Research ·
The category does not decide it
44 of 48 categories contain both ultra-processed and non-ultra-processed products.
We classified 48 of the commonest categories in a UK supermarket shop under NOVA. In 44 of them, the classification is not decided by the category at all. It is decided by which specific product you picked up.
44 of 48
categories that contain both group 4 and non-group-4 products
Only 4 categories in the set are uniformly ultra-processed: ready meals and protein bars. Everything else, including several people are confident about in both directions, contains products on either side of the line.
Method
48 categories were selected as the ones most likely to appear on a weekly UK shop. For each, the NOVA definition was applied to the formulations actually sold: what is typically in the product, and whether any of it is a substance used only as an input to manufacturing rather than as a food or a culinary ingredient. Each category has its own page showing the working, and every one is linked below.
A category was recorded as splitting when products meeting the group 4 definition and products not meeting it are both routinely available in a mainstream UK supermarket. It was recorded as not splitting where every mainstream version carries at least one group 4 marker.
What this is not
It also makes no claim about health. NOVA describes how a food was manufactured and this audit reports only that. The four groups are not a ranking and there is no target share. See what NOVA does not tell you.
Why it matters
Almost all public advice about ultra-processed food is given at category level, naming whole aisles rather than products. This audit is a reason to be careful with that framing, because in nearly every category such advice is wrong for some of the products in it.
The direction of the error is not consistent either. Instant coffee is group 1 and sounds industrial. A tub of spread looks like butter and is group 4. Peanut butter is usually group 3 and feels like it should not be. Reduced-sugar baked beans are more likely to be group 4 than the standard tin.
The practical consequence is that the only way to know what share of a shop was ultra-processed is to read the shop, item by item. That is a self-interested conclusion for this site to reach, which is why the method and the limits above are stated in full rather than summarised.
The pattern that recurs
One relationship turned up in enough categories to be worth naming: the version of a product marketed as the lighter or the healthier option is more often the group 4 one. Removing something from a food generally means adding something to do its job.
- Reduced-sugar baked beans gain a non-sugar sweetener.
- Low-fat yoghurt gains a thickener.
- Reduced-fat peanut butter gains a filler.
- Reduced-fat sausages gain water and a binder.
- Spreads exist to behave like butter with less fat, which needs emulsifiers.
That is an observation about manufacturing rather than a recommendation, and it is not a claim that either version is better for anybody.
Updating this
Formulations change, and a category that does not split today may split next year. This page is a snapshot, the working for every category is public, and any correction is welcome at founders@dupi.co.uk. Re-runs are recorded below rather than replacing the figure silently.
This audit has been re-run 5 time since it was first published. The set grew from 24 categories to 48, and the proportion held: all six additions split.
Every category, and the working
| Category | Splits? | Where the line falls |
|---|---|---|
| Bread | Yes | Group 3 without an emulsifier, group 4 with one |
| Breakfast cereal | Yes | Plain oats group 1, most boxed group 4 |
| Crisps | Yes | Ready salted group 3, flavoured group 4 |
| Yoghurt | Yes | Plain group 3, most flavoured group 4 |
| Hummus | Yes | Usually group 3, preservatives move it |
| Baked beans | Yes | Most tins group 4 on modified cornflour, short-list tins group 3 |
| Cheese | Yes | Block group 3, processed slices group 4 |
| Peanut butter | Yes | Usually group 3, no-stir versions group 4 |
| Plant milk | Yes | Usually group 4, plain versions group 3 |
| Sausages | Yes | Traditional group 3, most packs group 4 |
| Bacon | Yes | Dry-cured group 3, reformed group 4 |
| Chocolate | Yes | Plain dark group 3, most bars group 4 |
| Pasta sauce | Yes | Passata group 1, most jars group 4 |
| Ice cream | Yes | Traditional dairy group 3, most tubs group 4 |
| Fruit juice | Yes | Pure juice group 1, juice drinks group 4 |
| Tofu | Yes | Plain group 3, marinated and substitutes group 4 |
| Butter and spreads | Yes | Butter group 2, most spreads group 4 |
| Soup | Yes | Fresh chilled group 3, tinned and packet group 4 |
| Biscuits | Yes | Plain shortbread group 3, most packets group 4 |
| Pizza | Yes | Plain margherita group 3, most group 4 |
| Coffee | Yes | Instant and ground group 1, sachets group 4 |
| Granola | Yes | Oats and syrup group 3, fortified group 4 |
| Ready meals | No | Effectively always group 4, premium included |
| Protein bars | No | Group 4 by construction, the protein is an isolate |
| Tortilla wraps | Yes | Plain group 3, but almost all supermarket wraps group 4 |
| Mayonnaise | Yes | Full-fat group 3, light versions group 4 |
| Fish fingers | Yes | Fillet in plain crumb group 3, battered or reformed group 4 |
| Frozen chips | Yes | Plain group 3, coated and crispy group 4 |
| Popcorn | Yes | Kernels group 1, flavoured bags group 4 |
| Jam | Yes | Standard group 3, reduced-sugar group 4 |
| Pasta | Yes | Dried group 1, filled fresh group 4 |
| Rice | Yes | Plain group 1, flavoured pouches group 4 |
| Bagged salad | Yes | Leaves group 1, kits with dressing group 4 |
| Tinned fish | Yes | In water group 1, in brine group 3, with mayonnaise group 4 |
| Smoked salmon | Yes | Real smoke group 3, smoke flavouring group 4 |
| Honey | No | Always group 2. Honey-flavoured products are a different product |
| Eggs | Yes | Group 1 always; liquid egg with stabilisers group 4 |
| Ham | Yes | Carved group 3, reformed sliced group 4 |
| Nuts | Yes | Plain group 1, salted group 3, dry roasted group 4 |
| Dried fruit | Yes | Plain group 1, oiled or sugared group 3, coated group 4 |
| Noodles | Yes | Dried group 1, instant group 4 |
| Beer and wine | Yes | Fermented group 3, alcopops and pre-mixed group 4 |
| Flour | Yes | Wholemeal group 1; white is contested because UK law requires fortification |
| Frozen vegetables | Yes | Plain group 1, sauced or coated group 4 |
| Milk | Yes | Plain group 1 at any fat content, flavoured group 4 |
| Energy drinks | No | Group 4 by construction, sugar-free included |
| Gravy | Yes | Pan gravy group 3, granules and cubes group 4 |
| Crumpets | Yes | Traditional group 3, most packs group 4 on preservatives |
Common questions
- Is this peer reviewed?
- No. It is a classification exercise with a stated method, published with its limits and its working, and it should be read as that rather than as a study.
- How were the categories chosen?
- They are the ones most likely to appear on a weekly UK shop. That is a judgement rather than a sample, and it is stated as one.
- Can I cite this?
- Yes, with the method and the limits attached. It says that both answers exist within a category, not how common each is.
The only way to know is to read the shop.
One photo of a receipt does it line by line, which is the conclusion this audit points at and the reason to be sceptical of it.
Get the app