This is part of our series on packs: how to read them, what to trust, and what to take with a cautionary eye. Last time we were on what low, reduced and light actually mean. This one is about the row of stars.
The Health Star Rating scores a product between half a star and five. Australia and New Zealand brought it in back in 2014, and until now it has been voluntary, which is why it turns up on some products and not others.
It gets a hammering online. Some of that is deserved, and some of it comes from a misunderstanding of what the rating was ever trying to do.
What it is actually measuring
The Health Star Rating is not a verdict on whether a food is healthy in an absolute sense. It is primarily a tool for comparing the nutritional profile of similar packaged foods.
The score balances components worth limiting, energy, saturated fat, total sugars and sodium, against components worth encouraging, including fruit and vegetable content, fibre and protein. There are some differences in the way the algorithm treats broad categories of food, but the useful way to use the final rating is to compare like with like.
So a four star muesli bar is useful when you are standing in front of other muesli bars. It is not particularly useful compared with a piece of fish.
Which explains a lot of the outrage screenshots. Posting a highly rated packaged snack next to an unrated apple compares two things the system was never really designed to help you choose between. Fresh fruit and vegetables generally do not carry a rating because foods like these are not what the system was designed around.
The butter and margarine argument
This is the criticism that comes up most, and it deserves a proper answer rather than a dismissal.
Butter sits at the bottom of the range. A reduced-fat plant spread can score considerably higher. To a lot of people that reads as the system claiming a factory-made spread beats something a cow produced, and it is one of the biggest reasons people write the whole thing off.
On the particular question the rating is responding to here, replacing butter's saturated fat with predominantly unsaturated fat, it is broadly in step with the evidence. The scientific review prepared by the 2025 Dietary Guidelines Advisory Committee found that replacing butter with plant oils and spreads high in unsaturated fats decreases LDL cholesterol and may be associated with lower cardiovascular risk.
The old objection that margarine is loaded with industrial trans fats is also largely outdated in Australia. Manufacturers substantially reduced industrial trans fats decades ago, and FSANZ monitoring finds population exposure remains low. Some trans fats also occur naturally in dairy foods such as butter.
Worth knowing that trans fat will not usually appear separately on an Australian nutrition panel. It has to be declared when certain claims are made about cholesterol or particular types of fat, but otherwise you generally will not find a trans fat line on the pack.
Where the criticism lands
The critics have a real point underneath the butter argument, though it is not quite the one they are making.
Thinking on dairy fat has shifted. Cheese, yoghurt and milk now look broadly neutral overall for cardiovascular disease risk, including regular-fat versions, and the structure of the food appears to be part of the explanation.
One trial tested this directly. Participants consumed dairy fat either within cheese or in different forms where parts of the cheese matrix had been separated, including butter with calcium and casein added back. Over six weeks, total and LDL cholesterol were lowest when the dairy fat was eaten within the cheese matrix.
Similar dairy fat exposure, different food structure, different cholesterol response.
A meta-analysis of seven trials found the same general direction: compared with butter, dairy fat eaten as cheese produced lower total and LDL cholesterol.
That is the food matrix turning, and it is one of the more interesting developments in nutrition research.
Which makes butter the awkward case. It is dairy fat with most of that structure stripped away, so the finding that changes how we think about cheese does not automatically extend to butter.
A star rating cannot see any of this. It is reading selected components of a food, not its physical structure.
The other fair criticisms
A highly processed product may be able to improve its score by adding fibre or protein, even though the rating itself tells you nothing about the degree of processing.
Non-sugar sweeteners are not penalised in the way sugar is, so replacing sugar with them can improve a product's rating even though the rating itself tells you nothing about the use of those sweeteners.
And the system says nothing about how processed a food is, an issue that has received far more attention in nutrition research since the HSR was introduced.
None of those things necessarily makes the rating wrong. They are things it was never built to measure.
What it has got right
There is an effect most people never see, because it happens before the food reaches the shelf.
There is evidence that the system has prompted some manufacturers to reformulate products to achieve a better rating, including small reductions in nutrients such as sodium and sugar.
That changes what is available to buy, whether or not anyone reads the label.
For a voluntary scheme, that is a reasonable achievement.
The missing star trick
This one has a shelf life, so use it while it lasts.
Because the system has been voluntary, companies could simply leave the rating off when the score was poor. And the evidence suggests they have been much more likely to display good scores than bad ones.
In a FoodSwitch analysis of Australian supermarket products, only 16% of products calculated to receive half a star displayed their rating, compared with 61% of products receiving five stars.
So a missing star has been telling you something.
In a category where most comparable products are displaying a rating, the product with none may be worth turning over.
Two caveats. Some foods sit outside the system altogether, including alcohol, infant formula and certain special purpose foods, so an absent rating is not always a choice.
And this trick stops being useful if the rating becomes compulsory.
The voluntary bit may be on the way out
In February 2026, food ministers asked FSANZ to prepare a proposal on making the Health Star Rating mandatory after voluntary uptake reached only 39% of intended products in Australia.
FSANZ's first public consultation ran from 7 May to 5 July 2026, with another round of consultation still expected before a final decision.
So for now, the system remains voluntary. But that may not be the case forever.
How to actually use it
Compare similar products, not completely different foods.
Use the stars to narrow a shelf down to two or three options, then turn those over and read the per 100g column for whatever you actually care about, whether that is fibre, protein, sodium or sugar.
Stars to narrow, panel to decide.
That is a better use of the system than either trusting it completely or writing it off.
The takeaway: The Health Star Rating is best used to compare similar packaged foods rather than as a verdict on whether a food is healthy. Its treatment of butter versus unsaturated spreads is broadly consistent with the evidence on LDL cholesterol, but the algorithm cannot account for things such as food structure or degree of processing. It has also encouraged some product reformulation. Use the stars to shortlist within a shelf, then turn the pack over and read the panel.
The science behind this post:
- Australian Government Department of Health, Disability and Ageing (2026). Health Star Rating System. Food Ministers agreed in February 2026 to progress work towards mandating Health Star Ratings after voluntary uptake reached 39% of intended products in Australia and 36% in New Zealand. health.gov.au
- Food Ministers Meeting communique, 13 February 2026. FSANZ requested to prepare a proposal to mandate display of the Health Star Rating in the Australia New Zealand Food Standards Code, subject to two rounds of public consultation. foodministers.gov.au
- Food Standards Australia New Zealand. Proposal P1067: Health Star Rating System. First public consultation ran from 7 May to 5 July 2026. The Health Star Rating remains voluntary while the proposal is considered. foodstandards.gov.au
- Health Star Rating System Five Year Review Final Report (2019). Concluded the system was performing well overall and recommended it continue, alongside a package of reforms including algorithm refinements to address anomalies. healthstarrating.gov.au
- Health Star Rating System Five Year Review: Fats, oils and oil-based spreads. High saturated fat products such as butter score at the lower end of the distribution, while oils and spreads higher in polyunsaturated and monounsaturated fats can score considerably higher.
- The George Institute for Global Health / FoodSwitch analysis of Health Star Rating uptake in Australian supermarkets. Only 16% of products receiving 0.5 stars displayed their rating, compared with 61% of products receiving five stars, demonstrating selective uptake under the voluntary system.
- Australian Government and published evaluations of the Health Star Rating system. Evidence indicates adoption of the system has prompted some product reformulation, including changes in sodium and sugar content.
- Scientific Report of the 2025 Dietary Guidelines Advisory Committee, Part D Chapter 4: Food Sources of Saturated Fat. Replacing butter with plant-based oils and spreads high in unsaturated fatty acids decreases LDL cholesterol and may be associated with decreased cardiovascular disease risk. dietaryguidelines.gov
- Food Standards Australia New Zealand. Trans fatty acids. Population exposure to trans fats in Australia and New Zealand remains low, with substantial reductions in industrial trans fats in the food supply. foodstandards.gov.au
- Food Standards Australia New Zealand. Nutrition information panels. Trans fat declaration is required where certain claims are made about cholesterol or particular types of dietary fat but is not otherwise routinely required on the nutrition information panel. foodstandards.gov.au
- Feeney EL, et al. (2018). Dairy matrix effects: response to consumption of dairy fat differs when eaten within the cheese matrix. American Journal of Clinical Nutrition. Six-week randomised controlled trial examining approximately 40g of dairy fat per day in different dairy matrices. Total and LDL cholesterol were lowest when the fat was consumed within the cheese matrix.
- Systematic review and meta-analysis of randomised controlled trials on isoenergetic substitution of cheese with other dairy products (2023). Relative to butter, dairy fat eaten as cheese reduced fasting total cholesterol and LDL cholesterol across seven trials.
- Regular-fat and low-fat dairy foods and cardiovascular diseases: perspectives for future dietary recommendations (2025). American Journal of Clinical Nutrition. Summarises an international expert workshop and contemporary evidence indicating that milk, yoghurt and cheese are overall neutrally associated with cardiovascular disease risk, including regular-fat varieties.