 g | >1.5 g |
Anything between is amber.
Drinks use lower thresholds, at roughly half those values.
And salt versus sodium: salt is sodium chloride, and only about 40 percent of it is sodium.
To convert: sodium in grams × 2.5 = salt in grams.
A label saying 0.8 g sodium means 2 g of salt — which is a third of the daily maximum, and the sodium figure looks reassuringly smaller. Labels in the US and several other countries give sodium; the UK and EU give salt.
Daily reference intakes, for an average adult:
Energy 2,000 kcal · total fat 70 g · saturated fat 20 g · total sugars 90 g · free sugars 30 g · salt 6 g · fibre 30 g.
These are averages for labelling, not personal targets — an individual's needs vary considerably with size, sex and activity.
The claims, and what they legally mean
Regulated claims have defined thresholds, and the definitions are frequently not what the words suggest.
| Claim | What it actually means |
|---|---|
| Low fat | ≤3 g per 100 g |
| Fat free | ≤0.5 g per 100 g |
| Low sugar | ≤5 g per 100 g |
| Sugar free | ≤0.5 g per 100 g |
| No added sugar | None added — may be high in sugar |
| Reduced X | 30% less than a similar product |
| Light / lite | 30% less of something — check what |
| Source of fibre | ≥3 g per 100 g |
| High fibre | ≥6 g per 100 g |
| Source of protein | ≥12% of energy from protein |
| Low salt | ≤0.3 g salt per 100 g |
And "reduced fat" is the classic trap. Removing fat removes flavour and texture, so sugar, salt or starch is frequently added back. A reduced-fat product is often not lower in calories at all — compare the energy figures directly.
Unregulated words, which mean nothing:
Natural. Wholesome. Farm fresh. Artisan. Superfood. Clean. Immune-supporting. Made with real fruit.
"Made with real fruit" can mean a trace. "Superfood" has no legal definition and is banned as a health claim in the EU unless accompanied by an authorised specific claim.
And a specific one: "made with wholegrain" tells you nothing about how much. Look for a wholegrain as the first ingredient, or the percentage stated.
Dates
And confusing these causes a large amount of avoidable food waste.
Use by — a safety date. On perishable, high-risk foods: meat, fish, ready meals, some dairy. Do not eat it after this date, even if it looks and smells fine, because dangerous bacteria can be present without any sensory change.
Freezing stops the clock — freeze before the use-by date and it is safe.
Best before — a quality date. The food is safe afterwards; it may just be less good. Dried pasta, rice, tinned goods and biscuits are frequently fine long after it.
Use your senses for best-before foods, and never for use-by foods.
Sell by / display until — for shop stock control. Ignore it.
Once opened, use within X days — this overrides the printed date.
Ultra-processed food
And this is where the label-reading skill has become most useful, because it is the one dietary distinction with rapidly accumulating evidence behind it.
The NOVA classification sorts foods by degree of processing rather than by nutrients:
Group 1 — unprocessed or minimally processed. Fruit, vegetables, meat, fish, eggs, milk, grains, legumes, nuts.
Group 2 — culinary ingredients. Oil, butter, sugar, salt, flour.
Group 3 — processed foods. Group 1 plus group 2: bread, cheese, tinned fish, salted nuts, canned vegetables.
Group 4 — ultra-processed. Industrial formulations containing ingredients not used in home cooking: protein isolates, modified starches, hydrogenated oils, emulsifiers, colours, flavourings, sweeteners, thickeners and bulking agents.
The practical test: if the ingredients list contains several things you would never find in a domestic kitchen, it is ultra-processed.
Typical examples: soft drinks, packaged snacks, mass-produced bread, breakfast cereals, reconstituted meat products, instant noodles, most ready meals, and many products marketed as healthy.
The evidence:
Observational studies consistently associate high ultra-processed food intake with obesity, type 2 diabetes, cardiovascular disease, depression and increased mortality.
And a controlled feeding trial is the most striking finding. Participants lived in a research facility and ate either ultra-processed or minimally processed diets, matched for calories, sugar, fat, fibre and salt, with unlimited amounts available.
On the ultra-processed diet they ate around 500 kcal more per day and gained weight. On the minimally processed diet they lost it.
Same nutrients on paper. Different outcome.
Why, as far as is understood: it is softer and faster to eat, so more is consumed before satiety signals arrive; the energy density is higher; the combination of fat, sugar and salt is engineered to be highly palatable; and it produces a weaker satiety response than whole food containing the same nutrients.
Which is the deepest practical lesson in this Part: the nutrition table does not capture everything that matters about a food.
Additives and sweeteners
E numbers are simply the European identification system for approved additives. An E number is not evidence of anything sinister — vitamin C is E300 and it is on the list.
Approved additives have been through safety assessment. Some people react to specific ones, and certain colours require a warning about possible effects on attention in children.
Artificial sweeteners — aspartame, sucralose, acesulfame K, saccharin, steviol glycosides.
On safety: approved sweeteners have been extensively studied and remain permitted at normal intakes. Aspartame was classified by one WHO agency as possibly carcinogenic in 2023, while the same organisation's safety committee retained the existing acceptable daily intake — a distinction that was widely misreported. Possibly carcinogenic is a statement about the strength of evidence, not about the size of the risk.
On usefulness: replacing sugary drinks with sweetened ones reduces sugar and calories in the short term. The WHO advises against using non-sugar sweeteners for long-term weight control, because the evidence of lasting benefit is weak.
People with phenylketonuria must avoid aspartame, which is why the warning appears (Chapter 21.8).
The ten-second method
In a shop, on any packaged product:
1. Turn it over. Ignore the front entirely.
2. Read the first three ingredients. That is most of what is in it.
3. Look at sugar and salt per 100 g against the traffic light thresholds.
4. Count how many ingredients you do not recognise as food.
5. Compare against the alternative product per 100 g, not per serving.
That is it. It works for anything.
The general principle
The foods with the strongest evidence behind them mostly do not have labels at all.
Vegetables, fruit, legumes, nuts, wholegrains, fish, eggs, plain meat and plain dairy do not come with health claims on the front, because there is no packaging to print them on.
Which produces a rule of thumb that is imperfect and remarkably robust: the more aggressively a food is marketed as healthy, the more likely it is to be a processed product competing for attention.
Nobody advertises broccoli.
What the next page fixes
Chapter 24.4 covers what a healthy diet actually is — evidence only, no ideology, and an honest account of what is established, what is contested, and what nobody knows.