The Health Star Rating: five stars for sugar, half a star for butter

The stars on the front of your food aren’t a measure of how healthy it is. They’re the output of a formula that gives points back for things a factory adds, on a scheme companies join only if they want to.

What is the Health Star Rating and who actually built it?

The Health Star Rating is a voluntary front-of-pack scoring system, launched in Australia and New Zealand in June 2014, that rates packaged food from half a star to five. It was developed by federal, state, territory and New Zealand governments alongside public health groups, consumer groups and the food industry.

That last group is the part most men have never been told about. Industry didn’t observe the process. It sat inside it, on the technical committees that built the formula.

One example is a matter of official public record. Dr Gregory Gambrill was awarded a Member of the Order of Australia in June 2025 for service to public health and nutrition. His honours citation lists him as Industry Member of the Front-of-Pack Labelling Technical Design Working Group in 2012 and 2013, a member of the Health Star Rating Technical Advisory Group from 2014 to 2016, and Algorithm Development Lead in 2014.

The same citation lists his employment. He was Sanitarium’s Corporate Technical Manager from 2012 to 2018, and worked for Sanitarium continuously from 1981 to 2023.

Sanitarium is owned by the Seventh-day Adventist Church and was set up to make plant-based health foods. It was also one of the first companies to put the stars on its packs, and it now says more than 80% of its range carries four stars or higher.

Sanitarium didn’t design the Health Star Rating on its own, and nobody is claiming it did. The algorithm was adapted from an existing Food Standards Australia New Zealand nutrient profiling model and developed by a group that included government, public health, consumer and food-industry representatives. Ministers then approved it.

What’s true is narrower, and it matters. The citation, which is drawn from the material submitted for his nomination, records the algorithm development lead for 2014 as a senior technical manager at a packaged food company, and that company’s products are now marketed on the scores the algorithm produces. Nobody hid any of it. It was written into the governance from the start.

The first year set the tone. In February 2014 the government website that was to publish the ratings went offline within hours of launch, on instruction from the assistant health minister’s office, whose chief of staff held shares in a lobbying firm with food and beverage clients and later resigned. The industry body’s chief executive said he had called the minister that day to call the launch premature, but did not ask for the takedown. The site stayed dark for ten months.

The algorithm gives points back for things a factory adds

The scoring runs on subtraction, and that’s where it comes apart.

A product first collects baseline points for energy, saturated fat, total sugars and sodium. More points is worse. It then gets points taken back off for fruit, vegetable, nut and legume content, for protein, and for fibre. The final score is baseline minus those modifiers, and a lower score means more stars. The point tables are public, in the government’s own implementation guide.

Two of the three ways to claw points back are protein and fibre. Both are things you can tip into a mix. You don’t have to take the sugar out. You can leave the sugar exactly where it is and add protein and fibre until the arithmetic comes good.

Take Kellogg’s Nutri-Grain, which carries four stars. Per 100 grams it has 1640 kilojoules, 0.5g saturated fat, 24.1g sugars and 350mg sodium. Run those through the table and it collects 4 points for energy, none for saturated fat, 5 for sugar and 3 for sodium. Twelve baseline points.

Then the credits come off. It has 20.2g of protein per 100g, worth 10 points, and 5.2g of fibre, worth 5. Twelve minus fifteen is minus three, and minus three sits in the four-star band.

The bonuses are worth fifteen points against a total penalty of twelve. The credits are bigger than the problem.

Where does a breakfast cereal get 20 grams of protein and 5 grams of fibre? Turn the box over. The ingredients list runs: cereals (51%), sugar, wheat protein, maltodextrin, molasses, oat fibre, salt, barley malt extract. Isolated wheat protein and oat fibre are in there as ingredients in their own right, fifth and eighth on the list.

I’m not claiming all 20 grams came out of a bag. The algorithm scores the total protein and fibre in the finished product and makes no distinction about where any of it came from. Some of that is the wheat and the oats. Some of it isn’t, and the label won’t tell you the split. The formula doesn’t care. That’s the point: a manufacturer chasing a better score can move the total without touching the sugar.

One more thing in that calculation. Protein points are only awarded if a product’s baseline score comes in under 13. Nutri-Grain’s baseline is 12. Its sugar reading is 24.1g, and the next penalty tier starts above 24.6g. Six tenths of a gram more sugar per 100 grams would take the baseline to 13, disqualify all ten protein points, and drop the rating from four stars to two and a half.

It sits six tenths of a gram under the line.

Why does UP&GO have a 5 star health rating?

It’s solely based on the protein. Sanitarium’s UP&GO Protein Energize Choc Hit carries a five star rating, the maximum available, printed on the front of the carton.

Nearly every man who starts with me is drinking something like this. They see one of those chocolate smoothie milks, it says protein on the front, and they never turn it over to look at the carbs. They think that’s healthy. It’s marketing.

Its ingredients, in order: water, skim milk powder, soy protein, cane sugar, vegetable oils (sunflower, canola), wheat maltodextrin, milk protein concentrate, fructose, plant fibre, cocoa (0.5%), acidity regulator, vegetable gums, flavour, mineral, stabiliser, vitamins, salt. Cane sugar is fourth. Fructose is eighth. There is more maltodextrin in it than cocoa.

Here’s the arithmetic on its own nutrition panel, per 100mL:

Note what isn’t in that table. Dairy beverages are one of the two categories the guide rules ineligible for fibre points, so the gram of fibre per 100mL earns this drink nothing at all. Protein does the entire job on its own.

Before the protein score comes off, the drink sits at 1 point, which is three and a half stars. Protein alone carries it the rest of the way to five.

Two numbers on that panel are worth a second look. Energy is printed as 335 kilojoules per 100mL. The threshold below which energy scores zero is 335. Sodium is printed as 90 milligrams. The threshold below which sodium scores zero is 90. One kilojoule more and one milligram more on the label, and each picks up a penalty point.

Don’t overread it. Nutrition panels are rounded, so the printed figures aren’t necessarily the numbers the manufacturer ran the calculation on. What you can check is the carton. As declared, both readings sit on the thresholds rather than above them. I’m not going to tell you that was deliberate, because I can’t prove it and I don’t need to. I’ll tell you where the printed numbers land, and you can read them off the pack yourself.

The pack also carries a line most people never read. It says: “FORMULATED SUPPLEMENTARY FOOD. UP&GO is intended to supplement a diet which may be low in energy or vitamins and minerals.” That’s a regulated designation, not marketing. Read it next to the five stars on the same carton and decide for yourself what the badge is telling you.

The sugar figure on the front is per bar. The rating is per 100 grams.

Uncle Tobys Chewy Apricot muesli bars carry four and a half stars, and the front of the box advertises “Only 4.9g Sugar”.

That’s true. It’s also 15.6 grams of sugar per 100 grams, because a bar weighs 31.3 grams. The reassuring number on the front is calculated per bar. The star rating is calculated per 100 grams. Two different denominators on the same face of the same box, and the smaller one gets the big print.

The score works out like this. Per 100g the bar collects 4 points for energy, 1 for saturated fat, 3 for sugar and none at all for sodium, which is unusually low at 15mg. That’s 8 baseline points. Then 7.1g of protein takes 4 off and 11.5g of fibre takes 11 off. Eight minus fifteen is minus seven, which is the four-and-a-half star band.

Its fibre credit alone is worth more than its entire penalty score.

I’ll be fair to this one: a fair bit of that fibre is genuinely oats, and I’m not going to pretend a muesli bar is the same object as a flavoured milk drink. The problem isn’t that the bar is poison. It’s that a product which is 15.6% sugar by weight finishes half a star ahead of the Nutri-Grain and four stars ahead of a block of butter, and a man reading the front of the box is told the sugar figure is 4.9.

Why does butter get half a star?

Because butter has nothing to subtract. It’s roughly two thirds saturated fat, which loads it with baseline points, and it carries no fibre, no protein to speak of and no plant content to take any of them back off. Woolworths’ own-brand salted butter carries half a star.

Set that beside the drink above. Butter: cream and salt, half a star. A flavoured milk drink with cane sugar, fructose, maltodextrin and two added oils: five stars. Those two numbers come off different tables with different cut-offs, and the scheme’s own custodians will tell you they were never meant to be compared. Nothing on either pack says so. Same glyph, same shelf, and the category is printed nowhere. It’s the comparison the packaging invites and the one the system refuses to stand behind.

Be precise about what that does and doesn’t prove. It does not prove butter is better for you than margarine. That’s an argument about fats, and I’ve made my position on it elsewhere. What it proves is narrower and harder to argue with: the Health Star Rating has no measure of how processed a food is. Three ingredients or thirty, it can’t see the difference. It reads four nutrients and hands out a score.

The saturated fat penalty isn’t original to the Health Star Rating. It came in with the nutrient profiling model FSANZ already used for health claims, which was built on the low-fat consensus of its day, evidence that was a lot thinner than the confidence attached to it. That argument is still live, and plenty of people disagree with me on it. But you don’t have to settle the saturated fat question to see the problem here. The algorithm treats a nutrient count as a verdict on a food, and those are not the same thing.

The rating is voluntary, and low scorers wear it far less often

This is where a flawed formula becomes a marketing tool. It’s been measured twice.

Researchers looked at roughly 20,000 packaged foods sold here between 2014 and 2019. About 41% displayed a rating, and more than three quarters of those displayed ratings sat at three stars or above. Nestlé put stars on its Uncle Toby’s cereals, averaging 3.9. It left them off its confectionery, averaging 1.4.

By November 2025 the pattern held. A George Institute audit reported by the ABC found low-scoring products displayed a rating about 24% of the time, against 53% for higher scorers. A bad score roughly halves your chance of ever seeing a badge.

One honest qualification, because it cuts against the simple version of this story. The supermarkets are the exception. Coles, Woolworths and Aldi label their own private-label ranges broadly, which between them account for a majority of all the stars in the country. That’s why you can find a Woolworths butter wearing half a star at all. The selective display is a branded-manufacturer habit, not a universal one, and the retailers who label everything are the reason we can see the gap.

Milo is the cleanest illustration. In 2018 Nestlé carried 4.5 stars on Milo powder, a figure that only worked if you assumed three teaspoons stirred into 200mL of skim milk. On its own the powder rated 1.5. CHOICE noted about 13% of people used skim milk and 55% used full cream, and Nestlé removed the rating. Seven years on, Milo appears on the published list of major products carrying no Health Star Rating at all.

The system caught a product overstating itself, and the product stopped being scored. That exit is still open to every manufacturer in the country.

Eleven years of missed targets, and a mandate that still hasn’t landed

The scheme has never hit one of its own uptake targets. After ministers declined to make it compulsory, they set voluntary goals instead: 50% of products by November 2023, 60% by 2024, 70% by November 2025.

The February 2026 food ministers’ communiqué records the result. Uptake reached 39% in Australia and 36% in New Zealand. Ministers said the low uptake had “limited the system’s effectiveness and undermined consumer trust”, and voted by majority to ask FSANZ to prepare a proposal to mandate it.

That work is now underway as Proposal P1067. The first round of public consultation has closed, and FSANZ reported on 5 August 2026 that it received more than 380 submissions and is assessing them. If the proposal advances to a draft regulatory measure, a second consultation follows, expected later in 2026, and only then does it go back to ministers for a decision. A transition period for industry would follow any approval. No date has been set for any of it. My read is that we’re years from every packet carrying a star, but that’s my inference, not an announced timetable.

One line in that communiqué matters more than the headline. Ministers recorded “persistent misunderstandings about how to use it, particularly that the HSR is only intended for comparing similar products.” That’s the system’s own custodians conceding the central problem. The stars were built to rank cereal against cereal, not cereal against an egg. Nobody shopping reads them that way, and nothing on the front of the pack tells you to.

Meanwhile the sugar question sits where it has sat since the start. In 2019 CHOICE modelled what would happen if the algorithm penalised added sugars instead of total sugars: Nutri-Grain fell from four stars to one and a half, Milo cereal the same, Uncle Tobys Plus Protein the same. When ministers updated the algorithm in November 2020, they strengthened total sugars and sodium slightly and left added sugars out. Added sugars are still not an input to the formula.

What men over 40 should read on a packet instead of the stars

Read the ingredients list.

It’s the one part of the label an algorithm can’t dress up, because it has to say what’s in there in descending order by weight. When sugar is fourth on a drink and cocoa is tenth, you know what you’re holding whatever the front says. You’re looking for two things: how long the list runs, and how many items on it you wouldn’t keep in your own kitchen. Wheat protein, maltodextrin, plant fibre and stabiliser 452 aren’t pantry items. They’re formulation.

Then buy most of your food from the parts of the shop where nothing is selling you on the front of a box. Meat, eggs, fish, green vegetables, nuts. Jonathan, who came to me at 52 having carried the same 11 kilos since his mid-twenties, described what that turned into for him: “I go to the same supermarket and I go to exactly the same route, which is generally around the perimeter of the store where all the fresh produce is, and nothing in the middle of the store.” He lost the 11 kilos in 13 weeks and did no significant exercise to do it. Some of it can carry a rating (fresh fruit and veg get an automatic five stars, often on a shelf ticket), but none of it needs one and nothing was reformulated to earn it. Shop outside the contest and the referee stops mattering.

Nathan came to me after five years of adding a couple of kilos a year, training four or five days a week at CrossFit and getting nowhere. His description of his diet before we started: extremely healthy, clean 90% of the time, cooked from scratch, avoiding preservatives. He believed every word of it, and he dropped 10.6kg in twelve weeks once he changed what he was actually eating.

I’m not blaming a star rating for Nathan’s weight, and neither did he. He was reading signals that don’t measure what he assumed they measured. Once he knew what to look at instead, he changed it inside a fortnight. That decision was his, and it’s the only part of any of this you control.

A food that needs a badge to tell you it’s healthy has already answered the question

Real food doesn’t advertise. If the front of the pack is working that hard to sell you on the contents, look at the contents.

Your weight gets decided in the trolley before it gets near the kitchen, and the kitchen decides it long before the gym gets a say. Getting the food right is what moves the weight, and you can’t do that while you’re outsourcing the decision to a badge.

If you’ve been shopping off the front of the pack for twenty years and your waist has only gone one direction, the useful question isn’t which cereal to swap to. It’s what else you’ve been sold that you never thought to check.

I run a free 30-minute call. It’s the first step into my coaching program, so go in knowing that’s what it is. It’s worth booking if you’re carrying 10 kilos or more and you’re done experimenting. If you just wanted to know which cereal to buy, the ingredients list does that for free and you don’t need me. Bring the three or four things you eat most weeks, and we’ll go through what they’re actually doing to you. That part is useful whether or not you ever work with me. Book it here.

Frequently Asked Questions

Is the Health Star Rating accurate?

It’s accurate to its own formula and misleading as a guide to whether a food is good for you. It scores energy, saturated fat, total sugars and sodium as negatives, then gives points back for protein, fibre and plant content, so a manufacturer can add protein and fibre to a sugary product and lift its rating without removing any sugar. It uses total sugars rather than added sugars, and it has no measure of processing at all. Food ministers themselves noted in February 2026 that people misunderstand it, because it’s only designed to compare similar products.

Is the Health Star Rating a scam?

No, and it doesn’t need to be. It’s a real formula, published in full, that does exactly what it says: it scores four nutrients as negatives and gives points back for protein, fibre and plant content. The problem isn’t fraud, it’s arithmetic. A manufacturer can lift a score by adding protein without removing any sugar, the system has no measure of how processed a food is, and displaying the result is optional. Nobody hid any of that. It was built that way, in the open, with the food industry on the committees.

Why does UP&GO have a 5 star health rating?

Because its protein score pulls it down far enough to reach the top band. UP&GO Protein Energize Choc Hit is scored as a Category 1D dairy beverage and collects just 1 baseline point, because its printed energy and sodium readings sit on the thresholds below which they score zero. Its 7g of protein per 100mL then subtracts 4 more, taking it to minus 3 and the maximum rating. Dairy beverages aren’t eligible for fibre points under the rules, so its fibre earns nothing. Without the protein score it would sit at three and a half stars, despite containing cane sugar, fructose and maltodextrin.

Why does butter have a low health star rating?

Because it’s roughly two thirds saturated fat, which the algorithm treats as the main negative, and it has no fibre, protein or plant content to offset those points. Woolworths’ own-brand salted butter carries half a star. That reflects how the formula weights saturated fat, and the fact that a two-ingredient food has nothing to subtract. It is not a measure of how processed the product is, because the algorithm has no input for that.

Do all foods in Australia have to display a Health Star Rating?

No. The system has been voluntary since June 2014. Uptake reached 39% of products in Australia and 36% in New Zealand by November 2025, against a target of 70%. Food ministers asked FSANZ in February 2026 to prepare a proposal to mandate it (P1067). The first public consultation has closed with more than 380 submissions, which FSANZ said in August 2026 it was assessing, and a second consultation is expected before ministers make a final decision.

Why do some products have no health star rating at all?

Because displaying it is optional. Research found lower-scoring products displayed a rating about 24% of the time compared with 53% for higher scorers, so across the market the badge turns up far more often on products that score well. Nestlé removed the rating from Milo powder in 2018 after criticism of how it was calculated, and Milo still carries none. Supermarket private-label ranges are the exception and are labelled broadly.

Does the Health Star Rating account for ultra-processed food?

No. Level of processing isn’t an input, which is why an ultra-processed product can outscore a whole food on its protein score alone. The formula reads four nutrients off the panel and has no way to tell a two-ingredient food from a thirty-ingredient one.