Top banner ad available.

I suppose the political debate within the European Union about introducing a new food labeling system has already been harshly fought and lost, but it probably seems still worthwhile to introduce parts of the discussion in the information graphics community. (however, some sources still give it a small chance of success)

The "Traffic Light Labeling" or "Traffic-Light Nutritional Signposting" uses red, amber and green signals to show consumers, at-a-glance, whether a food product is high, medium or low in fat, saturated fat, sugars and salt. The traffic-light labels is specifically designed to enable quick comparisons between similar products, in an attempt to encourage consumers to make healthier food choices by choosing more products with green or amber lights than red.

The currently used food labeling provides the percentage guideline daily amounts (%GDAs) of a specific food product. It is claimed this %GDAs scheme creates a number of barriers for making healthy choices: first, they may suggest to the consumer that GDAs are daily targets to be aimed at, whereas in fact they are limits not to be exceeded. Second, the %GDA numbers provide values 'per portion' rather than per 100g or 100ml. In addition, brands often differ in their definition of what a 'portion' exactly constitutes. Lastly, the percentage amounts apply only to people of 'average' build', excluding an increasing amount of consumers.

In addition, a recent scientific study has shown that traffic light labels work much better than %GDA labels (or a mix of the two) across all socio-economic groups. This means that the traffic light scheme is less likely to widen dietary and health inequalities by only being useful to the most numerate or literate consumers.

People against the system (and most, if not all, major food brands) argue that the traffic lights cover a too broad numerical range to be meaningful. For instance, a single soup spoon of both full-fat spread or low-fat spread would would be both colored red. Secondly, some foods such as olive oil and walnuts would be displayed as red lights in spite of their health-giving qualities. In other words, if a consumer stucks to a green lights diet only, she would actually not be eating a balanced diet, as one needs a certain amount of carbohydrate and fat.

Taken these arguments into account, and from an information graphics point of view, is the traffic light metaphor the right way to communicate nutritional information to people?

See also:
. McDonalds Menu Charts
. Nutrition Data Visualization
. Food Measuring Plate
. How Does 200 Calories Look Like?
. Cheerios Graph Critique
. What the World Eats
. How Much Sugar is in Your Food?
. Eat Local Advertisement
. Beer Consumption per Country
. How to Feed the World

Parts of this post were based on an FT column by Michael Skapinker. Image above taken from Alex Segre's Flickr stream.


I believe the UK government are planning to have their own labelling scheme, or rather to extend the current one.

Sat 03 Jul 2010 at 12:03 AM
Matt Greenslade

What about the color blind?

Sat 03 Jul 2010 at 12:04 AM

I don't know if this still fits with current advice, but one 'rule of thumb' for a nutritious meal is supposed to be to divide the plate into quarters - 1/4 starch, 1/4 protein, 1/2 non-starchy veg. I know this wouldn't be much use for 'snack' products, but for ready meals etc. how about showing a circle with how the product compares to that 'ideal'? I have a picture in my head of how that might work... might post something later if I get time...

Sat 03 Jul 2010 at 12:47 AM

@Matt: The traffic light labeling system was invented by the UK Food Standards Agency, so most probably the same.

@Jan: Here you go: (via

Sat 03 Jul 2010 at 1:44 AM

Dreadful - 1. color blind inaccessible. 2.What's worse than a pie chart? A graphic that looks like a pie chart bu isn't. 3. It's all context. If the only thing I eat in a day is the item above, I will lose weight, so why are the 218 calories bad? 3.How about just highlighting the really bad stuff (trans fats for example, foods that are far in excess of calories than what you would expect - like milk shakes..

Still, I agree it may be an improvement on what they had before.

Sat 03 Jul 2010 at 2:55 AM

@Alex: Small adjustment though: the traffic light system does not propose to use a pie chart like representation. It just proposes, if I understand it correctly, to color-label the values between specific value bands. It is completely open how the colors are used (in a circle, row or column of values, ...).

Sat 03 Jul 2010 at 3:11 AM

Sounds like the goal of the changes is to aid in comparison between products - a goal that is rendered impossible by the incongruencies in the data. I don't think the change makes sense and I suspect that if passed, it will actually mislead people more than it helps them.

Sat 03 Jul 2010 at 8:31 AM

this is simply crazy. Example: light coke. it has Zero sugars, since the sweet flavor is made with aspartame. For our bodies aspartame can be much more difficult to digest.

Sun 04 Jul 2010 at 3:14 AM

It's a good system, has been user-tested in the UK with a range of consumer groups and performs well. But it has to be taken in context, as part of a wider programme of public health education. No labelling system applied to individual foods can possibly be expected to address the need for overall balance in the diet, but it can aid comparisons between different products of the same general type.

As for trans fats, the solution is simple. Ban them.

Sun 04 Jul 2010 at 8:11 PM

Stupid stupid stupid. Besides the visual flaws others mentioned above, you're trying to give advice to people of all different heights, physiques, and states of health. A high-fat, high-protein, high-calorie, nutrient-packed meal would be bad for a 45-year-old sedentary woman, but not for a athletic, growing teenage girl.

Just give the information and trust that people are mature and can take care of theirselves.

Mon 05 Jul 2010 at 11:51 AM

All of this labelling stems from a key assumption: that modifying food consumption based on labels will lead to better health outcomes. Flawed data-trawling epidemiological studies aside, where is the evidence? Show me the data!

Mon 05 Jul 2010 at 2:56 PM
Commenting has been temporarily disabled.