When I saw an “A” on Nestle’s Nesquick cereal, I wanted to know how a product with 75% carbohydrates, including 22% of sugar, could be considered a healthy product. Let’s look at how the Nutri-Score of “A” is calculated and see how just 0.1g of sugar and 2mg of salt more would result in a D rating.
Let's look first at the nutritional information, as it is the basis for the Nutri-Score. The product box lists the amount per 100g, which is also the amount used by Nutri-Score. Serving size is not relevant, so please eat everything only in 100g increments.
Be aware that many label systems don't include fibre in the carbohydrates section, but some do. Though fibre is technically a carbohydrate, from a nutritional standpoint it makes sense to list separately. The body processes the non-fibre carbs differently than fibre.
When choosing what to eat, you should also look at the ingredients, though the Nutri-Score makes little use of them. The quality of the food is nearly irrelevant when calculating a Nutri-Score, only the resulting nutritional summaries. This allows highly processed foods to share scores with lesser processed foods, despite sufficient evidence that excess refinement and processing is detrimental to the nutritional value.
Nesquick Ingredients, listed on box: 53.1% Whole grain flour, Sugar, Cornmeal, Glucose syrup, 5.6% reduced fat cacao, calciumcarbonate, sunflower or palm oil, salt, natural flavours, (added essential nutrients*)
*The box proudly states it contains these essential nutrients, which feels disingenuous given they’re added artificial sources, not because of using ingredients that contain them.
Nutri-Score Negative Calculation
Nutri-Score evaluates a product on its negative and positive components. The negative aspects are the things it considers bad for your health, or at least where increased quantities are bad for you. The positive aspects are those it considers good for your health.
We start by calculating the negative component (N) of the cereal. This is Nutri-Scores attempt to capture the bad aspect of a food product. It comprises four facets: energy, saturated fats, sugar, and sodium. Low numbers are good, high numbers are bad.
The energy score is calculated as the energy density. Nutri-Score's calculation takes the 1595kJ and divides it by 335, rounding the result down. This gives Nesquick a value of 4 for energy. All negative component scores Nutri-Score are fit into a 0 to 10 range.
Why does Nutri-Score divide by 335? Though not mentioned in the documents, I think I figured out why. The highest energy content1 of 100g of food, assuming it’s all fat, is roughly 3700kJ. If we divide this by 11, for the 11-step range of 0 to 10, we get 336. Maybe somebody thought 5 was nicer than 6. Everything in nutrition is an approximate, so even +/-5% wouldn’t make a difference… yet.
It’s not clear why energy is one of the key negatives of Nutri-Score. Energy is the primary reason, or one of, why we eat. Maybe we’re missing a recommendation of how many 100g packets of food product we should eat per day?
Let's assume we eat enough to reach recommended intake of 2000cal (~8400kJ) per day — which is roughly in the middle of the recommendations for age and activity level. In this case, we should eat 6 packets of Nesquick cereal, which, with it’s A rating, promises we’re following the healthiest diet possible.
Our friendly demon, saturated fats, are scored on the same 0 to 10 range, with 10g or more of fat getting the maximum value. With only 1.1g of saturated fats, Nesquick gets a score of 1 here.
Unlike with energy, I could not figure out why 1g increments are used here. Both the World Health Organization2 and American Heart Association3 recommend a maximum related to total calorie intake, 10% and 6% respectively. It’s like they took the upper amount, 10%, and multiplied by the 100g size to get the 10g max. But if saturated fats are the bad guys here, why does it max out at 10g? Wouldn’t 20g be much worse?
This is, of course, assuming that saturated fats are inherently bad, which we have only mixed evidence of.
Our less friendly demon sugar gets a much bigger allowance than saturated fat. To calculate its score, Nutri-Score takes the total sugar, 22.4g, and divides by 4.5g, to get a value of 4. 22.4g is just a tad below the cut-off of 22.5g, where it’d get a value of 5.
Refined sugars are definitely bad for you. There’s less contention on this than saturated fats. A Nutri-Score value of 10 for sugar would be 45g of sugar. I'm not sure why this was chosen as the upper-bound. The FDA recommends4 a maximum of 50g of added sugar a day, with the AHA saying half of that, yet Nutri-Score’s max range covers that in a single 100g serving! Why is this not rated on the same scale of saturated fat?
With a Nutri-Score of A we'd assume we could eat only Nesquick cereal. We saw before we'd need 6 x 100g units of the cereal to reach our caloric needs, but that puts us at 134g of sugar a day. This grossly exceeds the recommended maximum sugar intake.
This doesn’t even consider the true glycemic aspect: how fast, or even how much, the body converts this food into blood sugar. The popular glycemic load and index is something that attempts to track this, a key tool in obesity and diabetes management. The simple concept of “sugar” is simply not enough when faced with natural sugars, added sugars, and refined carbs, such as white flour.
Processed foods have too much salt in them. For this negative value, Nutri-Score takes the salt content, 88mg in this case, and divides by 90mg, rounding down. This gives us a value of 0 for salt.
Nutri-Score doesn’t actually use division, instead creating tables of values. In their provided Excel worksheet, they also create series of if-statement expressions to reproduce the tables. Their reluctance to use math worries me.
Looking at EU guidelines on salt5, I see a common limit of 5g per day. Nutri-Score seems to set this maximum as the value of 10. Again, I’m unsure why there is a maximum here. A company could keep adding salt and not suffer any further negative score.
Another side-note, the reported amount of salt is to be first divided by 2.5, then by 90mg. I could not figure out why this division of 2.5 is here. If genuine, why was the 90mg not just stated as 225mg, removing the unusual pre-factor?
Nutri-Score Positive Calculation
Recognizing that not all ingredients are created equal, there are some food constituents which are given a positive value (P). These are fruits and nuts, fibre, and protein. Here, high values are good, and low values are bad.
The positive values only range from 0 to 5, unlike the negative range of 0 to 10. It’s as if Nutri-Score understands it’s unlikely for a company to focus on using the good ingredients.
“Fruits, vegetables, legumes, nuts and rapeseed, walnut and olive oils”
There’s a large category of natural food products, including fruits, nuts, vegetables and more that are good things to eat. Nutri-Score, fortunately, excludes starchy foods, such as potatoes, from this group. The food product gets a positive value based on how much of these ingredients it contains.
For Nesquick we have it easy. It’s 0. It doesn’t contain any of these good ingredients.
This "Fruits and more" category is based on Eurocode 2 food classification, which, as far as I can tell, was never really adopted as a standard, and it’s hard to even find a reference anymore. I used the Wayback Machine6. We do, however, have this extensive Foodex2 system which codifies all the information we’d ever want to know about food. Why would an EU nutritional labelling system not use the EU food catalog?
Fibre is our friend, and processed food is often lacking in fibre. Nestle’s whole grain products are bursting with fibre — a point, that despite still being sugar bombs, is a positive. Nutri-Score takes Nesquick’s 8.6g of fibre and divides by 0.9g, giving us well above the maximum value of 5.
Why do we divide by 0.9g? This appears to be based on a 25g recommended minimum common to EU recommendations7 — a product with the maximum value and average caloric amount would reach that amount of fibre. If we eat our 6 servings of Nesquick per day, we’d reach nearly 50g of fibre, which is well above the EU recommended minimum of 25g.
Nutri-Score says protein is good. Take our product’s 8.5g of protein, divide by 1.6g, reaching the maximum of 5, but just barely (the cut-off from 4 is at 8g).
The factor of 1.6g again seems to line up with EU recommendations. Europe doesn’t have a protein problem, but it’s an important factor in many parts of the world.
Let's summarize the components:
Negative Component (N) Energy: 4 Saturated: 1 Sugar: 4 Sodium: 0 Positive Component (P) Fruits et al: 0 Fibre: 5 Protein: 5
Now we're going to add them together. This is where the negative components get a positive value and the positive components get a negative value. This results in a golf-like scoring, where a product is aiming for a minimal value.
Score = N - P = 9 - 10 = -1
Yay, Nesquick's -1 is at the maximum allowed for a score of A.
That’s cutting it close, so let’s look at some boundary values:
- 22.4g of sugar is just below the 22.5g cutoff, 0.1g would push it to a B
- 88mg salt, just below the bound at 90mg, 2mg would also push it to a B
Actually, there is a further confounding factor, which is why this product so neatly puts itself in the desired range. If the total N value (negative part) is 11 or greater, then the score may not consider the positive proteins points*. If both the next sugar and salt threshold are breached, N would be 11, and the adjusted final value would be 6, giving Nesquick a Nutri-Score of D.
*Even with N >= 11, the protein can be considered if the fruit and veggie content is 80% of more.
So, 0.1g of sugar, and 2mg of salt are the difference between Nutri-Score A and Nutri-Score D. That sounds like a lot of incentive to game the system, which sounds like it was Nestle's goal8.
This arbitrary cut-off is a strange attempt to not reward protein unless the rest of the product is “healthy”. In spirit, this rule kind of makes sense, but the calculation is terrible, and open to manipulation. It’s not surprising though, given that Nutri-Score is based on industrial-political consensus rather than scientific consensus.
It’s uncertain why Nutri-Score considers only four negative and three positive aspects to food. We’ve accumulated a wealth of information about nutrition, from food nutrients to metabolism in the last several decades.
I understand the need for a simple label for consumers. I don’t understand why the score can’t be based on valid scientific research, with key consideration given to macronutrients and metabolism.
Creator, Serge Hercberg, believes that “Numerous scientific studies performed over many years… validate the algorithm underlying the Nutri-Score calculation.” Yet nowhere in the publications for Nutri-Score could I actually find any referenced study — there is however a 52-page document on how to use the logo. Pick up any nutritional study from the past 20 years and you’ll find numerous concepts which are absent from the Nutri-Score system.
Nutri-Score’s problems are magnified by the ease at which the system can be gamed. Manufacturers have a wide variety of replacement products, from fillers, to sugar alcohols, to modified starches, and other processed creations that while themselves are ignored in the Nutri-Score value, can lower the negative values.
Nesquick has an A rating, yet is only 0.1g of sugar, and 2mg away from a D rating. It allows for eating five times the amount of recommended amount of sugar per-day.
All is not bad though, as the box the product comes achieves the same score as the product inside. Carboard9 has no sugar, low calories, maybe a touch of salt, and a lot of fibre. That nets a score of around -5 (better than the cereal itself), resulting in an A score.